Financial Asset Valuation Using Online Learning Techniques in Semi-Streaming Data

Online learning is a process toward an- swering an arrangement of inquiries when the knowledge of the geuinine result is given. Online learning majorly constitutes the algorithms Weighted Ma- jority as well as Randomized Weighted Majority. These are popularly called the mistake bound models, which means that the algorithms can specify the upper bound of its number of mistakes made in the prediction. This project introduces Financial asset valuation using online learning techniques combined with streaming data input, by means of the Apache Spark and Python collaboration (PySpark), Resilient Distributed Datasets (RDD) and Web scraping for processing the semi streaming data and Anaconda, Python3 in Jupyter Notebook to process the batch database. Our aim is to considerably reduce the regret bound for prediction of the values as well as to introduce a novel approach to streaming dataapplied on online-learning methods.