DSpace JSPUI
DSpace preserves and enables easy and open access to all types of digital content including text, images, moving images, mpegs and data sets
Learn More
Please use this identifier to cite or link to this item:
http://dspace.dtu.ac.in:8080/jspui/handle/repository/18743| Title: | COMPARATIVE STUDY OF MACHINE LEARNING ALGORITHMS FOR STOCK PRICE PREDICTION |
| Authors: | BASHEER E, FEBIN RAJ, GINSHU |
| Keywords: | STOCK PRICE MACHINE LEARNING MACHINE ALGORITHMS STOCK PREDICTION |
| Issue Date: | 31-May-2020 |
| Abstract: | The objective of the project is to do a comparative analysis of the machine learning algorithm to predict the price of a stock which is listed on the Indian stock markets which is the NSE and the BSE. We will be looking in to some of the machine learning algorithms that can be used to predict the stock price. The comparative study of these algorithms is to know which algorithm gives the best accurate result. In this report we will be talking about number of machine learning techniques that can be used to predict the stock price of the selected stock. Regression model, ARIMA, Long-short term memory, SVM, Random forest are some of the methods we use in the analysis and prediction of the stock price. A comparative study of these algorithms is also drawn out to best understand the accuracy of each of the methods. We will be using the data of sample stock price with which we will be implementing the machine learning algorithms to predict the stock value. We will be giving the detailed analysis of algorithm that we are using to predict the stock market price. Moreover, we will be using plots and charts to identify the trends in the stock price. Each method we use is distinct in each of the aspects regarding to the machine learning algorithms. There are number of algorithms used for the predictions, we will be focusing on certain algorithms which is best suited and gives the more accurate results from the data. We will also compare different algorithm used for prediction. We will be looking carefully in to the results that gives a comparative analysis of each of the method and a detailed inference on each of the algorithm and its results. Since each of the algorithms are different in their own ways a clear-cut interpretation and results will provide a good insight in to each of the algorithms also their limitation. The successful prediction of the stock price will be a great asset for the stock market institutions and this will also provide a solution to the real-life scenarios that the stock market investors will face |
| Description: | Prediction of the stock price is one of the most difficult scenarios we have in hand. There are many factors that should be incorporated and taken into account while predicting the stock price. These factors include physical, psychological, rational, irrational behavior etc. All this factor makes the stock price prediction more volatile and very difficult to predict with a high degree of accuracy. The main objective or the motivation behind this comparative analysis is to find out a clear-cut analysis of each of the algorithm which can be further implemented in the prediction of the stock price. This will in turn perform as a preliminary step to incorporate they dynamic changes in the market and also put it in to the algorithm which then gives a real time value. The stock market act as a third party who is responsible for successful completion a trade between the buyer and seller. They act as the intermediary to avoid default and cheating in the market, so the exchange will select the shares of trusted firms. Stock which are also called as shares generally represents ownership of a company to retail and institutional investors trading in the stock market. In this work we are trying to predict the future price of a stock and this prediction is expected to be robust, accurate and efficient. The proposed system is studied and planned such that the working is according to the real-life scenarios and should be well suited to the real-world settings. The system or the algorithm which we are using is expected to take all the technical factors and variables that might affect the price of a stock and its performance from the historical data. It has been found that number of these techniques have been previously studied up on in the prediction of the stock price. There are various methods used for the predicting the future value of the stock price and various way of implementing prediction system like Fundamental analysis, Technical analysis, Machine learning, Market mimicry, time series analysis etc. with the 9 advancement of technology and the upcoming of the digital era the prediction has moved up to another promising level. The most prominent and promising technologies include the application of Artificial Neural Network, Recurrent Neural Network, which actually is the implementation of machine learning algorithms. This technique of machine learning also involves the usage of artificial intelligence which have a great deal of impact in this field which in-turn teaches the system to learn and improve from the learning experiences without being programmed in time to time. There are also some traditional methods used in the prediction of stock price using machine learning algorithm such as the Backward propagation, which is also known as the Backward Propagation Errors. We will also be taking in to consideration of machine learning algorithms such as Linear Regression model, Arima model, The Prophet method, Long short-term memory etc. Researchers are now using many techniques in predicting the future stock price, these consist of many ensemble learning techniques. The highlight of these techniques is that they are very less time consuming and low price in predicting the future value price. Stock market prediction for short time window seems to be a random process. The stock price