A p-value is the probability, under a statistical test's null hypothesis, of obtaining a result at least as extreme as the observed result.
A p-value measures the statistical significance of an observed result under a test’s null hypothesis. In ANOVA feature selection, it helps judge whether an F-score reliably separates genuine and fraudulent webpages. section 2.1.2
Use p-values to qualify feature-ranking results:
Treat an F-test as statistically significant when its p-value is below 0.01. section 2.1.2
Do not rely on an F-score when its p-value exceeds 0.01. section 2.1.2
In the phishing dataset, NumNumericChars and PctExtResourceUrls exceed 0.01, while most other numerical features fall below it. section 2.1.2
How Fight Fraud with Machine Learning covers it
How deep it goes
Defines a practical significance threshold and inspects returned values, but does not derive p-values or test assumptions.
Examples
Uses phishing-page features, including NumNumericChars and PctExtResourceUrls, with scikit-learn p-value output.
What it assumes
Assumes numerical features, genuine-versus-fraudulent labels, and familiarity with statistical tests and feature ranking.
What you can do after
Evaluate whether ANOVA F-scores are statistically significant enough to support numerical-feature analysis.
How it connects
Connects p-values to null hypotheses, ANOVA F-tests, F-scores, and feature selection.
Statistics Slam DunkA p-value is the probability, under a specified null hypothesis, of obtaining a result at least as extreme as the observed result according to a test statistic.
Statistics Every Programmer NeedsA p-value is the probability, under a specified null hypothesis, of obtaining a result at least as extreme as the observed result.
Effective Data AnalysisA p-value is the probability of obtaining a test statistic at least as extreme as the observed value assuming the null hypothesis is true.
Grokking BayesA p-value is the probability, under a null hypothesis, of obtaining data at least as extreme as the observed data according to a specified test statistic.
bayes factoralternative to · A Bayes factor is the ratio of the marginal likelihoods of two competing models or hypotheses, quantifying their relative support from observed data.
correlation testingproduced by · Correlation testing is a statistical procedure for evaluating the association between variables, often against a specified null value.
Shapiro–Wilk testproduced by · The Shapiro–Wilk test is a statistical hypothesis test assessing whether a sample is consistent with a normal distribution.
hypothesis testingused by · Hypothesis testing is a formal statistical procedure for evaluating evidence from observed data about a specified hypothesis, often a null hypothesis.
This explainer was generated by Manning's knowledge graph from the text of Fight Fraud with Machine Learning. The section links point to the parts of the book it draws on.