Department: Engineering : Computer Science ![[clear]](close-x.png "Remove this limiter")
171 matches in the database.
These are records: 1 - 30.
Did you mean instcode:ucii?
[1] [2] [3] [4] [5] [6]
1.
ABEYSINGHE, RUVINI PRADEEPA.
SIGNATURE FILES FOR DOCUMENT MANAGEMENT.
Degree: MS, Engineering : Computer Science, 2001, University of Cincinnati
► A document management system, called SDMS (Signature-file Document Management System), has been…
(more)
▼ A document management system, called SDMS (Signature-file Document Management System), has been designed and implemented. This system was designed based on the signature file method to store documents in a database file and to retrieve them according to the user queries. We implemented an application, Email Organizer, based on the SDMS. This system will be useful for individuals as well as organizations where their emails need to be stored, indexed, categorized and then retrieved efficiently based on the content. User queries can be formulated by using keywords from the text that are interrelated by the Boolean operators AND, and OR. A novel algorithm based on linear feedback shift registers (LFSR) has also been developed and implemented for generating word signatures. The objective of incorporating this approach is to reduce its false drop ratio and to improve its signature generation speed. This algorithm has been compared with the hashing functions for performance. Experiments revealed that the LFSR algorithm is comparable to the hashing algorithm in terms of signature generation time and false drop ratio.
Advisors/Committee Members: Han, Dr. Chia-Yung.
Subjects: Computer Science
Keywords: SIGNATURE FILES; TEXT RETRIEVAL METHODS; INDEXING; DOCUMENT MANAGEMENT SYSTEM; PSEUDO RANDOM NUMBER GENERATION
More Like This
2.
Agarwal, Aarti Subhash.
Use of Query Control and Location for Routing in Mobile Ad Hoc Networks.
Degree: MS, Engineering : Computer Science, 2002, University of Cincinnati
► A mobile ad hoc network is a collection of wireless mobile nodes…
(more)
▼ A mobile ad hoc network is a collection of wireless mobile nodes dynamically forming a network without the use of any existing network infrastructure or centralized administration. As the nodes are mobile, the network topology is dynamic leading to frequent unpredictable connectivity changes. It is critical to route packets to destinations effectively without generating excessive overhead. This presents a challenging issue for protocol design since the protocol must adapt to frequently changing network topologies in a way that is transparent to the end user. A class of routing protocols called on-demand protocols has received a lot of interest because of their low routing overhead. In this thesis, we study techniques that can reduce this routing overhead even further. The on-demand protocols depend on query floods to discover routes whenever a new route is needed. Network-wide floods incur substantial overhead. Techniques have been proposed to contain the flood in a limited region where a route to the destination is highly likely to be found. Techniques have also been proposed to reduce redundant broadcasts. We propose various mechanisms to improve on these existing techniques. We propose adaptive mechanisms that utilize prior routing histories, mobility pattern and network load to choose the area in which the query flood should be contained. In addition, we propose a technique that utilizes the neighborhood information to reduce or eliminate redundant broadcasts. We evaluate their performances in isolation and in tandem. In the next part of the thesis, we turn our attention to use of location information for routing. In has been shown in prior work that availability of location information can substantially reduce routing overheads. However, equipping all mobile nodes with GPS or other positioning system is not a cost effective proposition. We develop and evaluate a localization technique that can localize mobile nodes even when only a fraction of nodes in the network has direct access to location information. The rest of the nodes localize themselves by hearing radio beacons emitted from the nodes that have access to direct location information. We evaluate the localization accuracy using our technique via simulations. We show that the accuracy is better than the radio range even when only about 10% of the network nodes have direct access to location information. To be useful for routing, each node in the network must learn the location of every other node. However, disseminating this information sufficiently frequently can incur large overhead. To counter this, we develop a mobility prediction and location tracking model based on the dead-reckoning navigation technique. Here, both current location and movement model of mobile nodes are disseminated via flooding. Every other node is now able to track the movement with some accuracy until the movement model changes bstantially. We use this technique along with a geographic routing protocol to solve the unicast routing problem in ad hoc networks. Simulation studies show that it performs better than well-known on-demand routing protocols.
Advisors/Committee Members: Das, Dr. Samir R.
Subjects: Computer Science
Keywords: Ad Hoc Networks; Routing; Dead Reckoning Model; Location based routing
More Like This
3.
ALI, NAWAB.
IMPROVING PAGING PERFORMANCE OF MEMORY-INTENSIVE APPLICATIONS WITH MEMORY ACCESS PATTERN GUIDED PAGE REPLACEMENT.
Degree: MS, Engineering : Computer Science, 2004, University of Cincinnati
► Memory-intensive applications have large data sets which cannot be accommodated in the…
(more)
▼ Memory-intensive applications have large data sets which cannot be accommodated in the available main memory. The Virtual Memory (VM) sub-system frequently moves parts of their data sets between the main memory and the secondary storage (swap disk). This results in a significant increase in the paging I/O traffic of the applications. Since the disk sub-system is orders of magnitude slower than the main memory, this increase in I/O traffic severely degrades the applications' performance. In this research, we reduce the paging I/O traffic of memory-intensive applications by exploiting the patterns in their memory access behavior. Most out-of-core applications exhibit clear patterns in the manner in which they access their virtual address space. We collect and analyze the memory access patterns of these applications and use this information to guide the page replacement algorithm. This technique, when coupled with memory prefetching and page clustering results in a substantial performance improvement for large, memory-intensive applications. We conducted trace-driven simulations on the NAS Parallel Benchmarks (NPB) to evaluate our paging optimization algorithms. The simulation results show a considerable improvement in the performance of our benchmarks. Our page replacement strategy reduced the average paging I/O traffic by 52%. On combining memory prefetching with our pattern guided page replacement algorithm, we observed an average performance improvement of up to 69%. These results are encouraging, and we hope to further improve the VM performance of memory-intensive applications by adding hardware support to our strategy.
Advisors/Committee Members: Hu, Dr. Yiming.
More Like This
4.
AMALAPURAPU, SUCHITRA S.
A STATISTICAL ANALYSIS OF AMINO ACID CHANGES IN THE HUMAN GENOME.
