The main aim of the project is to build an intelligent model capable of analyzing personal chat data to identify patterns that indicate user preferences, what they like and dislike. The algorithm is trained on datasets consisting of text conversations between two individuals, learning from the structure, sentiment, and topics of the chats. By the end of the training, the model can summarize user interests and aversions with high accuracy. This type of system has major applications in personalized marketing, recommendation engines, and user behavior analytics. The outcome is a functional tool that helps businesses and e-commerce platforms understand their users better, enabling them to offer more relevant services and products.
Over a twelve-week development cycle, students will engage in a sequence of tasks starting from foundational concepts to full implementation. The project begins with an introduction to natural language processing and machine learning fundamentals, particularly Naive Bayes and SVM algorithms. Students will create or curate a suitable dataset of simulated or anonymized chat conversations for model training. Next, they will install and configure machine learning libraries and tools using Python and platforms like Anaconda Navigator or Google Colab.
In the middle stages of the project, students will build a model that can preprocess text, extract relevant features, and classify segments of the conversation as indicating a “like” or “dislike.” The model will then be tested on unseen datasets to evaluate its accuracy and adjusted as needed to improve performance. Toward the end of the project, the team will prepare documentation, finalize testing, and deliver a presentation showcasing their system. Proper coding practices, ethical handling of sensitive data, and strict adherence to privacy protocols must be maintained throughout the project.
