🏡 Property Recommendation System ✨ A machine learning–powered tool to assist appraisers by recommending the top 3 most comparable properties in seconds.

📌 Overview This project leverages XGBoost to predict and recommend the best comparable properties for a given subject property. It reduces the time and subjectivity in manual property comparisons.

🚀 Key Features ⚙️ Developed a ML model using XGBoost 📊 Trained on real-world appraisal data 🧠 Automatically identifies top 3 comparable properties for each subject property 🕒 Saves time and increases consistency compared to manual appraisal 📍 Incorporates address similarity using cosine similarity 🧼 Standardizes and encodes property features for accurate predictions

How It Works -Load appraisal data, including both subject properties and comparable properties (comps).

-Extract and clean features from all properties, ensuring consistency between comps and subjects.

-Use cosine similarity to compare and score address similarity between properties.

-Train an XGBoost classifier to predict whether a given property is a valid comp for a subject property.

-For each subject property, rank all potential comps and recommend the top 3 based on predicted relevance.