Text and Image Classification using Conventional Machine Learning to Convolutional Neural Networks

G. Keerthi , C. Ravi Kishore Reddy , M. Sudhakar , K.Sushma

Image Processing is a broad area of science and innovation used in horticulture, pattern recognition, face or animal recognition and in many other fields. Conventional machine learning algorithms mostly applied on the categorical data and can be optimized with numerous parameters to minimize the cost function or certain loss. However, in the recent years the advancement of deep learning technologies used in many applications includes image processing. For example, Google Cloud Vision API enables several AI models to understand the content present in the images in a simple way. Many researchers across numerous disciplines have been integrating machine/deep learning in to their research to take care of issues that could not have been solved before. In this paper first, we took to give a careful investigation of machine learning algorithms to well-known convolutional neural networks (CNNs). Second, we have taken two different data sets from the Kaggle, one for categorical/text classification and other for animal classification. In addition to this, we experimented and evaluated the popular machine learning algorithms on these datasets and compared the results.

Volume 11 | Issue 7

Pages: 515-526