We noticed that there is a lack of accessibility and affordability for services to help people improve their state of mind or address their mental health issues. We wanted to create a platform that helps people navigate their mental health concerns and helps people understand themselves better through self-exploration.
What it does
MindFlight is s chat bot that uses natural language processing to look for cues and phrases in order to identify mental health issues or obstacles in the user's path to self-exploration.
How we built it
We combined the power of python along with the chatbotAI package, which created an efficient interface to create a bot using a natural language processor. We then wrote a script that integrated the research we had done with the chatbotAI package's system.
Challenges we ran into
A challenge that we encountered was the lack of documentation for the chatbotAI package. It uses an HTML template syntax combined with RegEx queries as an API for the chatbotAI. This syntax doesn't have a lot of documentation for it, which made it difficult to do things like extracting information from the user's dialogue.
Accomplishments that we're proud of
In the end, we were able to successfully integrate our research with the chatbotAI package despite the lack of clear documentation, through trial and error, as well as our problem-solving skills.
What's next for MindFlight
In the future, MindFlight will include a sentiment analysis network along with data retention which allows it to form an LSTM network, that is able to change and decide its direction based on the data it acquires in real-time. This will lower the barrier to access, and upgrade it from a Chomsky Type 3 grammar processor, to a Chomsky Type 1 grammar processor.