Master of Sciences, Case Western Reserve University, 0, EECS - Computer and Information Sciences

Link prediction is a constantly growing field, but the evaluation of newly developed algorithms requires a lot of computational resources that can be prohibitively expensive to perform on large networks. To resolve this issue, a possible approach is to reduce the computational complexity by randomly sampling the negative edges. Here, we investigate the effect of negative sampling on the evaluation of link prediction algorithms, propose models to estimate the sampling error based on the number of negative edges sampled, and suggest minimum values bounding the error to a desired amount. Across a wide-array of real networks, we show that the suggested values can appropriately bound the error and can speed up the evaluation process ~1000x times for large networks having $10^6$ nodes with minimal error. We anticipate that these results and our estimated model can help researchers keep the evaluation of link prediction methods accessible on large, real-world networks.

Committee: Mehmet Koyuturk (Advisor); Michael Lewicki (Committee Chair); Soumya Ray (Committee Member) Subjects: Computer Science

Doctor of Philosophy, University of Akron, 2008, Psychology-Industrial/Organizational

An experience sampling methodology (ESM) was utilized to demonstrate that emotional contagion is an important determinant of affect and attitude similarity within the supervisor-subordinate dyad. On a Friday afternoon, 41 manufacturing employees completed a series of trait-based measures (e.g., affect, job attitudes, emotional contagion susceptibility, emotional expressiveness, etc.), which served as control and moderator variables in the analyses. Beginning the following Monday, state-based measures were completed six times a day for two workweeks using personal digital assistants. Using Multilevel Random Coefficient Modeling, the current study demonstrates that all six state-based dependent variables (i.e., positive and negative affect (PA/NA), affective and cognitive job satisfaction, and affective and cognitive organizational commitment) exhibited significant within- and between-subjects variability. Second, state-based PA and NA predicted the state-based attitudes over and above trait-based versions of both affect and attitudes. There was also some indication that time-lagged state PA (i.e., collected during the previous measurement period) also significantly predicted the attitudes. More importantly, the current study offers evidence that emotional contagion continually operates, with moment-by-moment levels of supervisor affect and attitudes being passed down, which influence his or her subordinate's concomitant affect and attitude levels. Furthermore, a number of trait/dispositional characteristics of the subordinate, supervisor and dyad moderated the strength of this relationship.

Committee: Rosalie Hall PhD (Advisor); Paul Levy PhD (Committee Member); Aaron Schmidt PhD (Committee Member); James Diefendorff PhD (Committee Member); Steven Ash PhD (Committee Member) Subjects: Behaviorial Sciences; Management; Organization Theory; Organizational Behavior; Psychology; Statistics