Skip to content

Data Science Fundamentals, Best Practices

+analytics

Within the context of a project's declared

  • Problem Statement
  • Outcome Expectations, Underlying Aims
  • Deployment Goal


Underpinnings of ethical model development

 Comment
FeaturesFeatures profiling. Contextually justifiable? Relevant?
Bias AnalysisVia the distribution analysis of features; data bias, variance, and noise.1, 2
Model Analytics & InterpretationAn error metrics, model-algorithm input data, cost, and interpretation synthesis.
Model/Algorithm Mechanics An explanation of the mechanics of the model's underlying algorithm.
  • The algorithmic bias and variance of artificial-intelligence/machine-learning algorithms/architectures; the algorithms/architectures that underpin model development.
  • The strengths, weaknesses, usage constraints, of algorithms/architectures.
AuditabilityGit, etc.
Re-trainingArchitecture











  1. Understanding the Bias-Variance Trade-off ↩

  2. Noise: A Flaw in Human Judgment ↩