Search

We consider the team formation problem in open collaborative projects existing in large community setting such as the Open Source Software (OSS) community. Given a query specifying a set of required skills for an open project and an upper bound of team size, the goal is to find a team that maximizes the Degree of Acquaintance (DoA) and covers all the required skills in the query. We define the DoA in terms of the team graph connectivity and edge weights, corresponding to the local Clustering Coefficient for each team member and the strength of social ties between the team members, respectively. We perform a statistical analysis on historical data to show the importance of the connectivity and social tie strength to the overall productivity of the teams in open projects. We show that the problem defined is NP-hard and present three algorithms, namely, PSTA, STA and NFA, to solve the problem. We experiment the algorithms on a dataset from the OSS community. The results show the effectiveness of the proposed algorithms to find a well acquainted teams satisfying a given query.

In recent years, data uncertainty widely attracts researchers' attention because the amount of imprecise data is growing rapidly. Although data are not known exactly, probability distributions or expected errors are sometimes available. While most researchers on uncertain data mining are looking for methods to extract mining results from uncertain data, which is usually in the form of probability distributions or expected errors, it is also very important to lower the data uncertainty by making a part of data more certain to help get better mining results. For example, input values of some sensors in the sensor network are usually designed to be recorded more frequently than others because they are more important or more likely to change. In this paper, the issue of selecting a part of uncertain data and acquiring their exact values to improve clustering results is explored. Under a general uncertainty model, we propose both global and localized data selection methods, which can be used together with any existing uncertain clustering algorithm. Experimental results show that the quality of clustering improves after the selective exact value acquisition is applied.

In this paper, we investigate and utilize the characteristic of the group movement of objects to achieve energy conservation in the inherently resource-constrained wireless object tracking sensor network (OTSN). We propose a novel mining algorithm that consists of a global mining and a local mining to leverage the group moving pattern. We use the VMM model together with Probabilistic Suffix Tree (PST) in learning the moving patterns, as well as Highly Connected Component (HCS) that is a clustering algorithm based on graph connectivity for moving pattern clustering in our mining algorithm. Based on the mined out group relationship and the group moving patterns, a hierarchically prediction-based query algorithm and a group data aggregation algorithm are proposed. Our experiment results show that the energy consumption in terms of the communication cost for our system is better than that of the conventional query/update based OTSN, especially in the case that on-tracking objects have the group moving characteristics.