vrushaliap/Ethiopia_food_decisiontree

This repository contains a machine learning project analyzing the Ethiopian food market using the Decision Tree Classifier. The study explores pricing volatility, supply chain challenges, and market dynamics, achieving 96% accuracy in predictions. # Ethiopia_food_decisiontree # 🌾 Ethiopian Food Market Analysis using Decision Tree Classifier This repository contains a machine learning project analyzing the Ethiopian food market using the Decision Tree Classifier. The study explores pricing volatility, supply chain challenges, and market dynamics, achieving 96% accuracy in predictions. ## 📌 Overview This project analyzes the Ethiopian food market dataset to uncover key insights into pricing, distribution, and market trends. The **Decision Tree Classifier** was applied to predict market patterns with high accuracy, helping identify disparities and improve strategic decision-making. ## 🎯 Objectives - Apply **Decision Tree Classifier** to predict Ethiopian food market trends. - Identify significant patterns in product pricing and distribution. - Provide insights for stakeholders to address **supply chain and pricing volatility**. - Support better decision-making for **food security and sustainability**. ## 🧑‍💻 Methodology - **Dataset:** Ethiopia Weekly FEWS NET Staple Food Price Data (90,558 rows, 15 columns). - **Preprocessing Steps:** - Missing values (25% in "Value") filled with median. - One-hot encoding applied to categorical variables. - Outliers dected but were not removed if anyone wants to remove the outliers it can be removed using z-scores. - **Algorithm Used:** Decision Tree Classifier. - **Evaluation Metrics:** Accuracy, Sensitivity, Specificity, F1 Score. ## 📊 Results ### Model Accuracy The Decision Tree Classifier achieved **96.34% accuracy** with perfect sensitivity, specificity, and F1-score across folds. ### Market Insights Analysis revealed pricing disparities in different Ethiopian markets: - Essential imported goods (rice, sugar, oil) showed **low quantities**. - Beddenno market had consistently **higher costs** for basic items like diesel, oil, and whole grain. ## 📂 Repository Structure ├── data/ # Raw and cleaned datasets ├── notebooks/ # Jupyter notebooks for Decision Tree …