Neural Fuzzy Inference Hybrid System with Support Vector Machine for Identification of False Singling in Stock Market Prediction for Profit Estimation
Bhupinder Singh and Santosh Kumar Henge
The Stock market prediction is one of the biggest challenges in the global market. The volatility in the movement of stock prices deteriorates the interest of the investor and trader. The main reason of weakness of direction prediction accuracy of trader is due to buy or selling stock based on false signals that always result in loss of capital. The false singling-based perdition is the biggest problems in stock market prediction. Identification of false signals in the stock market prediction will remove some sort of noise with the implication of intelligent system based algorithms, which are used to build for solving the specific problem in specific domain and it not extendable for solving some specific uncertainties in same specific problem. The machine learning algorithmic-based neural networks, support vector machines and decision trees techniques will more helpful for detecting future stock values based on historical data and concurrent data. The blended technology of neural fuzzy inference hybrid system is deriving more flexible solutions for predicting the stock market values. This research has identified in-depth gaps in techniques that had not been explored earlier by previous researchers and it is proposing the blended technologies of neural fuzzy inference hybrid system along with support vector machines to reduce complexities in stock market prediction. This research is more helpful for the stock traders whose depends on intelligent trading system that help them to take efficient buy or sell decisions based on specific conditions.
Keywords: Stock predictability (SP), neural fuzzy inference hybrid system (NFIHS), support vector machine (SVM), cross-validation (CV), profit estimation (PE), time frames (TF), machine learning (ML), tuning parameters (TP)