Business problem
Why this project matters
Designed neural-network classifiers for wind-turbine generator failure prediction with class-imbalance handling and recall-focused evaluation for proactive maintenance.
Project scope
What the work covered
01Analyze sensor and failure-label distributions
02Prepare data and address class imbalance
03Train multiple neural-network configurations
04Compare precision, recall, F1, and confusion matrices
05Select a model aligned with maintenance risk
Workflow
From question to evidence
Business problem
→
Data preparation
→
EDA / feature work
→
Model or analysis
→
Evaluation
→
Business conclusion
Analysis dashboard
Portfolio evidence areas
These bars summarize the emphasis of the completed work; detailed numeric outputs and full notebook cells remain available in the linked repository and HTML report.
Notebook evidence
Selected analysis and report visuals
Notebook / HTML report snapshot 1
Notebook / HTML report snapshot 2
Notebook / HTML report snapshot 3
Notebook / HTML report snapshot 4
Notebook / HTML report snapshot 5
Notebook / HTML report snapshot 6
Images are loaded from the public project repository so the portfolio stays synchronized with the published notebook/report assets.
Conclusions
What the analysis demonstrated
Recall is the critical metric when missed failures have high operational cost.
Class imbalance can make accuracy misleading without threshold and confusion-matrix analysis.
The strongest evaluated configuration achieved validation recall of approximately 0.934.
GitHub README preview
Repository at a glance
renewind-inn-failure-prediction
Neural-network failure prediction for renewable-energy equipment with class-imbalance and recall-focused analysis.
Next step
Discuss a similar project
Use the project inquiry form with ReneWind preselected. Include the business objective, available data, timeline, and desired output.