As people get older, simple everyday life becomes more difficult. So watching my parents get older and seeing my grandparents have a more challenging time, it became clear to me that foundations like the Leland Home are essential to have. The Leland Home offers excellent opportunities for seniors looking for another home. However, going on the website, it isn't easy to navigate when thinking of a client over the age of 62. Implementing a simple chatbot to help the user navigate different site sections and give more information to the user is the fastest solution to the problem. Using mainly python and PyTorch, we are able to use deep learning methods to help parse, analyze, and understand the users' questions to the bot. Using a feet forward neural network as the model's base, we can then use a method to read in the user input through tokenization, stemming, and bag of words. These three processes will split up the sentence, cut the endings of words, and count words that matter to the model from our JSON file. Finally, we can train the model using PyTorch to create training data from the JSON file. Time was most valuable. Understanding the short stretch of the competition, I was not able to fully finish the project. I would have hoped to design the bot more intricately to give the user a step-by-step guide on anything someone navigating the site would need to know. Also, implementing some type of front-end would be an excellent completion of the project. Using Visual Studio code, I implemented most of the project. In addition, I learned more about python and its usefulness when implementing a machine learning model.
A Friendly Navigator
The Leland Home is a non-profit cause with a simple message to help people. However, the site is hard to navigate for their target audience. Implementing a simple bot to help the user is the solution!
Updates
Leave feedback in the comments!
Log in or sign up for Devpost to join the conversation.