STOCK MARKET TREND PREDICTION USING KNN

Abstract

Research on stock price prediction is both intriguing and difficult. A developed nation's power 
economy is the yardstick by which its economy is judged. The stock market is now a prestigious 
arena for traders to participate in due to the fact that it often yields high returns with little risk. 
The stock market is a great place for data miners and business researchers to work because of the 
abundance of ever-changing information it contains. Our goal in this research was to help users, 
managers, and investors make better investment choices by predicting stock prices using a 
company's stock data using a non-linear regression technique and the k-nearest neighbour 
algorithm. In order to train the module, this method takes a stock's daily high, low, open, and 
close values as well as its volume. The next step in testing is to get the user's initial stock value 
and use it as a test variable in the module. The module's output will be the stock's expected 
closing price. A visualisation graph may be used to analyse the discrepancies between the actual 
and anticipated stock closing prices. The findings were fair and logical since the kNN technique 
is robust with a minimal error ratio. Furthermore, the forecast results were near to, if not exactly 
parallel to, real stock values, depending on the data we used.