Analysis of Global Finance Using Web Scraping and Topic Modeling
This paper aims to exploit these databases thereby giving out a processed knowledge base and forming the outcome of topics and subtopics related to Global Finance by providing a modeled output in the form of visits, clicks, likes, dislikes and more.
Abstract
At present there is a humongous amount of knowledge that is present all around us for every field and it may be used to derive predictions to get better outcomes and profit. Organizations can use this data to improve their user experience and learn from user trends. This paper aims to exploit these databases thereby giving out a processed knowledge base and forming the outcome of topics and subtopics related to Global Finance. It focuses on providing a modeled output in the form of visits, clicks, likes, dislikes and more, these parameters can help organizations to grow immensely. Sentiment analysis is one of the branches of NLP, which is, in lay man's terms is a text mining operation to find the positive and negative sentiment of the data source, so if we could apply this to financial news, social media feed or use opinion mining for detecting the opinion of the masses on social media this will act as an excellent tool for individual investors to analyze market in no time. In addition, a machine learning model is developed to predict the stock prices. The model is based on support vector regression. The major challenge in this scheme was to detect the taxonomy into which the data is to be categorized. With IAB Taxonomy, a list of subtopics and topics was formulated to arrange the content in groups. We will use raw structured Data of URLs as an input, the Proposed Model will use this input to power on the program and produce an output by dividing the website data into topics and subtopic. In addition, various other parameters like visits, likes, dislikes and score will also be displayed. The analysis and implementation of the model presented in our study was completed using modeling and scraping methods keeping the kind of data we have in mind.