Degree: MS, Engineering : Computer Science, 2003, University of Cincinnati
► The advent of modern sequencing techniques has facilitated sequencing of the entire…
(more)
▼ The advent of modern sequencing techniques has facilitated sequencing of the entire human genome. Single Nulceotide Polymorphisms (SNP’s) are of tremendous interest to biologists since SNP’s in the coding region of the DNA have been found to be a major cause of various genetic disorders. A detailed analysis of all the existing SNP’s in the human genome could throw light on the causes and effects of inheritance of various genetic diseases. This information can be used by pharmacogeneticists to develop new drugs or techniques to cure genetic disorders, which could revolutionize disease treatment. We have made an attempt to produce a summary of the amino acid jumps related to the Single Nulcleotide Polymorphisms (SNP’s) in the human genome. The summary of observed amino acid jumps and theoretical amino acid jumps have been compared and the discrepancy in their results has been recorded. The same analysis can also be extended to the genome of other organisms.
Advisors/Committee Members: CHENG, DR. YIZONG.
Subjects: Computer Science
Keywords: SNP; bioinformatics; DNA; Contig; genetics
More Like This
5.
ANDRA, SHESHU KALYAN Chakravarthy.
TRACKING PHENOMENA WITH SENSOR NETWORKS.
Degree: MS, Engineering : Computer Science, 2005, University of Cincinnati
► The recent emergence of the sensor networks technology is significantly impacting the…
(more)
▼ The recent emergence of the sensor networks technology is significantly impacting the capabilities for automated distributed monitoring of environments. The large deployments of sensors for fine-grained data gathering makes it practically infeasible to transport all data to a central site for processing. Thus, effective localized communication and computation mechanisms need to be designed. Aggregation trees covering complete terrains have been shown to be an effective way of performing computation and communication in sensor networks. This thesis presents a reactive approach to constructing localized aggregation trees, limited to the regions of activity, wherein the sensors within close proximity that sense some phenomenon construct a local aggregation tree for their localized computational use, send the conclusion to the base station, and then the tree structure may be discarded. Four characteristics of a mobile phenomenon, namely geographical contour, location, velocity and predicted direction, were computed using this tree, which are used for tracking purposes. The working of the algorithms were shown through software simulations and effectiveness of the suggested approach was demonstrated.
Advisors/Committee Members: Bhatnagar, Dr. Raj.
More Like This
6.
ASHOK, RAMYA.
A DATABASE SYSTEM TO STORE AND RETRIEVE A CONCEPT LATTICE STRUCTURE.
Degree: MS, Engineering : Computer Science, 2005, University of Cincinnati
► Data mining techniques are utilized to discover noteworthy and unrecognized associations between…
(more)
▼ Data mining techniques are utilized to discover noteworthy and unrecognized associations between data items. Organization of discovered concepts in the form of a lattice-structure has many advantages from the perspective of knowledge discovery. It facilitates insights into dependencies among different concepts mined from a dataset. Theoretical foundations of lattices of concepts have been extensively studied [18, 19] and have been implemented with a number of different algorithms. Any of these mining tasks can generate a very large number of concepts and a database organization is required for an efficient storage and retrieval of parts of the lattice. The task of efficiently organizing and retrieving various nodes of a lattice is the focus of research described in this thesis. A concept is a pair that consists of a set of objects and a particular set of attribute values shared by the objects. For our task we assume that the large numbers of concepts are available in the form of a dataset generated as a result of any one of the various mining algorithms [17]. Our implementation facilitates the process of storing the concepts in the form of a Lattice Structure in a database system. A Lattice Structure is a representation of concepts in the form of a partial-order hierarchy. The placement of concepts in the form of nodes at different levels in a Lattice Structure is determined by one of many possible partial ordering relations. Such lattices are known as concept Lattices. A concept Lattice Structure facilitates navigation of data and concept node connectivity. The exploration of concept space is accomplished by traversing various paths in the Lattice Structure. Multiple concepts that share a common object and a set of attributes are known as overlapping concepts. The Lattice Structure systematically arranges the overlapping concepts to facilitate insights into the deeper structural relationships among concepts. This enables generalization of concepts, thereby assisting in meaningful information retrieval and knowledge discovery. By transforming concepts into a lattice hierarchy, we demonstrate that a Lattice Structure is a powerful mechanism of representing and retrieving conceptual entities. In the research described in this thesis, a database system has been implemented to store concepts to build a concept Lattice Structure. Querying techniques have been implemented to retrieve concepts from the database in order to demonstrate node navigation and to display retrieval paths. Lattices built with two different partial ordering relations have been used to demonstrate our system. The database can easily accommodate other partial ordering relations. Further, the time-complexity for datasets of varying sizes has been analyzed and empirical and analytical results are presented in the thesis. The complexity is shown for both the partial ordering relations mentioned above. The concept Lattices can be applied in application areas such as search engines like Yahoo, Google, Web Crawler, etc., organizing email messages, text-based searches and knowledge discovery in textual databases. These application areas are explained in detail in the next chapter and how our research work implementation differs in each of these application areas.
Advisors/Committee Members: Bhatnagar, Dr. Raj.
Subjects: Engineering, Mining
Keywords: Concepts; Hierarchy; Dataset; Lattices; Lattice Structure; Database; Data Mining
More Like This
7.
Aswathanarayanan, Subramanian.
A STUDY OF THE EFFECT OF SINGLE NUCLEOTIDE POLYMORPHISMS IN HUMAN GENOME ON THE SECONDARY STRUCTURE OF PROTEINS.
Degree: MS, Engineering : Computer Science, 2002, University of Cincinnati
► Geneticists throughout the world are involved in deciphering the genetic code of…
(more)
▼ Geneticists throughout the world are involved in deciphering the genetic code of various organisms including that of human beings. With the discovery of more single nucleotide polymorphisms (SNPs) everyday and the role played by them in human diseases, a lot of attention is being paid to understand their effect on the structure and biological function of proteins. This will not only help us in understanding the human biology better, but will also provide us with a way to fight genetic and hereditary diseases. In this thesis we have studied the effect of SNPs on the helix structure, which is one of the most important structures of all. The study has been done by taking in to consideration, the hydrophobicity of amino acids involved. Overall, our results indicate that most of the SNPs are structure preserving in nature. We have also investigated if change of secondary structure, especially helix, in genes might be the responsible for disease in the human body. Our results indicate that a break in helical structure of a gene might change its biological function.
Advisors/Committee Members: Cheng, Yizong.
Keywords: SNP; Single Nucleotide Polymorphism; Secondary Structure of Protiens; Human Genome
More Like This
8.
AULUCK, NITIN.
REAL-TIME SCHEDULING ALGORITHMS FOR PRECEDENCE RELATED TASKS ON HETEROGENEOUS MULTIPROCESSORS.
