Date of Award
Doctor of Philosophy (PhD)
Computer Engineering and Sciences
Philip K. Chan
This dissertation covers four data mining problems with applications in decision support based on user activity data. The first problem is an efficient approach to maximizing spread of information in social networks with applications in decision support for marketing where the goal is to find the best set of users, based on a limited budget, to maximize the word of mouth. The data for this problem is based on user activities in social networks that lead to formation of friendship (or follower-followee) graphs. The second problem is identifying action-outcome relationships to facilitate building a knowledge base of actions that could be used for decision support. The data for this problem is based on user experience about performing actions as expressed on social media. The third problem is automatic extraction of relevant product aspects in a summarized form as well as a list of pros and cons for each aspect. Identifying strengths and weaknesses of a product can be useful in the decision making process for the company that makes the product to improve the weaknesses and add desired features. We use real wold data sets based on user activities from social media to evaluate our proposed techniques. The fourth problem provides access control decision support for smartphone devices by distinguishing between device owner and others based on their typing patterns and device movements.
Ahmadzadeh, Ebad, "Data Mining Algorithms for Decision Support Based on User Activities" (2018). Theses and Dissertations. 627.