STOCK MARKET TREND PREDICTION USING HIGH-ORDER INFORMATION OF TIME SERIES

ABSTRACT:-

Stock market attracts thousands of investors’ hearts from all around the world. Forecasting the stock exchange data is an important financial subject which involves an assumption that the fundamental information which is publicly available from the past has some predictive relationships to the future stock returns. Stock market is one of the important things in any country’s economic point of view.   People need to understand the way stock changes and how a stock behaves in the next period of time. Predictions of Stocks are interesting not only for the trading community only but also for a computer enthusiastic public. When we think about prediction, it can happen in two ways: we can predict using previous data values and the other way is to look and understand the news and data in the Digital Media. In the previous case there is a problem with the unavailability of the data or some data which is available but we might get inaccurate predictions because of changing patterns. Our system will predict the stock prices for the next trading day and for the specific date and we use moving average technique to get a better prediction from the model. The Moving Average is the premier popular procedure among every such marker. The point of this investigation is to appear whether the moving normal pointer is absolutely worth the case by financial backers and examiners. The point of this instrument is to supply expectations and know if the future security is worth developing. Sentiment analysis is contextual mining of text which identifies and extracts subjective information in source material. Real-time sentiment analysis about the stock prices will become useful to the user to know what the opinion is about the company.  We have carried out sentiment analysis of tweets so that users can easily identify the opinion of people whether tweets are positive, neutral or negative by using graphs. The accuracy of the prediction is evaluated and gives a percentage of accurate result. Combining the accuracy and the prediction can be given to the user to acknowledge them the trend of the target stock with known accuracy. Also, users can check the future price of each company.

OBJECTIVES OF THE PROJECT:-

The objectives of the systems are:

•       Explore and Research about stock market prediction,
•     To predict stock price using the LSTM and ARIMA(Autoregressive Integrated Moving Average) model.
•       To build a dashboard for stock analysis.
•       To successfully scrap real-time stock prices from yahoo finance.
•       To display the stock price time-series graph of a given company.
•       To display future growth of the stock price time-series graph of a given company.
•       To display multiple company stock price time-series graphs for comparison.
•       To predict the stock price for the next trading day or the specific date.
•       To evaluate the accuracy of prediction and give a percentage of accuracy result.
•       To fetch the latest news related to the particular company.
•       To fetch the latest tweets related to the particular company.
•       To do sentiment analysis of the tweets so that users can easily identify the opinion of people whether tweets are positive, neutral, or negative by using graphs.
•       The accuracy of the prediction is evaluated and gives a percentage of accurate results. For that, we use model.predict_proba, which is a TensorFlow syntax to get the accuracy of the answer.

EXISTING SYSTEM:-

Prediction of stock prices is a very challenging and complicated process because price movement just behaves like a random walk and time varies. In recent years various researchers have used intelligent methods and techniques in the stock market for trading decisions. Here, we present a brief review of some of the significant researchers. prediction of the stock market using SVM. It is a good idea to use SVM as it always gives unique results and works well even at local minima, whether a stock trend can be predicted using the anomalies of the historical financial data or not. Sometimes data is insufficient because of that, accuracy is less.

PROPOSED SYSTEM:-

In our proposed system, Prediction of Stocks are very much interesting not only for the trading community but also for the computer enthusiastic public also. When we think about prediction, it can happen in two ways: prediction can happen because of previous data values and the other way is to look and understand the news and data in the digital media developing a model to predict the prices of the stock on

certain parameters using machine learning algorithms. Designing a web application which includes a list of companies for

which we will be predicting the price of the stock for a given date. Also, we did sentiment analysis of tweets so that users can easily identify the opinion of people whether tweets are positive, neutral or negative by using graphs.

MODULES:

• Fetching Raw Dataset: Day wise past stock price selected companies are collecting from the yahoo finance.
• Raw Data Pre-proccessing: Creating a clean data structure with 60 timestamp and 1 output which is fed to the neural network.
• Generating LSTM model: By adding data with LSTM which consists of a sequential input layer initiated using random weights and biases will predict the future values.
• Visualization: Using JavaScript prediction is visualized.
• Stock News and Tweets: By using twitter API and beautiful soap fetching stock related news and tweets
• Sentiment Tweets: Generating Tf-Idf vectors based on NLP to classify tweets into positive, negative and neutral.

ADVANTAGES  OF PROPOSED  SYSTEM:-

1. The accuracy of the prediction is evaluated and gives a percentage of accurate result.
2. System will predict the stock prices for the next trading day and for the specific date
3. Also, we did sentiment analysis of tweets so that users can easily identify the opinion of people whether tweets are positive, neutral or negative by using

HARDWARE AND SOFTWARE REQUIREMENTS

HARDWARE:-

1. Processor: Intel Core i3 or more.
2. RAM: 4GB or more.
3. Hard disk: 250 GB or more.

SOFTWARE:

1. Operating System : Windows 10, 7, 8.
2. Python
3. Anaconda
4. Spyder, Jupyter notebook, Flask.
5. MYSQL