wolverine3301 / sibyl Goto Github PK

some Ranking working only on categorical

• info gain, gain ratio, and gini are only working on categorical columns, there is a way to calculate them on numeric ones as well
  Exception in thread "main" java.lang.IllegalArgumentException at java.base/java.util.PriorityQueue.<init>(PriorityQueue.java:171) at info_gain.InformationGain.gain(InformationGain.java:37) at scout.CategoryRanker.makeGain(CategoryRanker.java:71) at scout.CategoryRanker.<init>(CategoryRanker.java:42) at scout.CategoryRan_test.main(CategoryRan_test.java:11)

add printing and access for contingency tables ,critical values , and degrees of freedom

• access and print for observed table

• access and print for expected table

• print and access to chi2 critical values

• print and acces to degrees freedom

• overall print

uninitialized attributes

make soft copy column list for different types of columns. also have counts for how many in each.
have them be updated on creation of the dataframe, when a column type changes, one is added/removed.

when testing naive bayes with cross validation set to 10 trails, the late trails have terrible scores differnt prior to the update.

there is a rounding error after a dataframe is standardized.
after standardization the mean should equal 0 and the std should equal 1

is 0 sometimes when it shouldnt

Exception: Comparison method violates its general contract
possibly something wrong in particle comparator.

while it does work and puts roughly the same number of items in each bin even when not evenly divisible, there seems to be a slight unbalance of items in bins when it should be able to put an exactly even number in each.

probably when false negatives is equal to 0

when using multiple dataframe manipulation methods such as Util.splitontarget, acquire, and replace columns, the counts and arraylist of columns become out of sync. the problem seems to be around setStatistics() which adds columns to their arraylist and updates their counts. the errors experienced are either columns not being added or duplicated columns added.

broke up a dataframe using aquire and the numeric indexing is not set. could be not initalized from the start or reset when aquire is used