Empowering Communities Through Data Science Research
Harness data science to drive social change and improve lives.
Machine Learning: Optimizing Participant & Mentor Matching
Abstract
This research project explores how machine learning can be used to predict personality traits to better understand participants in Aceing Autism. Understanding participants' personality traits allows Aceing Autism to tailor their programs and support strategies to each individual's unique needs. This can help Aceing Autism match mentors with like-minded students and address gaps in their personality profiles, leading to more effective and personalized learning experiences. A supervised machine learning approach was used, where personality survey data, taken from the Big Five Personality traits data set, was preprocessed and used to train classification models that label traits such as Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism as either high or low. The model achieved an accuracy of approximately 99 percent, showing that it can effectively identify personality patterns from structured data. Performance results based on personality score matches from the project suggest that machine learning can be a useful tool for analyzing personality traits and may help organizations like Aceing Autism better support their participants.
Results
With 1 being the highest, I can take any student and coach and test their personility
Conclusion
The results of this project demonstrate that machine learning models can be highly effective in predicting Big Five personality traits from structured survey data. After training and evaluating multiple classification models, the final model achieved an overall accuracy of approximately 99 percent across the five personality traits. This high level of accuracy indicates that the model was able to successfully learn patterns between survey responses and personality outcomes.
When comparing different models, simpler models such as logistic regression performed well due to the structured nature of the dataset and the relatively clear relationships between features and target variables. However, more complex models such as decision trees and random forests were able to capture non-linear relationships and interactions between features, which further improved prediction performance. The use of ensemble methods, particularly random forests, helped reduce overfitting and increased the stability of predictions across different subsets of the data.
Support for Aceing Autism Organization
This past summer I had the privilege of getting a paid internship for the first time in my life and I truly believe I should give some of the money back to the community especially for this good cause. Therefore I have started working on a fundraiser for the Aceing Autism organization.
Little bit about myself
President of South Asian Culture Club at Polytechnic
President of South Asian Culture Club at Polytechnic
President of South Asian Culture Club at Polytechnic
This is my third year running this club. We host events to spread and expose school to South Asian Culture. We celebrate holidays like Diwali and play cricket
Parliamentary Debate Captain
President of South Asian Culture Club at Polytechnic
President of South Asian Culture Club at Polytechnic
This is my first year heading Parliamentary debate team.
Social Media Manager for Pickleball club
President of South Asian Culture Club at Polytechnic
I have had the privilege of running Pickleball club at my school. I am currently working on starting pickleball for autistics kids at my school
Costco Internship
I created a Chatbot for Costco Australia that helps answer questions asked by logistics companies. This chatbot uses Chatgpt model that I trained to answer logistics related questions
3D Printing club
We are working on printing connectors for Geodome project for school
Coming soon: Pickleball club for Autistics Kids
I am currently working on starting a Pickleball club for Autistic Kids at Polytechnic