AI Forecast

A beginner's guide to Stocks and Investment; uses Artificial Intelligence to predict the stock market.

Comment

Inspiration

Harsimran: I've built a passion for the stock market and the cryptocurrency market just after I came around this world a year ago. Since then, I've thought about a way to make investing available for more people and a way to predict the future of these markets. Recently, I've been diving deeper and deeper into the field of machine learning and wondered if AI could help predict trends that humans can't.

Ridhima: Stocks and investments are concepts that many adults have heard about, but never taught about. As students, we are suddenly expected to know how to finance. This website was created for students; dimple and casual, this website contains one needs to educate themselves about the stock market and even make their first investment using our Artificial Intelligence: Stock Predictor. As a fellow researcher, I love writing academic, and informational papers and this website displays my skills: researching and explaining, helpfully and practically.

What it does

The AI Forecast allows the user to select a stock from the list they want to analyze. Then, it downloads the required data for the stock from Yahoo Finance and displays it on a table that is categorized by date, closing price, and various other values. Beneath that, there is a graph that displays the actual closing prices for the stock, along with what the LSTM model predicts the closing prices will be. Underneath that there is a graph of the predicted closing price values for 'n' years into the future of the selected stock. Linked to it is an informational site where one can read the articles to gain enough knowledge to maneuver on the AI Forecast.

How we built it

Harsimran: I used TensorFlow and Keras to create a 6-layer LSTM neural network. It was trained on stock closing prices from the year 2015 till today, downloaded from Yahoo Finance for 100 epochs. The front end is developed with Streamit. The graphs are shown thanks to Matplotlib.

Ridhima: I used pure HTML scripting (.htm file) to develop the website which was then linked to the Forecast via an API, as a courtesy to my teammate. I used CSS templates from the W3Schools website, which I then adapted to fit our UI needs. Furthermore, I used databases such as Google Scholar, and Investopedia to write the information which is cited in MLA.

Challenges we ran into

Harsimran: This is one of my first projects utilizing TensorFlow and Keras to create a model so it was a little bit difficult in terms of ML.

Ridhima: One of the main challenges I ran into was being a complete beginner regarding all aspects of this project. I have never worked with pure coding as I have always been assisted by game-developing UIs (id est, Unity, Unreal Engine). At first, I decided to make the website using NextJS and its addon Chakra UI. Having never created a website from scratch, or using the softwares listed above, I was stuck on every aspect of web design. Six hours into working on the project late at night, I decided to abandon the .js files and I coded the website via .htm files.

Accomplishments that we're proud of

Harsimran: I'm proud of the accuracy of my first LSTM model given the amount of time I had to create it.

Ridhima: Considering that I have never coded a website from scratch, I am very proud of the functional UI I have created as the face of our project. When I scrapped the NextJS and Chakra UI just a few hours from the deadline, I was doubting whether I would be able to complete my part of the project however, it all pulled together and we were able to combine our sub-projects into the AI Forecast.

What we learned

Harsimran: I also learned how to utilize Streamlit to create a UI on a local site and better how ML models work.

Ridhima: As funny as it sounds, I finally learned to use VS Code; as a C++ programmer, I downloaded the wrong extensions, and the compiler seized to work so I abandoned it until today, and I am proud to say that I understand the software now. Furthermore, I learned the formatting of HTML and how to host server sites, and many other things that were covered in the workshops provided by the Vivid Hack Team.

What's next for AI Forecast

Harsimran: I will increase the accuracy of my model and make changes to the UI. Make it will increase the accuracy of my model and make CH available also for cryptocurrencies and once I complete these things among other small changes, I will make Forecast publicly available in the form of a website.

Ridhima: More educational content on my behalf; I would love to start a research database exclusive to the AI Forecast. Furthermore, I would like to implement interactive elements to make the website more fun and engaging for our audience. Perhaps, it will follow a Quizlet, or Khan Academy format, where one can learn the information and test themselves on it.

Built With

Share this project:

Updates