Prediction of BSE Stock Data using MapReduce K-Mean Cluster Algorithm

Abstract

Bombay Stock Exchange (BSE) Limited, established in 1875 as the Native Share and Stock Brokers' Association is considered to be one of Asia’s fastest stock exchanges and oldest stock exchange in the South Asia region. On 31 August 1957, the BSE became the first stock exchange to be recognized by the Indian Government under the Securities Contracts Regulation Act 1956. In this paper, we developed a novel framework that can achieve parallel time series prediction using Hadoop. By implementing the proposed framework, the system should be able to deal with massive amount of time series data, either regular or irregular. The proposed system can handle the optimization, parameter selection and also model combination through K-mean clustering. In this paper, experiment is carried to forecast the company’s next bid accurately based on the other companies that have similar trend with it.