Skip to content
Open sourceAI / MLTime series · Regression

Appliances Energy Prediction

Comparing regression models for smart-home energy consumption

Role
Data science
Period
2025
Status
Open source
Links
GitHub ↗

By the numbers

19,735
sensor rows
0.76
R² (LightGBM)
4
models compared

Overview

Predicting appliance energy use from 19,735 ten-minute readings in the UCI dataset. Random Forest, XGBoost, LightGBM and linear regression were compared; LightGBM led with R² ≈ 0.76.

  • The right-skewed target was log-transformed; hour, weekday and weekend features were derived.
  • SelectKBest, RFE and PCA compared: SelectKBest gave a small gain, PCA hurt on this problem.

01

Problem

Appliance energy use is right-skewed with a few high peaks; indoor temperature/humidity sensors correlate strongly with outdoor weather. The goal was to predict ten-minute consumption and see which model family suits this structure.

02

Approach

Exploratory analysis on the 34-column dataset (time series, correlation heatmap, hourly box plots), log transform and time features. Linear regression as baseline; Random Forest, XGBoost and LightGBM trained with hyperparameter search. Evaluated on RMSE and R².

03

Architecture

  1. Data
    • UCI veri seti
    • 19.735 × 34
  2. Preparation
    • Log dönüşümü
    • Zaman öznitelikleri
    • SelectKBest
  3. Models
    • Linear
    • Random Forest
    • XGBoost
    • LightGBM
↓ Data flows top to bottom

04

Key decisions

  1. 01

    Model the target on a log scale

    Peaks dominated RMSE. The log transform balanced the distribution and let tree models learn the low-consumption regime too.

Outcome

LightGBM gave the best result with R² ≈ 0.76 and RMSE ≈ 0.217 on the log scale. SelectKBest feature selection lowered RMSE from 0.2138 to 0.2122; PCA raised it to 0.36, showing it is unsuitable for this problem.

Related projects

  • AI / ML
    Open source

    Traffic Accident Severity Prediction

    Route-risk prediction trained on 100,000+ accident records

    Route-risk prediction trained on 100,000+ historical accident records, issuing real-time alerts at 85%+ accuracy using OpenRouteService and weather APIs.

    100K+
    accident records
    85%+
    accuracy
    • Python
    • Scikit-Learn
    • XGBoost
    • Pandas
    • Flask
    • React
    Open
    GitHub ↗
  • AI / ML
    Open source

    Mobile Price Classification

    Classifying phone price tiers from hardware specifications

    A classification study predicting a phone's price tier from hardware specs such as RAM, battery, screen and camera; includes a data-collection script, Jupyter analysis and a containerized prediction service.

    • Python
    • Pandas
    • Scikit-Learn
    • Jupyter
    • Docker
    Open
    GitHub ↗
  • AI / ML
    Live

    ChefsStack

    Recipe platform with a vector-based recommendation engine

    A recipe and recommendation platform where suggestions come from vector similarity instead of hand-written rules, and user interactions flow through Kafka and Redis into personalized feeds.

    <50ms
    similarity query
    1.000+
    interaction events processed
    • Python
    • FastAPI
    • PostgreSQL
    • pgvector
    • Redis
    • Kafka
    Open
    Visit site ↗