Distinguished & Emeritus Professor, Iwate Prefectural University, Japan.

Biography:

Basabi Chakraborty received B.Tech, M.Tech and Ph. D degrees in Radio Physics and Electronics from Calcutta University, India and worked in Indian Statistical Institute, Calcutta, India until 1990. From 1991 to 1993 she worked as a part time researcher in Advanced Intelligent Communication Systems Laboratory in Sendai, Japan. She received another Ph.D in Information Science from Tohoku University, Japan in 1996 and worked as a postdoctoral research fellow until 1998. She joined as a faculty in the department of Software and Information Science, Iwate Prefectural University, Japan in 1998 and served as Professor and Head of Pattern Recognition and Machine Learning laboratory until her retirement in March, 2022. Currently she is a distinguished Professor and Professor Emeritus in Iwate Prefectural University. She also worked as Dean and Distinguished Professor in School of Computing, Madanapalle Institute of Technology and Science, A.P, India for three years from June 2022 to June 2025. Her main research interests are in the area of Pattern Recognition, Machine Learning, Soft Computing Techniques, Biometrics, Data Mining and Social Media Data Mining. She is a senior life member of IEEE, member of ACM, an active member of IEEE WIE affinity group and worked as a chair of IEEE JC WIE in the year of 2010-2011. She was the founder chair of Sendai section WIE affinity group in 2017-2018. Currently she is secretary of Sendai LMAG (Life member affinity Group).

Shapley based Approach for Optimum Feature Subset Selection

     Optimal feature subset selection is an important prerequisite for any classification or regression task involving high dimensional data to decrease computational cost as well as to increase performance of the classification model. To achieve this, feature selection algorithms select the discriminatory and relevant features and discard redundant and irrelevant features by evaluating individual features or the feature subsets using some suitable metric. Lots of feature evaluation measures have been developed till now that can be classified into three main categories, wrapper or model dependent, filter or model independent and hybrid. Shapley value, a concept from cooperative game theory, has recently been used in various machine learning applications. It is emerging as a powerful evaluation tool in Explainable AI (Artificial Intelligence) for assessing the contribution of an individual feature to the performance of the model. In this lecture, research works on Shapley value-based feature subset selection algorithms are reviewed. The role and performance of Shapley value in finding the solution of optimal feature subset from a large feature set has been critically examined using simulation experiments with more than 100 bench mark data sets from different application areas. Merits and demerits of Shapley based feature selection approach are summarized and future directions of research has been elucidated.