Project components

1. Development of an intent-based chatbot:

• A feed-forward neural network was used to train the intent-classification model

• Utilizing NLTK techniques such as tokenization, stemming and bag of words to pre-process the training data and as well as user-inputs.

Technique	Description	Usage
Tokenization	Process of breaking text into words or phrases (tokens).	allowing the chatbot to understand and process individual components of user input.
Stemming	Reducing words to their base or root form.	Improves text analysis by grouping similar words.
Bag of Words	Represents text as an unordered set of words.	transforming text into numerical representations that serve as feature vectors for training models to recognize user intents.

• Before training the FNN model, it is required to prepare the training data in a json format.

• The intents.json is a file used to train intent-based chatbots. It defines user intents with names and associated training examples, teaching the chatbot how to recognize and respond to different user inputs.

• This file is crucial for natural language understanding, contributing to the chatbot's accuracy and effectiveness in understanding user intent.

Below is a snippet example of the training data:



2. Integrating the Chatbot with APIs:

• The chatbot was integrated with 2 external APIs (themoviedb ,openweather)

• themoviedb API allows the chatbot to extract movie details.

• openweather API allows the chatbot to extract weather details.

• Additionally, I also integrated the chatbot with thewikipedialibrary to parse data wikipedia.