ML Model Repository

This repository contains machine learning models for various business use cases including sales forecasting and customer behavior prediction.

📁 Model Files

1. Ice Cream Sales Forecasting Models

ML-Job-1-child1_model.pkl

• Type: Linear Regression
• Purpose: Predicts ice cream sales based on weather and temporal factors
• Features: Date, DayOfWeek, Month, Temperature, Rainfall
• Target: IceCreamsSold
• Algorithm: Linear Regression
• Serialization Date: 2026-01-29T18:19:47.681000
• Scikit-learn Version: 1.3.0

ML-Job-2-child3_model.pkl

• Type: Lasso Regression
• Purpose: Ice cream sales prediction with regularization
• Features: Date, DayOfWeek, Month, Temperature, Rainfall
• Target: IceCreamsSold
• Algorithm: Lasso Regression (alpha=1.0)
• Serialization Date: 2026-01-29T18:33:09.378000
• Scikit-learn Version: 1.3.0

ML-Job-2-child3-1_model.pkl

• Type: Lasso Regression
• Purpose: Updated ice cream sales prediction model
• Features: Date, DayOfWeek, Month, Temperature, Rainfall
• Target: IceCreamsSold
• Algorithm: Lasso Regression (alpha=1.0)
• Serialization Date: 2026-01-29T18:42:52.078000
• Scikit-learn Version: 1.3.0

2. Customer Spending Prediction Model

ML-Job-3-child3-1_model.pkl

• Type: Lasso Regression
• Purpose: Predicts customer average spending based on purchase frequency
• Features: Name, AverageFrequency
• Target: AverageSpend
• Algorithm: Lasso Regression (alpha=1.0)
• Serialization Date: 2026-01-29T18:58:15.438000
• Scikit-learn Version: 1.3.0

🛠️ Model Pipeline Structure

All models use a scikit-learn Pipeline with the following preprocessing steps:

Preprocessing Pipeline

1. ColumnTransformer for feature-specific transformations:

   • Numerical Features: Imputation + Scaling
     • SimpleImputer (median strategy)
     • StandardScaler (for Lasso models) or passthrough (for Linear Regression)
   • Categorical Features: Imputation + Encoding
     • SimpleImputer (constant strategy with "missing")
     • OneHotEncoder (first category drop, handle_unknown='ignore')

2. Regressor: Linear or Lasso regression

📊 Model Input Formats

For Ice Cream Sales Models:

{
  "input_data": {
    "columns": ["Date", "DayOfWeek", "Month", "Temperature", "Rainfall"],
    "index": [0, 1, 2],
    "data": [
      ["2025-06-15", "Sunday", "June", 75.5, 0.0],
      ["2025-06-16", "Monday", "June", 72.0, 0.1],
      ["2025-06-17", "Tuesday", "June", 78.2, 0.0]
    ]
  }
}