Our goal is to use posts from this mental health forum to identify people with similar backgrounds/concerns so they can identify with each other. We thought that this way, people undergoing depression/mental issues who simply need attention can receive attention from individuals who understand what the person is going through. We hope that this will eventually lead to support groups more tailored to what the a certain person is needing.
We divided the project into two main groups: A group that can gather the required data to analyze and another group focusing on analyzing the data in order to figure out how to group people together.
Our group is gathering the data from mental health forums. Our main focus is to gather posts and the usernames of those who wrote each post. Other useful information that we could collect includes dates, likes/hugs, and mentions within posts. Since forums are so large, we will be using a mix of Selenium and Beautiful Soup to move through pages and collect the relavent data.
Our group is looking at unsupervised machine learning as a primary way to group people together. By using an algorithm such as KMeans, we think that we may be able to appropriately group specific people together. Our main objective of this project was to learn Natural Language Processing as well, and working with text data allows us to do so. We aim to be able to extract key words from the posts that might lead to a stronger link between two people (such as words relating to relationship problems or specific situations) and to be able to subtract slightly less meaningful parts of the post in order to cut down on computation requirements.
- Areeba Ahmed (aahmed70)
- Danielle Curammeng (dc18)
- Shaw Kagawa (skagawa2)
- Lanxin Liu (lanxinl2)