🥕 Vegetable Price Prediction

A machine learning project to predict the price per kg of vegetables based on environmental conditions, seasonality, and product condition using various regression techniques including Linear Regression, Support Vector Machine (SVM), and Random Forest Regressor.

---

📊 Dataset Overview

The dataset contains information about vegetables collected from a local market, with the following features:

• Vegetable: Type of vegetable (e.g., tomato, potato, cucumber)
• Season: Seasonal category (e.g., summer, winter)
• Month: Month of observation
• Temp: Average temperature
• Deasaster Happen in last 3month: Whether a natural disaster occurred recently
• Vegetable condition: Quality (e.g., fresh, average, scrap)
• Price per kg: Target variable

---

🔧 Preprocessing Steps

• Data Cleaning: Fixed typos (e.g., "scarp" → "scrap")
• Handling Missing Values: Replaced blank or missing months with mode
• Encoding: Used one-hot encoding for categorical variables and ordinal encoding for months
• Train-Test Split: 70% training, 30% testing

---

📈 Models & Performance

Model	R² Score	Mean Squared Error
Linear Regression	0.81	582.61
Support Vector Machine (SVM)	0.19	3430.13
Random Forest Regressor	0.91	271.64

✅ Random Forest performs best in both R² and MSE.

---

🚀 Installation & Run Locally

1. Clone this repo:

   git clone https://github.com/your-username/vegetable-price-prediction.git
   cd vegetable-price-prediction

2. Install dependencies:

pip install -r requirements.txt

3. Run the notebook:

 jupyter notebook

🛠️ Tools Used

Python

Pandas, NumPy

Scikit-learn

Matplotlib, Seaborn

Jupyter Notebook

📌 Key Takeaways

Vegetable pricing is influenced by seasonal and environmental factors.

Machine learning can effectively forecast prices with relatively small datasets.

Feature encoding plays a crucial role in model accuracy.

📸 Sample Visualization

Feature correlation for better understanding input relationships