Metrics and Analysis

Base Interfaces

Metric	Base class for LensKit metrics.
ListMetric	Base class for metrics that measure individual recommendation (or prediction) lists, and whose results may be aggregated to compute overall metrics.
GlobalMetric	Base class for metrics that measure entire runs at a time.
MetricFunction	Interface for per-list metrics implemented as simple functions.
RankingMetricBase	Base class for most ranking metrics, implementing a k parameter for truncation.

Bulk Analysis

RunAnalysis	Compute metrics over a collection of item lists composing a run.
RunAnalysisResult	Results of a bulk metric computation.

Basic Statistics

ListLength	Report the length of the output (recommendation list or predictions).
TestItemCount	Report the number of test items.

Top-N Accuracy

NDCG	Compute the normalized discounted cumulative gain [JarvelinKekalainen02].
RBP	Evaluate recommendations with rank-biased precision [MZ08].
Precision	Compute recommendation precision.
Recall	Compute recommendation recall.
RecipRank	Compute the reciprocal rank [KV97] of the first relevant item in a list of recommendations.

List and Item Properties

MeanPopRank

Compute the _obscurity_ (mean popularity rank) of the recommendations.

Item Distributions

ExposureGini	Measure exposure distribution of recommendations with the Gini coefficient.
ListGini	Measure item diversity of recommendations with the Gini coefficient.

Prediction Accuracy

RMSE	Compute RMSE (root mean squared error).
MAE	Compute MAE (mean absolute error).

Rank Weights

The rank weighting classes (RankWeight and descendants) provide flexible rank weights for use in evaluation metrics. The rank-weighted top-N metrics (Top-N Accuracy) use these for weighting the recommendations.

RankWeight	Base class for rank weighting models.
GeometricRankWeight	Geometric cascade weighting for result ranks.
LogRankWeight	Logarithmic weighting for result ranks, as used in NDCG.