Gradient Boost Decision Tree Classifier Based Big Data Analysis for Twitter Sentiment Analysis Using Adaptive Neural Network

Social media in nowadays has a lot of data to improve, leading to legacy organizations as a tag for big data and failing to manage this data. Large volume data management is now an issue. Big data growth is huge. Due to various properties, it is very difficult to manage. This manuscript focuses on analysing big data in a Twitter environment using Hadoop. We categorize big data according by the characteristics huge data, value, diversity and speed. This action involves tokenization, removal of blocked, ignored word, and numerous deletions. Beginning Word Words HDFS (Hadoop Distributed File System) Passes to reduce repetitive words and delete using graph and reduction techniques. After that, emoticons and nonemoticons features are obtained. The resulting properties rank their subject. Next, the classification is done using DLMNN (Deep Learning Modified Neural Network) which is based on Improved Elephant herd optimization for Weight optimization. Finally, the result achieved was verified using the analysis method cross-validation method. The final contribution is made with the help of an efficient Sentiment Analysis (SA) and Twitter Data Awareness Classification a Gradient Boost Decision Tree Classifier (GBDT). Finally, GBDT classifies the data has highperformance to categorize the classes negative, positive, or neutral. In addition, we propose a very effective impression result based on Emotional analysis like this has proven to be very effective and accurate.