7 skills.
Use when asked to detect anomalies/outliers in data — statistical, machine-learning, and deep-learning detection methods, supervised vs. unsupervised approaches — as distinct from trend-analysis's focus on overall patterns over time rather than individual deviant points.
Use when asked to apply Bayes' theorem — updating a probability estimate given new evidence, P(A|B) = P(B|A)·P(A)/P(B) — as one specific probabilistic technique within the broader statistical-analysis toolkit (see statistical-analysis), foundational to Bayesian inference and Bayesian networks.
Use when asked to test whether two categorical variables are significantly associated — contingency tables, observed vs. expected frequencies, the chi-square test statistic — as one specific inferential-statistics technique within the broader toolkit (see statistical-analysis).
Use when asked to model or solve a constraint satisfaction problem (CSP) — variables, domains, and constraints solved via backtracking, forward checking, or constraint propagation — as distinct from Monte Carlo methods (see monte-carlo-analysis), which use random sampling rather than systematic constraint solving.
Use when asked to solve a mathematical or engineering problem via repeated random sampling — the general Monte Carlo computational method — as distinct from monte-carlo-testing's specific application of the same underlying technique to software reliability testing.
Use when asked which statistical technique fits a data question — descriptive vs. inferential statistics, regression, time series, factor/cluster analysis, data mining — as the umbrella technique-selection layer above more specific methods like bayes-theorem-analysis, chi-square-analysis, and monte-carlo-analysis.
Use when asked to analyze data over time to identify patterns and forecast — regression, moving averages, exponential smoothing — as distinct from anomaly-analysis's focus on individual deviant points rather than the overall directional pattern.