Cloud-Based Machine Learning and Sentiment Analysis

Google Drive Link

This thesis document contains three aspects of data science projects aimed at improving tools and techniques used in analyzing and evaluating data. The first research study involved the use of a standard cybersecurity dataset and cloud-based auto-machine learning algorithms were applied to detect vulnerabilities in the network traffic data. The performance of the algorithms was measured and compared using standard evaluation metrics. The second research study involved the use of text mining social media, specifically Reddit. We mined up to 100,000 comments in multiple subreddits and tested for hate speech via a custom designed version of the Python Vader sentiment analysis package. Our work integrated standard sentiment analysis with Hatebase.org and we demonstrate our new method can better detect hate speech in social media. Following the sentiment analysis and hate speech detection, in the third research project, we applied statistical techniques in evaluating the significant difference in text analytics, specifically the sentiment-categories for both lexicon-based software and cloud-based tools. We compared the three big cloud providers, AWS, Azure, and GCP with the standard python Vader sentiment analysis library. We utilized statistical analysis to determine a significant difference between the cloud platforms utilized as well as Vader and demonstrated that each platform is unique in its analysis scoring mechanism.