Stock Price Prediction Project: Viva Questions

1. General Questions about the Project

1. What is the aim of your project?

o The aim is to predict stock prices using historical data and machine learning. The
project provides visualizations, real-time analysis, and AI-driven predictions to
assist users in making informed investment decisions.

2. What societal benefits does your project offer?

o It promotes financial literacy, bridges the technology gap by introducing AI tools,

and reduces misinformation by providing data-driven predictions.

3. Why did you choose this project topic?

o The project demonstrates the practical application of AI in finance, addressing

real-world challenges like stock market analysis and promoting technological
inclusivity.

4. Who are the potential users of your project?

o Potential users include students, beginner investors, finance enthusiasts, and

educators.

5. What challenges did you face during this project?

o Challenges included managing data preprocessing, ensuring model accuracy, and

integrating the predictive model with the user interface.

2. Technical Questions about the Code

6. Which libraries have you used, and why?

o Libraries include:

▪ pandas and numpy: For data manipulation and numerical computations.

▪ matplotlib: For data visualization.

▪ keras: To load the pre-trained machine learning model.

▪ yfinance: For retrieving historical stock price data.

 ▪ [Link]: To normalize the data using MinMaxScaler.

▪ streamlit: To build an interactive web app.

7. What role does the MinMaxScaler play in your project?

o It normalizes the data to a range of [0, 1], which ensures the model trains
efficiently and predictions are scaled back accurately.

8. How is the data split for training and testing?

o The data is divided into 85% for training and 15% for testing. Additionally, the last
100 days of the training data are appended to the test data for prediction
continuity.

9. Why did you use a rolling average in your visualization?

o Moving averages (100-day and 200-day) are used to smooth fluctuations in stock
prices, making it easier to identify trends over time.

10. Explain how you prepare the test data for predictions.

o The last 100 days of training data are appended to the test data. The combined
dataset is normalized, and a sliding window approach is used to create input-
output pairs for the model.

11. What does the model's output represent?

o The model predicts the normalized closing prices, which are then scaled back to
the original range using the inverse of the scaling factor.

3. Machine Learning-Specific Questions

12. What kind of machine learning model is used in your project?

o The code uses a pre-trained Long Short-Term Memory (LSTM) model, which is
effective for time series predictions like stock prices.

13. What are the key evaluation metrics for your model?

o Metrics like Mean Squared Error (MSE) and Root Mean Squared Error (RMSE) are
used to evaluate the accuracy of the predictions.

14. Why did you choose LSTM over other algorithms?

 o LSTMs are specifically designed to handle sequential data and can capture
temporal dependencies, making them ideal for time series data like stock prices.

15. How does the sliding window method work in this project?

o It creates overlapping sub-sequences of the dataset to generate input-output

pairs for training and testing the model.

16. What are the limitations of using LSTM for stock price prediction?

o LSTMs require significant computational resources, can overfit on small datasets,

and may struggle with highly volatile or unpredictable data.

4. Questions about Streamlit

17. How does Streamlit enhance your project?

o Streamlit allows users to interact with the model via a web interface, making it
easy to input stock tickers, view visualizations, and compare predicted vs. actual
prices.

18. What features are implemented in the Streamlit app?

o Features include:

▪ Selection of stock ticker.

▪ Visualization of historical prices and moving averages.

▪ Display of model predictions alongside actual values.

19. Can users interact with the app in real-time?

o Yes, users can input stock tickers and view real-time predictions and
visualizations.

20. What challenges did you face while integrating Streamlit with your project?

o Challenges included ensuring the app's responsiveness, managing data inputs

dynamically, and integrating visualizations effectively.

5. Questions on Data and Analysis

21. What challenges did you face with the stock data?
 o Handling missing data, ensuring the data is normalized, and managing the
temporal nature of stock price predictions.

22. Why did you use Yahoo Finance for data retrieval?

o Yahoo Finance provides easy access to historical stock data through its API, which
integrates seamlessly with the yfinance library.

23. What types of patterns were you looking for in the stock data?

o Patterns such as trends, seasonality, and volatility to help the model make
accurate predictions.

24. How do you handle missing or inconsistent data?

o Missing data can be interpolated or dropped depending on the extent of the

inconsistency.

6. Debugging and Testing Questions

25. How did you test the accuracy of your model?

o By comparing predicted prices with actual prices using visualizations and

calculating error metrics like RMSE.

26. What would you improve in your project?

o Enhancements could include adding real-time predictions, sentiment analysis

from news or social media, and improving model accuracy with hyperparameter
tuning.

27. What issues did you face during model training?

o Issues included overfitting, slow training times, and sensitivity to

hyperparameters.

28. How did you debug errors in your code?

o By breaking down the code into smaller units, testing each component
individually, and using debugging tools like logs and exceptions.

7. Advanced Questions

29. What is the role of feature engineering in your project?

 o Feature engineering involves creating meaningful features, such as moving
averages and normalized prices, to improve model predictions.

30. How can hyperparameter tuning improve the model?

o Tuning parameters like learning rate, batch size, and number of LSTM layers can
optimize model performance and reduce overfitting.

31. What are alternative approaches to predicting stock prices?

o Alternatives include ARIMA models, Random Forests, Gradient Boosting, and

hybrid models combining LSTM with attention mechanisms.

32. How would you integrate sentiment analysis into your project?

o By analyzing news articles and social media data using natural language
processing (NLP) to gauge market sentiment and combine it with stock price
predictions.

33. What ethical considerations are involved in using AI for financial predictions?

o Ethical considerations include ensuring transparency, avoiding misuse of

predictions for market manipulation, and preventing over-reliance on AI-driven
advice.

34. How do you ensure your model generalizes well to new data?

o By using techniques like cross-validation, regularization, and testing the model on

unseen data.

35. What deployment challenges might you face?

o Challenges include scaling the app for multiple users, ensuring low latency for
real-time predictions, and managing data privacy and security.

36. What is the future scope of your project?

o Future scope includes adding support for more complex models, integrating
macroeconomic indicators, and deploying the app for public use.