Degree: PhD, Engineering : Computer Science, 2005, University of Cincinnati
► The problem of real-time scheduling has been very well researched for the…
(more)
▼ The problem of real-time scheduling has been very well researched for the single processor case and the identical multiprocessor case. There are a number of efficient algorithms and schedulability tests that enable the system designer to model his/her application as a set of real-time tasks. However, it may very well be the case that the multiprocessor system consists of processors with different speeds and configurations. This heterogeneity makes the scheduling problem much more complex. Real-time scheduling on heterogeneous multiprocessors has not received much attention in the scheduling literature. We propose several algorithms for scheduling a set of precedence related tasks on heterogeneous multiprocessors. In the first section, we propose a reliability driven algorithm for scheduling periodic tasks on heterogeneous systems. We relax the assumption that all the processors in the system are equally reliable. We introduce a new metric called cost of dependability. The periodic tasks are assigned to processors in such a way that the reliability of the system is maximized. We observe that our scheme generate schedules that are consistently more reliable than the schedules generated by algorithms that do not take the reliability of the processors into account. In the second section, we propose a duplication based algorithm for scheduling a set of soft deadline tasks on a system of heterogeneous multiprocessors. We observe that communicating tasks assigned to different processors have to incur a delay by using the inter-processor communication channel. By duplicating any of these two tasks on either assigned processor, we can cut down on that delay. Hence tasks can start (and hence finish) earlier. This leads to a larger number of tasks meeting their deadlines and hence an increase in the guarantee ratio of the real-time application. The next section extends the concept of task duplication to a set of hard deadline tasks. The deadlines are hard in that even a single deadline miss could lead to catastrophic results. By employing task duplication, we observe that a larger number of task sets are able to finish execution before their deadlines. Hence, the success ratio of the application experiences an increment. Based on our simulation results, we observe that our proposed algorithm offers a higher success ratio as compared to the other algorithms that do not use task duplication. It may very well be the case that the real-time application consists of a mix of tasks with hard and soft deadlines. In the next section, we propose an integrated scheme for scheduling a mix of hard and soft deadline tasks on heterogeneous multiprocessors. We observe that our scheme offers a better schedulability as compared to the other algorithms in the literature when sufficient processors are available and when communication is a dominant factor in the system. Moreover, our algorithm is scalable in that the application can be scheduled even if sufficient processors are not available for the initial cluster generation.
Advisors/Committee Members: Agrawal, Dr. Dharma P.
Subjects: Computer Science
Keywords: Real-Time Systems; Task Scheduling; Heterogeneous Multiprocessors; Hard and Soft Deadlines; Scalability; Task Duplication
More Like This
9.
BALAPA, MANOHAR.
A HEURISTIC FLIGHT PATH PLANNER FOR A SMALL UAV ATTEMPTING TO FIND A SINGLE TARGET IN MINIMUM TIME.
Degree: MS, Engineering : Computer Science, 2008, University of Cincinnati
► The work presented here is a part of ongoing research into flight…
(more)
▼ The work presented here is a part of ongoing research into flight path planning for autonomous fixed wing aircraft. The problem of a single Unmanned Aerial Vehicle attempting to find a single stationary target in minimum time is examined. The problem is tailored to a specific type of UAV i.e. a low cost high maneuverability propeller driven fixed wing aircraft of wingspan 2-20 meters. The objective area to be searched for the target is split into cells the size of the area occupied by the target and occupancy probabilities for each cell are assigned. A methodology for extending the problem to the case of a target that moves but is constrained to move only in the objective area is presented. The search problem is set up as an infinite horizon optimization problem. The problem is converted to a finite horizon shortest path problem. The NP hard shortest path problem is then solved using a two step look ahead heuristic algorithm, which ultimately yields an approximately optimal solution of the search problem itself. Finally simulation results are presented which show the heuristic algorithm generated path’s efficacy compared to a zigzag path and a path generated by a greedy algorithm.
Advisors/Committee Members: Frenandez, Dr. Emmanuel.
Subjects: Computer Science
More Like This
10.
Banerjee, Torsha.
Energy Efficient Data Representation and Aggregation with Event Region Detection in Wireless Sensor Networks.
Degree: PhD, Engineering : Computer Science, 2008, University of Cincinnati
► Unlike conventional networks, wireless sensor networks (WSNs) are limited in power, have…
(more)
▼ Unlike conventional networks, wireless sensor networks (WSNs) are limited in power, have much smaller memory buffers, and possess relatively slower processing speeds. These characteristics necessitate minimum transfer and storage of information in order to prolong the network lifetime. In this dissertation, we exploit the spatio-temporal nature of sensor data to approximate the current values of the sensors based on readings obtained from neighboring sensors and itself. We propose a Tree based polynomial REGression algorithm, (TREG) that addresses the problem of data compression in wireless sensor networks. Instead of aggregated data, a polynomial function (P) is computed by the regression function, TREG. The coefficients of P are then passed to achieve the following goals: (i) The sink can get attribute values in the regions devoid of sensor nodes, and (ii) Readings over any portion of the region can be obtained at one time by querying the root of the tree. As the size of the data packet from each tree node to its parent remains constant, the proposed scheme scales very well with growing network density or increased coverage area. Since physical attributes exhibit a gradual change over time, we propose an iterative scheme, UPDATE_COEFF, which obviates the need to perform the regression function repeatedly and uses approximations based on previous readings. Extensive simulations are performed on real world data to demonstrate the effectiveness of our proposed aggregation algorithm, TREG. Results reveal that for a network density of 0.0025 nodes/m2, a complete binary tree of depth 4 could provide the absolute error to be less than 6%. A data compression ratio of about 0.02 is achieved using our proposed algorithm, which is almost independent of the tree depth. In addition, our proposed updating scheme makes the aggregation process faster while maintaining the desired error bounds. We also propose a Polynomial-based scheme that addresses the problem of Event Region Detection (PERD) for WSNs. When a single event occurs, a child of the tree sends a Flagged Polynomial (FP) to its parent, if the readings approximated by it falls outside the data range defining the existing phenomenon. After the aggregation process is over, the root having the two polynomials, P and FP can be queried for FP (approximating the new event region) instead of flooding the whole network. For multiple such events, instead of computing a polynomial corresponding to each new event, areas with same data range are combined by the corresponding tree nodes and the aggregated coefficients are passed on. Results reveal that a new event can be detected by PERD while error in detection remains constant