In this day and age “Investment” has become a necessity and an important factor for companies and individuals. Investment is something which is called a monetary asset purchased with the idea that the investment will provide a profit in the future. Before investing people study Financial performance, Background and experience in the industry, Company uniqueness, Effective business model, Large market size of a particular company, but they do not focus on minute fingerprints of financial distress. The main of this research is to draw down the factors of investment under a single umbrella and generate and investment advice.
This study will mainly focus on developing two models to refine the investment process. The first model we have proposed is a stock market prediction based on deep learning techniques. Here we have used a Realtime dataset from Yahoo finance for a particular company where clients want to invest. Here we have used different deep learning techniques. But for this research sequential LSTM has outperformed all the models with minimum rmse (root mean squared error) score. With the help of this are able to predict the stock closing price of the company for up to one month. The second model will be of Financial Distress Predication based on various Bagging and Boosting techniques with the integration of various SMOTE techniques were used. But for this research Balanced Bagging Method with ADASYN has outperformed from all the models with an accuracy of 93%. ADASYN Adaptive Synthetic Sampling Method is a modified version of SMOTE which performed best with our bagging and boosting model, we have used ADASYN to deal with the class imbalance problem. This empirical research is carried out based on real world financial data of 3476 Chinees company with over 84 financial and non-financial features.
Keywords: Investment, Financial Distress Prediction, Stock Market Prediction, Deep learning, SMOTE, ADASYN, Bagging and Boosting, Ensemble Methods, Adaboost-SVM, LSTM.