Machine Learning Based Automatic Pattern Analysis for Banking Data with Improved Feature Selection

Abstract

A very famous adage of Adam Smith, “All money is a matter of belief”. It is, of course, the beginning of the first use of money was observed when the supply of demanded product was available in the hands of others. It can also be said that the introduction of money has introduced us to a system called business. Initially, this system gave rise to the internal economy but later it spread to the whole world. But in the whole world there is a new system emerged for the flow of economy which is known as bank to everyone. And through this bank, a country may be importing or exporting every day. Only the economy of a country is considered good when export is more than import. Due to everyone's attention of better economy, export can be increased and import can be optimized. So, to solve this problem statistics can help us greatly. As Statistics, has been using in determining the existing position of per capita income, unemployment, population growth rate, housing, schooling medical facilities and so on. In this study not only statistics but also machine learning tools were used to analyze and forecast the financial banking data specifically import data. Basically, import data is known to a country as an economic data. So when we can predict about imports, then deciding how much of the export will be good for economics can easily be determined. In this study we have worked with import data of Bangladesh Bank for analysis and forecast the import of Bangladesh, based on collected data to strengthen the economic condition.