From c382b655367ba7ba419da6bb6a63a6b9b5478036 Mon Sep 17 00:00:00 2001
From: K Om Senapati <komnoob123@gmail.com>
Date: Sat, 29 Jun 2024 03:36:05 +0000
Subject: [PATCH] Added README.md for Student Stress Levels Detection project

---
 .../Model/README.md                           |  0
 Student Stress Levels Detection/README.md     | 71 +++++++++++++++++++
 2 files changed, 71 insertions(+)
 delete mode 100644 Student Stress Levels Detection/Model/README.md
 create mode 100644 Student Stress Levels Detection/README.md

diff --git a/Student Stress Levels Detection/Model/README.md b/Student Stress Levels Detection/Model/README.md
deleted file mode 100644
index e69de29bb..000000000
diff --git a/Student Stress Levels Detection/README.md b/Student Stress Levels Detection/README.md
new file mode 100644
index 000000000..082681da1
--- /dev/null
+++ b/Student Stress Levels Detection/README.md	
@@ -0,0 +1,71 @@
+## Student Stress Levels Detection
+
+### 🎯 **Goal**
+
+To understand the influence of the headache, study load, academic performance etc on students stress level.
+
+### 🧵 **Dataset**
+
+https://www.kaggle.com/datasets/samyakb/student-stress-factors/data
+
+### 🧾 **Description**
+
+By using Data Analysis, Data Visualization and then applying various Regression Algorithms from Machine Learning to predict a students stress level.
+
+### 🧮 **What I had done!**
+
+1. Checked the shape of dataset, types of columns and basic statistics
+2. Then checked the histplots and boxplot of all the features
+3. Then checked distribution of stress level
+4. Then checked reg plots of other features with stress level
+5. Then checked pair plots and correlation heatmap of all the features
+6. Then divided the dataset into features and target and normalised features and splitted dataset into training and testing sets
+7. Then trained on each model and checked the MAE and R2 on testing dataset
+8. Then selected Decision Tree Regressor and did hyper parameter tuning on that and got the best model and saved it.
+
+### 🚀 **Models Implemented**
+
+- Linear Regression
+- Decision Tree
+- Random Forest
+- Gradient Boosting
+- Support Vector Regressor
+- XGBoost Regressor
+
+### 📚 **Libraries Needed**
+
+- matplotlib
+- seaborn
+- numpy
+- pandas
+- scikit_learn
+- xgboost
+
+### 📊 **Exploratory Data Analysis Results**
+
+![Box Plots](./Images/box_plots.png)
+![Reg Plots](./Images/reg_plots.png)
+![Stress KDE](./Images/stress_level_kde.png)
+
+### 📈 **Performance of the Models based on the Accuracy Scores**
+
+| Model                   | R^2 Score | MAE  |
+|-------------------------|-----------|------|
+| Linear Regression       | 0.23      | 1.01 |
+| Decision Tree           | 0.91      | 0.13 |
+| Random Forest           | 0.90      | 0.18 |
+| Gradient Boosting       | 0.74      | 0.55 |
+| Support Vector Regressor| 0.60      | 0.60 |
+| XGBoost Regressor       | 0.91      | 0.13 |
+
+
+### 📢 **Conclusion**
+
+Both Decision Tree and XGBoost have similar performance in terms of R^2 score and MAE.
+Selecting **Decision Tree Regressor** as its easier to interpret as you can visualize the tree and understand the decision-making process.
+
+### ✒️ **Your Signature**
+
+K Om Senapati <br />
+[![GitHub](https://img.shields.io/badge/GitHub-kom--senapati-blue?style=flat&logo=github)](https://github.com/kom-senapati)
+[![Twitter](https://img.shields.io/badge/Twitter-kom__senapati-blue?style=flat&logo=twitter)](https://twitter.com/kom_senapati)