The following project is a AI chatbot for people struggling with stress, depression and anxiety. Simple ML model was trained in the project on a selfmade dataset from scratch on a simple json file , the model after a lot of parameter tuneing was set to preform 500 epocs with a batch size of 22 and a RNN model was used with 3 layers of 320, 200 and 194 nodes respectively, the model acheived an accuracy of 96% with a loss of only 8 to 10% Simple custometkinter gui was made just for generaic GUI representation, with dark mode as well. Python speech recognition algorithms were also used for an intractive user machine intraction.
The dataset for this project is quit limited at the moment and can slove only upto 5% of the given problems, to make this project better in the near future, many approaches are being studied at the moment to enhance and make this project an actual success.
