Project Goals
Analysie the all the stock market that is NSE (National Stock Exchange) in India.
This project aims to analysis the all the stock market that is NSE (National Stock Exchange) in India. The National stock exchange has the characteristics that all the stock markets have in common is the imprecision and uncertainty that are related with their long term and short term future state. These characteristics are used to creates the risky and volatile to invest and difficult to determine the appropriate the suitable stock for an investor. It also makes the maximum profit on their investments. The stock market analysis is used to trying to forecast the stock market trend because the input and output data which is forecasted the relationship between them is find out using analysis and training. The analysis is divided into two types like technical analysis and fundamental analysis. The prediction of future prices based on the past market data is done using technical analysis. The fundamental analysis is used to involve the company analysis, economic analysis and industry analysis. The stock value and stock market is affected by the various factors like bank exchange rate, movement of stock markets, political events, commodity price index, psychology of investors, general economic conditions and firm’s policies etc. Here, we are using the neutral network to predict the financial markets by using the both analysis.
This project goal is to replace the existing method to predict the stock market data because the existing method use the statistical parameter value of tth day and using this method the closing price of the day is find out . But, it does not provide the effective solution for prediction problem. So, this project is helpful in increasing the prediction by using the older values of statistical parameters in various methods like Support Vector Regression (SVR), Artificial Neural Network (ANN), Support Vector Machine (SVM) , Random Forest and Naïve-bayes method. This project also overcomes the above stated problem and it needs to successfully predict the stock index value for next month or week. It also needs to improve the prediction efficiency.
According to this paper (Patel et al., 2015) describes the problem of Indian stock markets in predicting the direction of movement of the stock price index and stock price. It easily compares the four models of prediction such as Random Forest, Naive Bayes, Artificial Neural Network(ANN) and Support Vector Machine (SVM). During stocking ten technical parameters are computed using input data. It focuses on the representation of technical data calculated from input data in the deterministic form or it shows a specific trend. With the help of these technical parameters calculated using input data increases the accuracy of each prediction models. It easily evaluates the input approaches to predict the stock data. In this paper technical parameters are calculated for ten years which is from 2003 to 2013 of two stocks exchanges. Overall performance of random forest is better than remaining three prediction models.All the prediction models performance are displayed and it improves the technical parameters of the trend deterministic data. Due to various uncertaininties the prediction of stock prices index and stock is very difficult. So, investors before investing their money in a stock they uses two types of effective analysis approaches. The two type of analyses are fundamental analysis and technical analysis. These are providing the effective performance on improving the prediction of data on India stock markets.
Related Work
This paper (Patel et al., 2015) is used to describe the mixture of machine learning techniques that help in predicting stock market index prices. In this paper mainly focus is on the prediction of future prices of stock market index. This paper uses the two Indian stock market indices like S&P and CNX Nifty stock exchange sensex. These indices are used for evaluating experimentally. This indices’ contains the 10 years of historical data and it used to predict the data on Indian stock markets. This paper propose two station of fusion approaches by using the SVR-SVR fusion prediction models because this models performs the single stage scenarios where SVR, Artificial neural network and Random forest. These models are used to improve the prediction of data and these are selected by technical indicators or investors. The values of stock indices are calulated by technical parameters which uses the stocks with hight market capitalization value. The statistical information is derieved using technical parameters from the value of stock prices. The prices of stock are of high market capatilization. After the results it was concluded that performance of two state hybrid prediction model is better than single state prediction models. The performance is significantly improved when RF and ANN are hybridized with SVR.
The data set contains Indian eequity market for NSE (National Stock Exchange) of India’s benchmark stock market index (Kara, Acar Boyacioglu and Baykan, 2011). For 22 sectors of economy it has well diversified 50 stock index which is used for differentpurposes like index absed derivatives, index funds and portfolios for bench-marking.The data contains the 8 variables: index, date, time, open, high, low, close and id. It also contains the historical data from 2013 to 2016; the number of trading data of each minute of has given each date. The currency of the price is Indian Rupee (INR).
- index : market id
- date: numerical value
- time: factor
- open: numeric (opening price)
- high: numeric (high price)
- low: numeric (low price)
- close: numeric (closing price)
The data in data set for each year increases and decreases. It helps in finding the design parameters of prediction models by taking the data of equal proportion for each of the four years. For the better representation of all the data sets in parameter setting, it uses sampling method .It uses the two technical indicators to moving the averages and it is simple technical analysis tool. It is used to predicting the short term future. The indicators values are used to trend the deterministic input set and it is used to predict the data and provide the improved performance of all the models (Chong, Han and Park, 2017).
Data
This project replaces the existing method to predict the stock market data because the existing method uses the statistical parameter value and it does not provide the effective solution for prediction problem (Rai and Seeru, 2017). So, we are using the below method to provide the improved performance for predictions models:-
- Artificial neural network (ANN)
- Random forest
- Naïve – bayes
- Support vector machine (SVM)
Here, we are using the prediction models are discussed in below.
The artificial neural network is basically demonstrated their capability in prediction and financial model. The ANN model helps in predicting the movement of stock price index. Generally, it contains the input layer, output layer and the hidden layer. These layers are connected to each other. In ANN model, at each layer at least one neuron must be employed in artificial neural network. In ANN model, the technical indicators act as input of the artificial neural network model. The output of the ANN model contains two type of patterns for stock price direction (STOCK MARKET PREDICTION USING MACHINE LEARNING, 2017). The ANN is the dense network of interconnected neurons and it gets activated depending on the inputs. The ANN parameters are momentum constant, number of epochs, number of hidden layer neurons and value of learning rate. The values of these parameters are tested in parameter setting experiment. The value of learning rate for the ANN model varies between 0.1 and 0.9 (Masoud, 2014).