and is less than a threshold of 10%. As the node density increases, accuracy and delay for event detection are found to remain almost constant, making PERD highly scalable. Whenever an event occurs in a WSN, data is generated by closeby sensors and relaying the data to the base station (BS) make sensors closer to the BS run out of energy at a much faster rate than sensors in other parts of the network. This gives rise to an unequal distribution of residual energy in the network and makes those sensors with lower remaining energy level die at much faster rate than others. We propose a scheme for enhancing network Lifetime using mobile cluster heads (CH) in a WSN. To maintain remaining energy more evenly, some energy-rich nodes are designated as CHs which move in a controlled manner towards sensors rich in energy and data. This eliminates multihop transmission required by the static sensors and thus increases the overall lifetime of the WSN. We combine the idea of clustering and mobile CH to first form clusters of static sensor nodes. A collaborative strategy among the CHs further increases the lifetime of the network. Time taken for transmitting data to the BS is reduced further by making the CHs follow a connectivity strategy that always maintain a connected path to the BS. Spatial correlation of sensor data can be further exploited for dynamic channel selection in Cellular Communication. In such a scenario within a licensed band, wireless sensors can be deployed (each sensor tuned to a frequency of the channel at a particular time) to sense the interference power of the frequency band. In an ideal channel, interference temperature (IT) which is directly proportional to the interference power, can be assumed to vary spatially with the frequency of the sub channel. We propose a scheme for fitting the sub channel frequencies and corresponding ITs to a regression model for calculating the IT of a random sub channel for further analysis of the channel interference at the base station. Our scheme, based on the readings reported by Sensors helps in Dynamic Channel Selection (S-DCS) in extended C-band for assignment to unlicensed secondary users. S-DCS proves to be economic from energy consumption point of view and it also achieves accuracy with error bound within 6.8%. Again, users are assigned empty sub channels without actually probing them, incurring minimum delay in the process. The overall channel throughput is maximized along with fairness to individual users.
Advisors/Committee Members: Agrawal, Dr. Dharma.
Subjects: Remote Sensing
Keywords: wireless sensor; data aggregation; polynomial regression; energy efficiency; mobile; faulty cognitive radio
More Like This
11.
BATRA, SHALINI.
DISCOVERY OF CLUSTERS IN SPATIAL DATABASES.
Degree: MS, Engineering : Computer Science, 2003, University of Cincinnati
► Spatial data mining is discovery of interesting relationships and Characteristics that may…
(more)
▼ Spatial data mining is discovery of interesting relationships and Characteristics that may exist implicitly in databases. Data mining systems aim to discover patterns, find unexpected results and correlations and extract useful information recorded in databases. The branch of data mining that deals with location of data is spatial data mining. Spatial data mining is a highly demanding field and has huge data that is collected from various applications like Remote Sensing, Geographical Information Systems (GIS), Medical Equipments, Video Images, Computer Cartography, Environmental Assessment and Planning. In this thesis we consider spatial databases and discover spatial clusters of similar characteristics. The clusters discovered are at different levels of granularity. There has been relevant work done in the area of spatial data mining and there exist algorithms for finding clusters in spatial databases. We find regions of similar characteristics in spatial databases with no prior information from the user. We use quad tree data structure to summarize spatial locality. In a quad tree, there are always four children per node and number of leaf nodes are always in power of 4. This structure allows effective information traversal in the tree. We consider each data point to be a leaf node in quad tree. We define parameters to measure similar characteristics among nodes. Each node contains interesting spatial facts discovered for the sub tree lying under it. Further, we develop an algorithm to simultaneously examine any associations across the clusters formed by different attributes in the same spatial context.
Advisors/Committee Members: Bhatnagar, Dr. Raj.
Subjects: Computer Science
Keywords: spatial data mining; clustering; quad-tree; clusters
More Like This
12.
BEMALKHEDKAR, KALYAN.
AN OBJECT ORIENTED FRAMEWORK FOR BITSTREAM GENERATION FOR MULTI-TECHNOLOGY FPGA CHIPS.
Degree: MS, Engineering : Computer Science, 2005, University of Cincinnati
► Technologies in FPGA (Field Programmable Gate Array) chips have rapidly evolved over…
(more)
▼ Technologies in FPGA (Field Programmable Gate Array) chips have rapidly evolved over the past few decades. The architectures of these FPGA chips are constantly evolving. To increase their applicability, newer analog, digital and mixed signal components are been integrated. The software tool chain used to program application specific functionality onto these chips also needs to evolve simultaneously to support these new components and architectures. This software tool chain starts with high level design entry and is completed by the generation of bitstream for programming the chip. In this thesis, we present a novel object oriented framework approach for bitstream generation: XBits. XBits takes in a placed and routed circuit description and generates the bitstream for an FPGA chip whose architecture is specified in our custom Architecture Description File Format. This format allows us to specify varying FPGA architectures and components which can be easily integrated into rest of the framework.
Advisors/Committee Members: Tomko, Karen.
Subjects: Computer Science
Keywords: FPGA Bitstream Generation Object Oriented Framework
More Like This
13.
Bhargava, Sonali.
Generic and Scalable Security Schemes for Ad Hoc Networks .
Degree: MS, Engineering : Computer Science, 2002, University of Cincinnati
► An Ad Hoc network is a collection of wireless, mobile nodes that…
(more)
▼ An Ad Hoc network is a collection of wireless, mobile nodes that dynamically form a network without the use of centralized, fixed network infrastructure. Inherent characteristics of an Ad Hoc network such as dynamic topology and limited physical security poses severe security challenges to the network. Hence, these networks demand much stronger security mechanisms than the traditional, wired and static networks. Well established contemporary routing protocols seem to adapt to the dynamic conditions as well. However, they provide either no security mechanisms at all, or have only partial solutions for protecting the dynamic routing framework. It is hard to achieve security and robustness in the routing protocols at the same time in such networks. Several issues have to be understood and addressed before devising a security mechanism. Moreover, challenges involved in addressing attacks differ from one protocol to the other. This thesis targets at securing reactive routing protocol AODV. The routing protocol is vulnerable to two kinds of attack: External and Internal attack. We have discussed some existing external attacks and possible malicious behavior from compromised nodes. To mitigate the attacks, we propose a dual level security model. On the first level, we have External Attack Detection Model(EADM), that secures the network with authentication and confidentiality that rely on mutual trust between nodes. And on the second level, Intrusion Detection Model (IDM) identifies the misbehaving nodes using the knowledge base and Response Model (RM) isolates these nodes from the network.