movement over a long time will usually develop a linear curve. The general tendency of the people is to buy the stock as the price may go high or rise in the near future. The main problem with people not investing in stock market is due to the uncertainty associated with the market value. These refrain people from investing in the stock. So there should be a proper mechanism that will help the people in predicting the price in real life scenarios. The methods used to predict the future price of the stock includes some time-series analysis using regression and other models. Also, some technical analysis and some machine learning algorithms are used for the price prediction. The dataset of the stock market prediction model includes details like the Opening price, High price, Low price, Closing price(OHLC), volume data etc… that are needed for the prediction of the price of the stock. We are also incorporating a classification technique that will also be a first hand in predicting the stock values. The aim is to design a model that gains from the market information utilizing machine learning strategies and gauge the future patterns in stock value development |
| URI: | http://dspace.dtu.ac.in:8080/jspui/handle/repository/18743 |
| Appears in Collections: | MBA |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 2K18-MBA-908 2K18-MBA-909.pdf | Prediction of the stock price is one of the most difficult scenarios we have in hand. There are many factors that should be incorporated and taken into account while predicting the stock price. These factors include physical, psychological, rational, irrational behavior etc. All this factor makes the stock price prediction more volatile and very difficult to predict with a high degree of accuracy. The main objective or the motivation behind this comparative analysis is to find out a clear-cut analysis of each of the algorithm which can be further implemented in the prediction of the stock price. This will in turn perform as a preliminary step to incorporate they dynamic changes in the market and also put it in to the algorithm which then gives a real time value. The stock market act as a third party who is responsible for successful completion a trade between the buyer and seller. They act as the intermediary to avoid default and cheating in the market, so the exchange will select the shares of trusted firms. Stock which are also called as shares generally represents ownership of a company to retail and institutional investors trading in the stock market. In this work we are trying to predict the future price of a stock and this prediction is expected to be robust, accurate and efficient. The proposed system is studied and planned such that the working is according to the real-life scenarios and should be well suited to the real-world settings. The system or the algorithm which we are using is expected to take all the technical factors and variables that might affect the price of a stock and its performance from the historical data. It has been found that number of these techniques have been previously studied up on in the prediction of the stock price. There are various methods used for the predicting the future value of the stock price and various way of implementing prediction system like Fundamental analysis, Technical analysis, Machine learning, Market mimicry, time series analysis etc. with the 9 advancement of technology and the upcoming of the digital era the prediction has moved up to another promising level. The most prominent and promising technologies include the application of Artificial Neural Network, Recurrent Neural Network, which actually is the implementation of machine learning algorithms. This technique of machine learning also involves the usage of artificial intelligence which have a great deal of impact in this field which in-turn teaches the system to learn and improve from the learning experiences without being programmed in time to time. There are also some traditional methods used in the prediction of stock price using machine learning algorithm such as the Backward propagation, which is also known as the Backward Propagation Errors. We will also be taking in to consideration of machine learning algorithms such as Linear Regression model, Arima model, The Prophet method, Long short-term memory etc. Researchers are now using many techniques in predicting the future stock price, these consist of many ensemble learning techniques. The highlight of these techniques is that they are very less time consuming and low price in predicting the future value price. Stock market prediction for short time window seems to be a random process. The stock price movement over a long time will usually develop a linear curve. The general tendency of the people is to buy the stock as the price may go high or rise in the near future. The main problem with people not investing in stock market is due to the uncertainty associated with the market value. These refrain people from investing in the stock. So there should be a proper mechanism that will help the people in predicting the price in real life scenarios. The methods used to predict the future price of the stock includes some time-series analysis using regression and other models. Also, some technical analysis and some machine learning algorithms are used for the price prediction. The dataset of the stock market prediction model includes details like the Opening price, High price, Low price, Closing price(OHLC), volume data etc… that are needed for the prediction of the price of the stock. We are also incorporating a classification technique that will also be a first hand in predicting the stock values. The aim is to design a model that gains from the market information utilizing machine learning strategies and gauge the future patterns in stock value development | 1.46 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.