Machine Learning, Its Limitations, and Solutions Over IT

Machine learning is an investigation of computer algorithms and sample data to build a mathematical standard model for making decisions without programming. Machine learning means the computer system is performing a task without being programmed for that task. We studied the machine learning concept in detail by exploring the relationship of machine learning with other fields, machine learning approaches, machine learning models, and limitations of machine learning. There are three main approaches for machine learning study supervised, unsupervised, and semi-supervised. Other than this Reinforcement machine learning, self-learning, feature learning, sparse dictionary learning, deviation detection, and robot learning. An artificial neural network, decision trees, support vector machines, regression analysis, Bayesian network, genetic algorithms, and some training models.

We found some limitations of machine learning such as bias, accuracy, ethics, and high cost of installation. To validate these limitations, we further conducted a sample survey of400 no's buyers in the Nashik city through Google form to answer two problems: Do buyers feel biased while buying on E-commerce mobile apps? And does buyer feel looted while dealing with health care issues at the hospitals? From the sample survey results, we concluded that buyers feel the bias experience while using E-Commerce apps and buyers feel looted with bad ethics experience while dealing with Health care services. We further proposed some solutions over the limitations of machine learning such as self-declaration form by E-commerce, standardisation of the proposed medical bills, and customized solutions over hardware instalment.