Advisors/Committee Members: Agrawal, Dr. Dharma P.
Subjects: Computer Science
Keywords: Ad Hoc Networks; security; AODV; IP sec; wireless networks
More Like This
14.
BIAN, HAIYUN.
FINDING INTERESTING SUBSPACE CLUSTERS FROM HIGH DIMENSIONAL DATASETS.
Degree: PhD, Engineering : Computer Science, 2006, University of Cincinnati
► Data mining focuses on finding previously unknown yet potentially useful, hidden patterns…
(more)
▼ Data mining focuses on finding previously unknown yet potentially useful, hidden patterns from large amounts of data. Clustering is one of the most commonly used unsupervised data mining techniques, and it has been successfully applied to find groups of similar data points in many applications. However, conventional clustering algorithms sometimes fail to find meaningful clusters when the dataset has dozens of attributes because the high dimensionality makes the data space very noisy. Subspace clustering is a solution to this problem that can find clusters in subsets of all the dimensions. Different subspace clusters may be formed in different subsets of dimensions, and a single data point may belong to multiple subspace clusters. A subspace clustering algorithm not only searches for the clusters, but also finds the subspaces where each individual cluster exists. Allowing overlapping of clusters in the object space and in the attribute space increases the complexity of the search algorithms exponentially and also makes the interpretation of relationships among clusters very difficult. In this thesis, we propose new subspace clustering algorithms that can find overlapping subspace clusters satisfying certain quantitative and qualitative properties. These properties an be defined by the domain users so that the search focuses only on those clusters that have some significance for the users. Molding of the search to find only clusters with specific properties has the advantage that the property itself, or its derivatives, can be used to prune away the uninteresting hypotheses at an early stage of the search. Various pruning strategies are presented in the thesis for different clusters properties to make the search more efficient. In many situations, the total number of subspace clusters having the desired properties is very large, which not only adds burden to the search, but also makes the analysis on the results very difficult. In this thesis, we present ways to impose a lattice structure on all the found clusters, and we show that the lattice facilitates the discovery of other knowledge embedded in the data. We also propose another solution to this problem by creating a condensed representation of all the clusters, that is, we find only a subset of all the clusters from which all other clusters having the desired properties can be inferred. For validation of our algorithms, we tested our algorithms on both the synthetic and the real pplication data. The results suggest that the algorithms are very useful in many application domains, such as with gene expression data and some standard datasets from the machine learning repository. The emerging infrastructure of distributed databases requires algorithms to be designed for mining meaningful patterns in data located at different sites. Due to security and privacy concerns, it is not always feasible to send all datasets to a centralized site to accomplish the mining task. An alternate solution is to have each site perform some computation locally, and exchange minimum amount of information with the other sites. We focus on finding subspace clusters in horizontally partitioned databases. The global computation is decomposed into localized computations on each participating site. We present the detailed decomposition algorithm as well as the format of message exchanges between the sites. Both theoretical and empirical validation of our proposed scheme is provided, showing that our algorithm can find all target patterns from the distributed datasets. Overall, the research presented in this thesis provides many insights into theoretically and empirically characterizing the problem of subspace clustering. The subspace clustering algorithms proposed in this thesis are expected to be useful in solving data mining problems in many applications.
Advisors/Committee Members: Bhatnagar, Dr. Raj.
More Like This
15.
Billups, Robert Brent.
COMPUTER ASSISTED TREATMENT EVALUATION.
Degree: MS, Engineering : Computer Science, 2001, University of Cincinnati
► Many problems require experimental means to determine the best solutions. Experimenters must…
(more)
▼ Many problems require experimental means to determine the best solutions. Experimenters must use statistics to analyze the results of the experiments. Proper statistical evaluation of the results could be easier, with some assistance. This paper proposes Computer Assisted Treatment Evaluation, a program to assist people in the statistical evaluation of the results of experiments in the area of computer algorithm analysis. This provides a framework upon which even more extensive evaluation can be built. An experiment evaluating several sorting algorithms provides an example.
Advisors/Committee Members: Purdy, Carla.
Subjects: Computer Science
Keywords: heuristics; statistics; experimental; algorithms
More Like This
16.
BRANT, MICHAEL J.
BINDING HASH TECHNIQUE FOR XML QUERY OPTIMIZATION.
Degree: MS, Engineering : Computer Science, 2006, University of Cincinnati
► XML is a format that allows the storage and exchange of information…
(more)
▼ XML is a format that allows the storage and exchange of information across the World Wide Web. XML is a semi-structured markup language containing recursively-nested elements. Typically the volume of data in an XML file is too large to be human readable, therefore XML query processing (retrieving and combining subtrees) needs to be automated. An XML query processor has to choose an efficient method for a particular query and XML document. In this thesis we develop a method called Binding Hash (BH) that performs a subset of XML query operations. The BH Method focuses on improving performance of the most time-consuming XML query operation called structural join. A performance study indicates that the BH Method outperforms similar techniques for queries that are deeply nested. The BH Method is flexible since it can be integrated into a larger system and selected by an optimizer when it achieves the best performance.
Advisors/Committee Members: Davis, Dr. Karen J.
Keywords: XML Query Processing; XML Query Optimization; Semi-structured data; XPath
More Like This
17.
Cabarcas, Daniel.
An Implementation of Faugère's F4 Algorithm for Computing Gröbner Bases.
Degree: MS, Engineering : Computer Science, 2010, University of Cincinnati
► Gröbner Bases are an important tool for analyzing systems of polynomial equations.…
(more)
▼ Gröbner Bases are an important tool for analyzing systems of polynomial equations. They allow the system of equations to be solved exactly and therefore have gained popularity in many areas of science and technology. However, finding Gröbner Bases is a computationally intensive task, thus, several algorithms have been developed for this goal. Faugère invented an elaborate algorithm to compute Gröbner Bases in 1999 called F4, which has become a benchmark due to its efficiency. We have implemented F4 from scratch in C++. In this thesis we revisit the theoretical foundation of the algorithm, provide details of our implementation, and compare it with other software that computes Gröbner Bases.
Advisors/Committee Members: Franco, John.
Subjects: Computer science
Keywords: Groebner Bases; Implementation; Algorithm; Gröbner Bases
More Like This
18.
CALENDER, CHRISTOPHER R.
APPROXIMATE N-NEAREST NEIGHBOR CLUSTERING ON DISTRIBUTED DATABASES USING ITERATIVE REFINEMENT.
