16A Perspective Analysis of Regularization and Optimization Techniques in Machine Learning

Ajeet K. Jain1*, PVRD Prasad Rao2 and K. Venkatesh Sharma3

1Department of Computer Science and Engineering, Koneru Lakshmaiah Education Foundation, Vaddeswaram, AP, India (Association: CSE, KMIT, Hyderabad, India)

2CSE, CVR College of Engg., Hyderabad, India

3CSE, KLEF, Vaddeswaram, AP, India


In the present days of big data analytics, the magnitude of data generated is escalating at an exponential rate and requires high-end machines with many-core CPUs (GPUs) to handle. To deal with such scenarios, conventional optimization and regularization methods have been practiced to solve for optimum solutions. The structural information contents in big data vary in their representations, and the algorithms have to deal with their sparsity of underlying datasets. Thus, in engineering science applications, the methodologies play a significant role in finding out extremums with constraints. However, due to their paramount size, the traditional algorithms have to deal with contemporarily non-convex nature of data and have to covenant with manifold parameters. Consequently, the needs for dealing with larger datasets is equally parallel growing and have opened up variety of new techniques as well as delving into innovative research directions. In turn, this necessitates us to look further at various methodologies of optimizations and regularization algorithms focused therein from machine learning ...

Get Computational Analysis and Deep Learning for Medical Care now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.