The support vector machine is categorised into two types- Support vector classification and Support vector regression. The SVM model uses the high dimensional feature space. It is a learning system. The SVM model contains the points that are assigned to one of the two disjoint half spaces. It help in identifying the maximum margin of hyperplane. The separation of margin is used to minimize the separation between negative and positive examples on the vector machines. The maximum margin hyper plane find by SVM model is a final decision boundary. (Sentiment Classification using Machine Learning Techniques, 2016). SVM model helps in mapping the high dimensional feature spaces and input vectors. It uses the kernel function to determine the predicted data efficiently.
The random forest is the most popular technique for prediction model. It uses decision tree learning for classification. It is very efficient and accurate method as compared to other classification methods. The algorithm used by this model is categorized as ensemble learning algorithm.It uses the decision tree as the base of ensemble learning. Ensemble learning algorithm is a single classifier and it helps in determining the class of test data. It uses sampling method with replacement in which according to given data set the n trees are learnt. The random forest avoids the problem of overfitting.
Methodology
The naive bayes classifier is used to assume the class conditional independence. In this method prediction of probability of the data that belongs to a particular class is done. The bayes theorem concept is used in predicting the probability of data as it helps in calculating the posterior probability. In this method the test data is classified with the highest probability. In the class conditional independence method the values of attributes of one class is independent of the other class. It uses the training set to calculate the prediction of the data. It serves the theoretical justification for other classifiers.
The prediction models are used in calculating the design parameters of predictor. The design parameters are calculated by taking equal samples of data from each of the four years. For better representation of data of overall stock market it uses the sampling method that help in parameters setting the data. It is used to predicting the short term future. The indicators values are used to trend the deterministic input set and it is used to predict the data and provide the improved performance of all the models. Our prediction model should follow particular pattern for prediction of trend for a particular year if that year is a part of cycle day a bullying one.The values of stocks for particular year can be affected by external affairs like economy of the country . They are not isolated , but they create a trend. In this method it send the result containing the good performances by all models. For improved performances of all models it uses continuous valued input (Twala, 2011). Here, we are using the following prediction models like RF, ANN and SVM. These approaches are predicts data for 30 days ahead of the time are carried out. Different combination of model parameter are carried out by each prediction model. But in this method the best parameter combination is used which give minimum prediction error The visual representation tells how effective is this proposed approach for the prediction task. By seeing visual representation the reason for the improved performance may be justified. It is successfully identified the transformation from the technical parameters and it describing the day to day closing price. The results of Nifty are shown below (Zhang, Johnson and Wang, 2012).
Conclusion
This project is successfully analyzed the all the stock market that is NSE (National Stock Exchange) in India. The National stock exchange has the characteristics that all the stock markets have in common is the imprecision and uncertainty that are related with their long term and short term future state. The stock market analysis is used to trying to forecast the stock market trend because the forecasting is used to find the relationship between the output and input data through the analysis and training. This project goal is to replace the existing method to predict the stock market data because the existing method use the statistical parameter value of tth day and to predict the closing price of the day, the output of this method is used as input. But, it does not provide the effective solution for prediction problem. So, this project is helpful in increasing the prediction based by using older values of statistical parameters in various methods like random forest, SVR (support vector regression), ANN (Artificial neural network ), SVM( Support vector machine) and naïve-bayes method. It shows the all prediction models performance and it improves the technical parameters of the trend deterministic data.
References
Chong, E., Han, C. and Park, F. (2017). Deep learning networks for stock market analysis and prediction: Methodology, data representations, and case studies. Expert Systems with Applications, 83, pp.187-205.
Kara, Y., Acar Boyacioglu, M. and Baykan, Ö. (2011). Predicting direction of stock price index movement using artificial neural networks and support vector machines: The sample of the Istanbul Stock Exchange. Expert Systems with Applications, 38(5), pp.5311-5319.
Masoud, N. (2014). Predicting Direction of Stock Prices Index Movement Using Artificial Neural Networks: The Case of Libyan Financial Market. British Journal of Economics, Management & Trade, 4(4), pp.597-619.
Patel, J., Shah, S., Thakkar, P. and Kotecha, K. (2015). Predicting stock market index using fusion of machine learning techniques. Expert Systems with Applications, 42(4), pp.2162-2172.
Patel, J., Shah, S., Thakkar, P. and Kotecha, K. (2015). Predicting stock and stock price index movement using Trend Deterministic Data Preparation and machine learning techniques. Expert Systems with Applications, 42(1), pp.259-268.
Rai, J. and Seeru, F. (2017). Stock market forecasting using machine learning with market sentiment. Global Sci-Tech, 9(4), p.218.
Sentiment Classification using Machine Learning Techniques. (2016). International Journal of Science and Research (IJSR), 5(4), pp.819-821.
STOCK MARKET PREDICTION USING MACHINE LEARNING. (2017). International Journal of Advance Engineering and Research Development, 04(5).
Twala, B. (2011). Predicting Software Faults in Large Space Systems using Machine Learning Techniques. Defence Science Journal, 61(4), pp.306-316.
Zhang, M., Johnson, G. and Wang, J. (2012). Predicting Takeover Success Using Machine Learning Techniques. Journal of Business & Economics Research (JBER), 10(10), p.547.
Order an Essay Now & Get These Features For Free:
Turnitin Report
Formatting
Title Page
Citation
Outline