Degree: MS, Engineering : Computer Science, 2004, University of Cincinnati
► There are often times where an application needs to find the N…
(more)
▼ There are often times where an application needs to find the N nearest neighbors. Imagine the scene of an accident and someone needs to know all of the nearest fire and police departments. It may not be enough to only know one station; they may need to know the five nearest stations, or even more. That is where finding the N nearest neighbors is the perfect algorithm. Finding the N nearest neighbors is not a new concept, but doing this on distributed databases while reducing the traditional communication costs is new. There are existing ways to accomplish this by using a centralized method in which each individual database site transmits all of its data to one central machine, a learner machine, that then performs the necessary calculations. However, communication costs are very expensive, much more expensive than computational costs, and that is why there is a need for an algorithm that can perform finding the N nearest neighbors using distributed databases while minimizing communication costs. In this approach, more power is given to the individual database sites, and less to the central learning machine. What has been attempted is to allow the individual database sites to perform calculations on their data in hopes of gaining some kind of summary information that will allow a minimal amount of information to be transmitted to the learner machine rather than every point. And in fact, that is exactly what this newly proposed algorithm does. It has the individual database sites perform many calculations and then transmit only the relevant information to the learner machine. The learner machine also does computations and gathers information from all of the individual database sites. Once it has gathered some global information, it then sends out the important, relevant global information to each of the database sites. Then, in turn, each individual database site can return only the points that it feels will be relevant. This drastically reduces the amount of communication and, as one will see, works very well using real-world examples.
Advisors/Committee Members: Bhatnagar, Dr. Raj.
Subjects: Computer Science
Keywords: nearest neighbor; cluster; K-means; clustering; approximate nearest neighbor; approximate k-means
More Like This
19.
CARVER, TIMOTHY A.
A STUDY OF ADAPTIVE COMPUTER ASSISTED INSTRUCTION.
Degree: MS, Engineering : Computer Science, 2001, University of Cincinnati
► This thesis provides a standard procedure for developing Adaptive CAI lessons, and…
(more)
▼ This thesis provides a standard procedure for developing Adaptive CAI lessons, and fully utilizes the interactive features of web-based learning through Arthur. Arthur is an Adaptive Computer-Assisted-Instruction (CAI) program. Offering multiple presentation methods, Arthur is considered Adaptive CAI because it can shift from one presentation mode to another based on the learner’s demonstration of understanding the material. As Arthur is still in the experimental stages, it is not easily used by a novice or intermediate user. This presents a problem for the faculty member who is not an expert computer user. Developing CAI media is difficult enough, developing Adaptive CAI media is even more so. Without expert computer skills, a faculty member alone could not create lessons for this (and most other CAI) paradigms. The main idea of this thesis was to divide the work between a Faculty member as the expert in his/her subject, and an Instructional Designer who works with a multimedia team and the technology. This division of the work has demonstrated success with conventional CAI, and worked well in this study.
Advisors/Committee Members: Han, Dr. Chia Yung.
Subjects: Computer Science
Keywords: adaptive; computer; assisted; instruction; learning
More Like This
20.
CHAO, YUE.
Fault Detection and Diagnosis of Electro-Mechanical Systems Based on Real-time Data.
Degree: MS, Engineering : Computer Science, 2008, University of Cincinnati
► The purpose of our research is to build a computer-based system that…
(more)
▼ The purpose of our research is to build a computer-based system that utilized real-timeoperating data to detect and diagnose malfunction in electro-mechanical systems. The methodology is to use assessment chart, which uses small number of data and control chart, to evaluate the performance of mechanical system. This method can be applied to a subsystem or be extended to the entire electro-mechanical systems. In order to achieve our research goal, our system is implemented in the following steps: 1) build a model which can characterize a mechanical system, 2) review the expert experience and control strategy of such a mechanical system to establish the assessment chart based FDD system, 3) monitor the actual status of the system in real time to assure the system is operating efficiently, and 4) make decision on the collected real-time data to detect any abnormality in the performance of the target system. This approach is applied to a real building HVAC system with genuine data. The result of our experiment shows that our strategy successfully detects most of faults introduced in the building system.
Advisors/Committee Members: Han, Chia-Yung.
Subjects: Computer science
More Like This
21.
CHEN, MINGQING.
A RULE BASED EXPERT SYSTEM FOR IAQ.
Degree: MS, Engineering : Computer Science, 2005, University of Cincinnati
► Indoor air quality (IAQ) refers to the quality of the air inside…
(more)
▼ Indoor air quality (IAQ) refers to the quality of the air inside buildings as represented by concentrations of pollutants and thermal (temperature and relative humidity) conditions that affect the health, comfort, and performance of occupants. To ensure IAQ there is a need for systems that provide monitoring and diagnosing capabilities. Currently, tools for determining IAQ are based on a set of fixed rules and limited procedures only professionals can apply. Space occupants or owners usually are not able to apply the rules easily according to what they need in their spaces under given circumstances, nor are they able to adapt the existing rules forming their own rules for their own specific cases. This thesis presents a generic model for IAQ procedures. The model can be implementation in a computer-based IAQ system. The major features of this model are the capability to allow users to specify the problem, to identify key parameters for monitoring, detecting abnormalities, troubleshooting the detected problem, and suggesting actions for controlling the environment. The procedures used for each application, the customization of existing rules to specific cases, can be retained to establish new rules that will further become part of the knowledge base of a rule-based expert system framework. The implementation is demonstrated and two case studies are also presented.
Advisors/Committee Members: Han, Chia-Yung.
Subjects: Computer Science
Keywords: Indoor Air Quality; Expert System
More Like This
22.
CHEN, XINBEI.
A KNOWLEDGE MANAGEMENT MODEL FOR E-BUSINESS WORLD.
Degree: MS, Engineering : Computer Science, 2002, University of Cincinnati
► Today, more than ever, organizations have to work harder to keep up…
(more)
▼ Today, more than ever, organizations have to work harder to keep up with the pace of changes and increased global competition in the Internet-based environment, the e-business world. To succeed in multi-channel, high speed information processing environment, organizations need to leverage the knowledge they have at their disposal; they need to harness the knowledge inherent at both the individual and the corporate level; they also need to operate their business in real-time, making adjustments to changes in market conditions. In this thesis we demonstrate how knowledge management can and should contribute to leading and managing e-business-driven changes in business and operating processes, and indicate the rudiments of action agenda that decision makers will deploy to build a knowledge-management-based approach to transform their business processes. One accomplishment of the thesis is to build up a generic knowledge management framework PEDS, the Online Performance Evaluation and Decision Support System, and demonstrate it, from the system architecture and development process to business modeling. We also show that PEDS can be applied to a variety of business contexts and has a great potentiality as an enhancer to the e-business world.
