In recent years, identifying experts has gained significant attention in the research area. The main motivation behind it is to facilitate the process of locating the correct individual capable of answering our queries. There has been a lot of focus on building expert recommendation systems. The main focus of these systems is to effectively build an expert profile in order to facilitate recognition. We argue that definition of an expert is a very subjective term and it has a major dependency on the individual initiating the search. There has also been a lot of research on personalizing search results. The two main methods applied in the design of these techniques are (1) Using explicit feedback (ratings etc.) (2) Using implicit feedback (mouse movements etc.). We propose TAK, a learning-based framework for accurate retrieval of experts based on tacit knowledge of the user placing the request. We focus on defining the tacit knowledge of the user based on implicit features like experience and education to deduce the preference of the user and generate more specific and targeted suggestions. The increasing usage of social media for everyday communication has made it a suitable repository of user specific information. Thus, we base our study on LinkedIn, which is a social media application pervasively being used for exchanging information and locating qualified individuals. We use crowd preference knowledge to create a learning-based framework and augment the result with the expert profile created from LinkedIn to provide expert recommendations to the user. This enables the user to make an informed decision. A comparative analysis of the results of the proposed method to the method applied by LinkedIn proves that the former provides more popular suggestions to the latter. It further proves that cultivated tacit knowledge with years of experience has an impact on expert selection decision.
