Expert finding in an enterprise social network.

The study addresses expert finding problem - identifying some professional community participants with a relatively higher level of certain knowledge or skills. This study considers the search for experts in an enterprise social network - the company's intranet portal, introduced to speed up communications and improve its integrity; build a convenient knowledge base; increase manageability, speed of execution and transformation processes. Since the value of a corporate social network increases for its users with the growth of the system, the success of such systems depends directly on the user involvement degree, i.e. regularity, intensity and diversity of the system use. In this study, expert finding problem is discussed in the context of user involvement increase. The search for experts is based on the data analysis on user actions in the corporate social network of an oil and gas production company. Experts are characterized by their activity in various modules of the system, regularity of activity and the level of communication network involvement. Expert search and typing are carried out using clustering algorithms (hierarchical and k-means) based on the users’ characteristics in each month. As a result, an expert path was obtained for each user - a sequence of transitions from cluster to cluster over the all user activity periods. Possible areas of the expert finding results applications include: 1) searching for the most active experts, asking them to involve new users, jointly developing requirements for the functionality improvement; 2) creation of smart tips or recommendations focused on offering functionality interesting for a user expert group; 3) searching for the most requested functionality, from which users most often begin to get acquainted with the system, or problem functionality users have difficulty with.