Advisors/Committee Members: Han, Dr. Chia-Yung.
Subjects: Computer Science
Keywords: knowledge management; information process; component based development; decision making system; PEDS
More Like This
23.
CHEN, YUNLI.
PERFORMANCE ANALYSIS AND ENHANCEMENT OF MAC PROTOCOLS FOR IEEE 802.11 WLANS.
Degree: PhD, Engineering : Computer Science, 2004, University of Cincinnati
► An ad hoc network is a basic WLAN. In the IEEE 802.11…
(more)
▼ An ad hoc network is a basic WLAN. In the IEEE 802.11 WLANs, CSMA/CA is a widely used MAC layer protocol and considerable work has been done on the performance evaluation of this protocol. But most research is confined to saturation performance of single-hop ad hoc networks. In this work, we employ a linear feedback model to valuate the performance for CSMA/CA according to the Poisson distributed traffic in both single-hop and multi-hop ad hoc networks. To the best of our knowledge, this is the first attempt to evaluate the performance for CSMA/CA protocol analytically under a general assumption about the traffic. This work has also shown that the performance degrades dramatically in multi-hop ad hoc networks when the number of competing nodes increases, which implies that the scalability is still a major problem in ad hoc networks. The IEEE 802.11e working group proposed EDCF to support integrated data and voice (or video) communications. In the literature, there is very limited analysis about EDCF since it is a new protocol. In this work, we evaluate the performance of EDCF by dividing the traffic into two groups, namely real-time packets and non real-time packets, and use an analytical model to quantify the performance of both IFS priority and CW priority in the EDCF. Since there are multiple stations contending for one channel in the system, the queue model for EDCF cannot be a regular M/G/1 model. We redefine the service time and the waiting time so that we can still use M/G/1 model in the analysis. To validate the accuracy of our analytical results, we have done extensive simulations and we observe that EDCF does provide service differentiation between different traffic categories. But due to the inherent contention characteristic of CSMA/CA MAC protocol, the delay cannot be guaranteed, even for the highest priority traffic. A lot more work need to be done for the MAC protocol to effectively support service differentiation in the future. To enhance the performance of 802.11, we proposed optimal contention window scheme. It is observed that this protocol can greatly improve the performance.
Advisors/Committee Members: Agrawal, Dr. Dharma P.
Subjects: Computer Science
Keywords: WLAN; 802.11; CSMA/CA; performance analysis
More Like This
24.
CHHABRA, MONICA.
Modeling and Analysis of Ligand Docking to Norovirus Capsid Protein for the Computer-Aided Drug Design.
Degree: MS, Engineering : Computer Science, 2008, University of Cincinnati
► Noroviruses have been recognized as the most important cause of non-bacterial epidemic…
(more)
▼ Noroviruses have been recognized as the most important cause of non-bacterial epidemic acute gastroenteritis, affecting individuals of all ages. With the identification of trisaccharides' binding site(s) for norovirus (both VA387 and Norwalk virus strains), the door has turned open to find biologically active chemicals with better or equal binding affinity compared to trisaccharides. In this thesis as a first objective, trisaccharides binding site(s) on noroviruses were identified via computational docking and validated with experiments results. In addition to experimentally identified and computationally predicted binding site, a second stable binding site for Norwalk virus was also computationally predicted. Completion of the first goal paved the way for the second and most important aim of the research, which was to computationally identify lead candidates from a library of two million drug-like compounds that bind to the viral receptor pocket identified earlier, thus inhibiting the binding of host histo-blood group antigens (HBGA). Delivering on the second objective, a selection of 255 potential drug-like compounds were obtained successfully. Finally, an approach to score and cluster chemicals based on binding energy and binding similarity (specificity) to trisaccharides was developed and successfully implemented.
Advisors/Committee Members: Cheng, Yizong.
Subjects: Bioinformatics; Biomedical research; Computer science; Molecules; Virology
Keywords: norovirus; docking; norwalk virus; va387; nlv; virtual high throughput screening; computer aided drug design; antiviral drug; drug design; gastroenteritis; cadd
More Like This
25.
CHINTALA, VENKATA MANOJ.
HANDOFF SCHEMES FOR THE IEEE 802.11 WIRELESS LANs.
Degree: MS, Engineering : Computer Science, 2007, University of Cincinnati
► Handoff in the IEEE 802.11 wireless local area networks (WLANs) occurs whenever…
(more)
▼ Handoff in the IEEE 802.11 wireless local area networks (WLANs) occurs whenever the mobile station (MS) changes its association from one access point (AP). Previous studies have shown that the handoff schemes employed in the IEEE 802.11 networks do not meet the strict delay constraints placed by many multimedia applications. The problem is further compounded by the fact that limited coverage areas of APs employed in existing IEEE 802.11 WLANs create frequent handoffs. In this research work, four novel handoff schemes namely, FHAP (Fast Handoff by Avoiding Probe wait), APFH (Adaptive Predictive Fast Handoff), DPH (Dynamic Predictive Handoff) scheme and AP-DPH (Adaptive Preemptive Dynamic Predictive Handoff) scheme are proposed for the IEEE 802.11 WLANs to reduce handoff latency and improve the QoS of multimedia applications. The performance of all the schemes are analyzed through extensive simulations.
Advisors/Committee Members: Zeng, Dr. Qing-An.
Subjects: Computer Science
Keywords: Handoff IEEE 802.11; discovery phase; preemptive; Neighbor Graph; Improving QoS Multimedia Applications
More Like This
26.
CHOWDHURY, KAUSHIK ROY.
MULTI-CHANNEL MEDIUM ACCESS PROTOCOLS FOR WIRELESS NETWORKS.
