Machine Learning Based Automatic Pattern Analysis for Banking Data With Improved Feature Selection
Although in recent era of modernization, each and every affairs is dependent on technology, we are compelled to accept the economy as the key to all the affairs. Above all, the economy has only played a role in the implementation and expansion of all the technologies including computational problems. In order to implement any technology, investment of money is necessary. So, how this economy can be strengthened has become a major issue of the country. So, existing technologies, especially computational technologies, can be used to strengthen this economy as a result we can achieve unprecedented success. Even, Computational Economy has become a demandable subject in the whole world which is enhancing day by day. Some computational problems have been used in this study from Machine Learning. In this study Principal Component Analysis is used to decrease the feature vector. On the other hand Artificial Neural Network, Support Vector Machine and Recurrent Neural Network have been used to forecast the import amount of Bangladesh. Eventually, K-fold cross validation is used to separate the training and testing data set for the predictive models.
Keywords - Financial Informatics, Economic Informatics, Data Mining, Imports and Exports, Machine Learning, Artificial Neural Network, Support Vector Machine, Recurrent Neural Network, Computational Economics.