Elevating Education

Inspiration

Education is the first step in change, yet many students struggle silently. I was inspired by the desire to bridge the gap between potential and performance. Elevating Education was made to provide institutions and teachers with the tools to identify learning obstacles within students and successfully produce personalized recommendations.

What it does

1. Analyze learning patterns using student data
2. Predicts potential learning difficulties using ML and LSTM time-series models
3. Provides personalized solutions based on rule-based logic
4. Visualized trends and progress over time through numerous different plots and graphs
5. Offers confidence scores to support education decision-making
6. Predicts future test scores using past data ## How I built it
7. Created mock student performance data simulating the real world.
8. Trained a Random Forest Classifier with SMOTE and undersampling techniques to handle the imbalanced classes.
9. Built an LSTM neural network to predict future scores.
10. Implemented a logic-based system and flagged learning issues early. ## Challenges I ran into I had to create meaningful mock data to simulate real students' data. I also had to balance overfitting in the ML model due to a class imbalance in the student dataset. ## Accomplishments that I am proud of I combined rules-based logic, ML, and deep learning into a single predictive system. I am also proud of using different libraries to visualize long-term learning trends and make successful predictions with confidence scores while also being ethical and educator-friendly. ## What I learned I learned how to preprocess, balance, and model student data effectively, and how to use an LSTM network for time-series prediction. ## What's next for Elevating Education Integrate real-time data from LMS platforms, create a user-friendly frontend app, and add chatbot support to help students get automated suggestions and teacher explanations.

Built With

• matplotlib
• numpy
• pandas
• python
• scikit-learn
• seaborn
• smote
• tensorflow