Chat Nova

• 

  It the interface of our Chat bot

Inspiration

We sought to build a straightforward yet powerful AI chatbot which should be able to communicate with users in a substantial manner. Drawing inspiration from virtual assistants, we built Chat Nova to answer standard questions, help users budget, and credit scores, and general user interactions.

What it does

Chat Nova is a Python-based text chatbot developed with Streamlit and machine learning. It is able to: 1 . Welcome users and answer general questions. 2 . Give budgeting advice and credit score advice. 3 . Provide friendly and courteous conversational experiences. 4 . Identify various intents with TF-IDF vectorization and Logistic Regression.

How we built it

1 . Utilized NLTK for tokenization and processing of text. 2 . Used TF-IDF Vectorization to extract features from input text. 3 . Trained a Logistic Regression model to classify user intents. 4 . Created an interactive Streamlit UI for easy chatbot interaction.

Challenges we ran into

1. Training the model to accurately translate user input to the correct intent.
2. Optimizing responses so that it could avoid repetition and randomness.
3. Managing our users input differences successfully.

Accomplishments that we're proud of

1. Built a working chatbot in a specified time limit.
2. Integrated machine learning for intelligent selection and response.
3. Built an intuitive user interface using Streamlit.

What we learned

1. Enhanced our knowledge of Natural Language Processing (NLP).
2. Developed hands-on experience with ML models for text classification.
3. Learned how to combine AI with web applications using Streamlit.

What's next for Chat Nova

1. Widening the chatbot's knowledge base with increasingly diverse subjects.
2. Using real-time APIs (such as for weather information).
3. Improving the model using deep learning-based NLP for improved response accuracy.

Built With

• python
• streamlit