Incomplete data, characterised by missing, corrupted, or unobserved values within a dataset, presents a pervasive challenge across machine learning domains. It typically arises from hardware sensor failures, non-responses in surveys, alignment mismatches in multi-modal systems, or data collection dropouts. Left unaddressed, missing values break conventional mathematical operations (such as matrix multiplications in fully connected layers), leading to biased parameter estimation, reduced statistical power, and severe degradation in downstream prediction or classification accuracy.
|
|
|
|
Dr Suvra Jyoti Choudhury earned his B.Tech in Computer Science and Engineering from the West Bengal University of Technology, India, followed by his M.Tech in Information Technology from the Indian Institute of Engineering Science and Technology, Shibpur, India, in 2009 and 2014, respectively. He completed his PhD in Computer Science from the Indian Statistical Institute, Kolkata, in 2022. Currently, he is an Assistant Professor at the Indian Institute of Information Technology, Nagpur. His research interests include image processing, artificial neural networks, and machine learning.
|
|
|
Admin office: 196, T.T.K. Road, Alwarpet, Chennai - 600018
Campus: 5655, Central ExpressWay, Sri City, Andhra Pradesh - 517464
www.krea.edu.in
|
|
|
|
|
