Research Project: Investigating Risk Factors Associated with Self-Harm and Harm to Others using Machine Learning
This GitHub repository contains the research project that aims to investigate the risk factors associated with self-harm and harm to others using various machine learning algorithms. The study utilizes Multilayer Perceptron Regression, Random Forest, Decision Tree, and Linear Regression algorithms to predict self-harm and harm to others based on data collected through a survey conducted on Google Forms.
The main objective of this research project is to gain insights into the predictive performance of different machine learning algorithms, interpret the models' results, and establish statistical associations between risk factors and self-harm/harm to others.
The results of this study suggest that machine learning algorithms can effectively predict self-harm and harm to others based on the collected data. The best-performing algorithms have been identified, and their strengths and limitations in predicting self-harm and harm to others have been discussed. The interpretability techniques used in the study, such as feature importance rankings, partial dependence plots, and LIME, have provided valuable insights into the key features associated with self-harm and harm to others, shedding light on potential risk factors.
Moreover, the statistical analysis results reveal significant associations between risk factors and self-harm/harm to others, providing evidence of the relationship between various variables and these outcomes.
It's important to consider the study's limitations when interpreting the findings and implications. These limitations include potential biases in the data, sample size limitations, and generalizability concerns.
The research findings significantly affect mental health professionals, policymakers, and other stakeholders. The insights gained can inform the development of interventions, policies, and practices related to self-harm and harm to others, potentially contributing to the field of mental health research. Practical recommendations based on the research findings can be made to address these critical issues effectively.
In conclusion, this research project, utilizing machine learning techniques, contributes to our understanding of the risk factors associated with self-harm and harm to others. The findings can be used to inform mental health research and clinical care, leading to improved interventions and prevention efforts. Further research can build upon these findings to better understand and prevent self-harm and harm to others, ultimately contributing to the field of mental health research.
- /data: Contains the data collected from the survey on Google Forms.
- /code: Includes the Python scripts and notebooks used for data preprocessing, model training, and interpretability analysis.
- /results: Holds the outputs of the machine learning models and interpretability techniques.
- /docs: Contains additional documentation, references, and supplementary material related to the research.
- Clone the repository to your local machine using
git clone https://github.com/shouryamanekar/PRSH_ML. - Navigate to the
/codedirectory and follow the Jupyter notebooks to understand the data preprocessing and model training steps. - Explore the
/resultsdirectory to view the outputs and visualizations generated during the analysis. - Refer to the
/docsdirectory for additional information, references, and further reading.
Feel free to reach out to the project contributors for any questions or collaborations.
As of the last update to this repository, the research project is based on data collected through the survey. For future improvements and updates, consider expanding the dataset, incorporating new machine learning algorithms, and addressing the limitations identified in the research.