Assess trained models beyond aggregate metrics for CRISP-DM workflows.
npx skills add https://github.com/thbraet/claude-template --skill assess-model
Please help me install this Agent Skill. Skill: assess-model Source: https://github.com/thbraet/claude-template/tree/main/skills/assess-model Command: npx skills add https://github.com/thbraet/claude-template --skill assess-model
This skill performs deep assessment of trained models beyond aggregate metrics, analyzing where models succeed and fail, identifying systematic biases, evaluating overfitting risk, and translating technical performance into business impact to guide model selection.
Use a target model name or MLflow run to generate the assessment notebook and summary document at the expected locations.
High-intent search queries and answers about installing and using this skill.
Model assessment in CRISP-DM involves analyzing subgroup performance, residual patterns, and overfitting checks to identify systematic biases. This skill generates a structured notebook and summary document translating technical metrics into business impact.
You can check overfitting and subgroup bias using MLflow runs by applying deep model assessment. The skill analyzes training artifacts to evaluate systematic biases, identify residual patterns, and report subgroup performance beyond standard aggregate metrics.
Yes, you can translate error analysis into business impact. The skill evaluates deep model assessments and produces a structured summary document with evidence-based model selection guidance to inform stakeholders before deployment.
To run deep model assessment, you need Python data science libraries, access to training artifacts like notebooks and MLflow runs, and a target model name. The skill outputs a notebook at notebooks/4.4-model-assessment.ipynb and a summary document.
You should move a model to the CRISP-DM Evaluation phase after completing comprehensive error analysis, overfitting checks, and business-impact translation. This skill provides the deployment guidance and structured documentation needed to advance models.
manutejAutomate ML lifecycle management with MLflow experiment tracking and model registry.
thbraetGenerate a CRISP-DM deployment retrospective markdown document at docs/crisp-dm/6-deployment/6.4-review-project.md.
superfhpTrack machine learning experiments and manage models with wandb.
wachawoDecomposes aggregate metric data into dimensional segments to identify root causes.