Degree: MS, Engineering : Computer Science, 2006, University of Cincinnati
► As wireless technology becomes increasingly pervasive, there is a need for decentralized…
(more)
▼ As wireless technology becomes increasingly pervasive, there is a need for decentralized medium access control (MAC) protocols that satisfy user requirements at an acceptable cost of complexity and energ consumption. Hardware improvements have provided for multi-channel transceivers and leveraging this benefit necessitates further work in higher layer protocols. As devices get smaller leading to large networks formed of tiny sensor motes, the allocation of available channels without a central authority and their subsequent arbitration is a challenging task. In our work, we address this by first devising a channel allocation protocol (DCA) that ensures interference-free communication. We then propose a multi-channel MAC, C-MAC that enables energy efficient data transfer without tradeoffs in latency or throughput, and allows nodes to remain in their low-power sleep period for the maximum possible time. Apart from sensor networks, we have also analyzed the case for wireless local area networks (WLANs) that are the default providers of community and enterprise level internet connectivity. A judicious choice of the operational channel based on current network conditions can greatly improve performance in randomly deployed or hotspot areas. Our analytical estimation of interference allows each access point (AP) to independently arrive at the best channel, thus resulting in fewer re-transmissions due to neighboring traffic. For all the proposed schemes, a thorough performance evaluation has been undertaken and results reveal major improvement in performance for both WLAN and sensor networks, thus proving the viability of multi-channel communication models.
Advisors/Committee Members: Agrawal, Dr. Dharma P.
Keywords: Graph coloring,; Medium Access Control,; Multi-channel,; Sensor networks,; Wireless LAN.
More Like This
27.
CHUGH, SHRUTI.
AN ENERGY EFFICIENT COLLABORATIVE FRAMEWORK FOR EVENT NOTIFICATION AND DATA AGGREGATION IN WIRELESS SENSOR NETWORKS.
Degree: MS, Engineering : Computer Science, 2004, University of Cincinnati
► Wireless sensor networks consist of a large number of low power devices…
(more)
▼ Wireless sensor networks consist of a large number of low power devices equipped with RF links for communication that have numerous military, civil and environmental monitoring applications. The energy constraints due to limited battery power present several design challenges. In this thesis we propose a cluster based framework and a localized and deterministic schedule based on TDMA/FDMA MAC protocol for the nodes in a neighborhood to communicate with each other. We also present a three phase collaboration algorithm that entails exchanging messages to determine the local maximum in a region in terms of received energy from the source. This helps in minimizing the redundancy in reporting events and saves energy. We also introduce a robust data aggregation approach for queries targeting selective regions of the network. Extensive simulations show that there is a significant reduction in the overall network traffic and in the energy expended by the nodes.
Advisors/Committee Members: AGRAWAL, Dr. DHARMA P.
Subjects: Computer Science
Keywords: Wireless Sensor Networks; Collaborative Processing; Event Notification; Data Aggregation
More Like This
28.
DAHL, JORGEN L.
A COMMUNICATION LIBRARY FOR PEER-TO-PEER COMMUNICATION IN MESSAGE-DRIVEN PROGRAMS.
Degree: MS, Engineering : Computer Science, 2002, University of Cincinnati
► Distributed message-driven applications, such as distributed simulation, in tightly coupled local area…
(more)
▼ Distributed message-driven applications, such as distributed simulation, in tightly coupled local area networks often exhibit peer-to-peer communication patterns. However, high-performance communication research during the past decade have been focused on reducing the software overheads in client-server communication. Another line of research has been concentrated on achieving the performance potential of commodity systems. This thesis proposes to implement a communication library based on the MPICH, TCP,UDP, and MVICH commodity protocols. A reliable protocol for peer-to-peer communication is implemented and used in conjunction with the unreliable UDP protocol. Synthetic workloads that exhibit both peer-to-peer and client-server communication patterns are studied and performance results of using the communication library for these workloads are presented. The experiments performed in this thesis show that the implemented UDP based communication layer executing in user space outperforms the VIA based MVICH communication layer as well as MPICH and TCP for a number of synthetic workloads.
Advisors/Committee Members: Wilsey, Dr. Philip W.
Subjects: Computer Science
Keywords: peer-to-peer communication; cluster communication; reliable protocol; UDP
More Like This
29.
DARIRA, RAVI SURESH.
A DESIGN FRAMEWORK FOR PROPERTY MAPS.
Degree: MS, Engineering : Computer Science, 2006, University of Cincinnati
► Property Maps (PMaps) are bitmap index structures that improve performance by reducing…
(more)
▼ Property Maps (PMaps) are bitmap index structures that improve performance by reducing the number of extra tuples retrieved when query sets are executed. PMap design algorithms based on heuristics and possible enhancements to the design algorithms were proposed by Grommon-Litton [G00]. Gupta conducted performance study of PMaps and recommended ways to improve their efficiency [G02]. In this thesis, we study the design algorithms and enhancements and build a customized, integrated software tool to generate efficient PMaps. To make the tool easier to use, we develop additional pre-processing modules. In order to evaluate the effects of the enhancements we perform experiments by implementing both the original and the enhanced algorithms using sample PMaps. Results indicate that enhancements significantly improve query performance.
Advisors/Committee Members: Davis, Dr. Karen C.
Subjects: Computer Science
Keywords: PMap
More Like This
30.
DESAI, PRANAY A.
SEQUENCE CLASSIFICATION USING HIDDEN MARKOV MODELS.
Degree: MS, Engineering : Computer Science, 2005, University of Cincinnati
► The field of Bio-Informatics is fast growing with research in various related…
(more)
▼ The field of Bio-Informatics is fast growing with research in various related topics. One such topic is protein sequence classification. This thesis uses this topic as motivation to develop a methodology that uses Hidden Markov Models (HMMs) to classify sequences. Hidden Markov Models are a concept in probability theory widely known for their application in the speech recognition. The three phases of HMMs: training, decoding, and evaluation, are used to classify sequences into clusters that have known similar functional properties. The training phase of HMMs uses a cluster of sequences to learn a model that is most likely to generate the sequences in the training cluster. The decoding and evaluation phases of HMMs use the generated model to calculate the likelihood of an unknown sequence belonging to the same sequence and generating a most-probable path the sequence traverses. The thesis presents background on HMMs along with detailed explanations of the algorithms used to implement all three phases of HMMs. The primary focus of this thesis is on the training phase of HMMs. During the implementation of the training phase we discovered that the phase has a numerical and computational weakness relating to those structures in which some silent states are included as part of the model. The results presented in this thesis test the training algorithms, show their workability and weaknesses, and point towards the silent states related weakness.
Advisors/Committee Members: Bhatnagar, Dr. Raj K.
Keywords: HMM, Hidden Markov Models, Sequence Classification, Viterbi Algorithm, Forward Algorithm, Backward Algorithm
More Like This
[1] [2] [3] [4] [5] [6]
