Parameters to be used for the underlying contrast algorith,
Parameters of the tabular contrast
Given our scored subsets try to reason about single relevancy scores for the features.
Given our scored subsets try to reason about single relevancy scores for the features. We need to do this, since the scores we calculated are based on a subset of features. So we don't know which feature has which influence on the score. But since we do have quite a number of different subsets and their scores we can combine that knowledge to deduce which features might have which influence
subsets and their relevancy scores
dimensions to analyze
map of dimensions and their relevance
Parameters to be used for the underlying contrast algorith,
Parameters of the tabular contrast
Given a feature search space, rank the features according to their decreasing usefullness
Given a feature search space, rank the features according to their decreasing usefullness
search space
ranked features
Measure and return runtime for a function
Measure and return runtime for a function
result type of f
function to measure
tuple of the result of f and the runtime
Measure the runtime of a function and print the result to an output logger
Measure the runtime of a function and print the result to an output logger
result type of f
function name to use for printing
function to measure runtime of
result of the execution of f
A tubular algorithm that uses a table of cells to keep track of estimated relevancy and redundancy scores.
The table is quadratic with one row & column for each feature. Initialy all entries are set to zero / are empty. During the search the cells are filled with estimations based on random tries.