Inspiration

The stock market is an excellent platform to earn and invest money. It is also a risky option that increases your greed and leads to drastic decisions. This is majorly due to the volatile nature of the market. There is no proper prediction model for stock prices.

What it does

You need to enter the ticker of the company/stock. For example, Google - 'GOOGL', Tesla - 'TSLA', Reliance Industries - 'RELIANCE'. Once the ticker is entered, a dataset is created automatically containing the High price, Low price, Open price, Close price, Volume and Adj Close. Opening price of the past 10 years is visualized in the form of a graph. My source of information was yahoo, but it's an individual choice. You can even upload a csv file containing data instead of using online resources.

Then with the available data a LSTM neural network is created and trained. Then provide the date for the model to predict the price.

Finally a plot with predicted values is displayed.

How I built it

I build it using python. I used machine learning architecture of LSTM while also making use of prominent Python Libraries such as Tensorflow, Keras, numpy, Pandas, etc.

Challenges I ran into

This is a very complex task and has uncertainties.

Accomplishments that I'm proud of

Most of the times, predicted patterns are close to actual patterns.

What I learned

some machine-learning algorithms, converting data-frames to array and vice-versa, Architecture of LSTM.

What's next for Stock-Price-Prediction

Adding GUI.

+ 42 more
Share this project:

Updates