?? Empowering Mental Health Through AI: My Journey at MIT ??

As part of my AI studies at MIT, I am thrilled to share a recent project where we leveraged advanced machine learning techniques to analyze mental health treatment-seeking behavior. This work has the potential to impact lives positively by providing insights that can lead to better mental health support in the workplace.

Before I continue I want to thank and congratulate David Garcia , Lordan Bulanadi and Michael O'Shea for the great teamwork.

?? The Importance of Mental Health at Work: Mental health is crucial for overall well-being, especially in the workplace where stress and pressures are common. While it's essential to care for everyone, certain groups are more vulnerable to mental health issues. By identifying these groups and understanding their needs, we can create a more inclusive and supportive environment. These findings can drive in-depth work and actions, making it easier for employees to discuss mental health openly and seek the help they need.

My Personal Motivation: Having experienced anxiety crises in the past, I feel a strong inclination to embrace a programmatic project aimed at uncovering and helping people in need. This personal experience fuels my passion for using data-driven approaches to make a difference in mental health care.

?? Project Overview: We embarked on an in-depth analysis of a dataset from Kaggle containing information about employees' mental health status and workplace environment. Our goal was to identify the key factors influencing the likelihood of seeking mental health treatment.

Methodology:

Data Cleaning and Preprocessing:

• Feature Selection and Dropping Irrelevant Columns: Ensuring only relevant data is used.
• Handling Missing Values: Imputing missing data to maintain dataset integrity.
• Encoding Categorical Variables: Converting categorical data into numerical format using techniques like One-Hot Encoding.
• Standardization: Scaling data to ensure consistency and improve model performance.

Principal Component Analysis (PCA):

• Dimensionality Reduction: Reducing the number of features while retaining 90% of the variance.
• Explained Variance: Identified that 21 principal components were needed to explain the variance in the dataset.
• Data Visualization: Simplifying the dataset to visualize complex relationships.

Clustering Analysis:

• Elbow Method: Used to determine the optimal number of clusters.
• K-means Clustering: Applied to group similar data points into 3 distinct clusters, uncovering hidden patterns and personas within the data.`

Random Forest Analysis:

• Feature Importance: Trained a Random Forest classifier to rank features by importance.
• Model Accuracy: Achieved an accuracy of 98% on the test set, indicating reliable predictions.
• Handling Non-linear Interactions: Capturing complex relationships between features without overfitting.

?? Key Findings:

• Work Interfere: The most significant predictor of seeking treatment.
• Age: Younger employees are more likely to seek treatment.
• Family History: A strong indicator of treatment-seeking behavior.
• Care Options and Benefits: Availability at the workplace encourages seeking treatment.
• Remote Work: Analyzed the impact of remote work on mental health treatment-seeking behavior using causal inference techniques.
• Population Seeking Treatment: 50.48% of the population in our dataset sought mental health treatment.

Personas Identified from Clustering Analysis:

Persona 1: Cluster 0

• Demographics: Older than average, balanced gender distribution.
• Work Characteristics: Mostly not self-employed, works in medium-sized companies, less likely to work remotely, likely in tech companies.
• Mental Health: Moderate treatment levels, less likely to have a family history of mental illness, less likely to experience work interference due to mental health.
• Probability of Seeking Treatment: 45%

Persona 2: Cluster 1

• Demographics: Slightly younger than average, more males.
• Work Characteristics: More likely to be self-employed, high remote work presence, works in smaller companies, likely in tech companies.
• Mental Health: High treatment levels, average family history of mental illness, more concern about mental health consequences at work.
• Probability of Seeking Treatment: 55%

Persona 3: Cluster 2

• Demographics: Near average age, slightly more females.
• Work Characteristics: Mostly not self-employed, works in medium-sized companies, less likely to work remotely, likely in tech companies.
• Mental Health: Highest treatment levels, higher family history of mental illness, slightly above average work interference due to mental health.
• Probability of Seeking Treatment: 65%

?? Why This Matters: Understanding these factors can help organizations design targeted interventions and support strategies to improve mental health outcomes for their employees. By combining machine learning with causal inference, we can move beyond correlation to understand the underlying causes and make data-driven decisions that benefit people's lives.

I am proud of the work we have done and excited about the potential impact of these insights on mental health in the workplace. This project is a testament to the power of AI in making a difference in the healthcare sector.

#AI #MachineLearning #HealthcareAnalysis #MentalHealth #MIT #DataScience #RandomForest #PCA #Clustering #MentalHealthAwareness #Kaggle