Book Image

Big Data Analytics with Java

Book Image

Big Data Analytics with Java


Overview of this book

This book covers case studies such as sentiment analysis on a tweet dataset, recommendations on a movielens dataset, customer segmentation on an ecommerce dataset, and graph analysis on actual flights dataset. This book is an end-to-end guide to implement analytics on big data with Java. Java is the de facto language for major big data environments, including Hadoop. This book will teach you how to perform analytics on big data with production-friendly Java. This book basically divided into two sections. The first part is an introduction that will help the readers get acquainted with big data environments, whereas the second part will contain a hardcore discussion on all the concepts in analytics on big data. It will take you from data analysis and data visualization to the core concepts and advantages of machine learning, real-life usage of regression and classification using Naïve Bayes, a deep discussion on the concepts of clustering,and a review of simple neural networks on big data using deepLearning4j or plain Java Spark code. This book is a must-have book for Java developers who want to start learning big data analytics and want to use it in the real world.
Table of Contents (21 chapters)
Big Data Analytics with Java
About the Author
About the Reviewers
Customer Feedback
Free Chapter
Big Data Analytics with Java
Ensembling on Big Data
Real-Time Analytics on Big Data

Logistic regression

This is a popular classification algorithm where the dependent variable (outcome) is categorical. Even though it has the word regression in its name, it is a classification technique. Using this technique, we can train a model on some training data and the same model we can later use on new data to classify it into different categories. So, if you want to classify data into categories such as 1/0, Yes/No, True/False, Has Disease/No Disease, Sick/Not Sick and so on, logistic regression is a good classifier model to try in these cases. As per these examples, logistic regression is typically used for binary classification, but it can also be used for multiclass classification too.

The approach used by this algorithm is quite simple. We apply the data from the dataset onto a mathematical optimization function and this function will later make the data fall either in a 0 category or 1 category. Later on when we get a new piece of data we apply the same function to that new...