apriori module

class apriori.Apriori(transactions, uniques, min_sup=2.0, min_conf=1.5)

Bases: object

Apriori algorithm to find frequent itemset and extract the association rules

addrule(left, right)

Adds the rule if it is greater than min conf

Parameters:

• left – Left hand side itemset (list)
• right – Right hand side itemset (list)

Returns:

apriori_gen(Fk, k)

Used to generate size k candidates based on the frequent itemsets Fk

:param Fk list of k-1 itemsets, k:number of items that should have the new itemsets :return: new_candidates: list of size k-itemsets,

apriori_run()

Main frequent itemset generation algorithm - Apriori.

Returns:Frequent itemsets (dict)

confidence(left, right, left_precalculated=None)

Calculates the confidence of the rule

Parameters:

• left – left handside of the rule (list)
• right – right handside of the rule (list)
• left_precalculated – Existing suppport count value

Returns:

Confidence value (float)

diffelems(list1, list2)

Extracts the difference elements in lists, symmetrically.

For example:

list1 = [‘a’,’b’,’c’], list2 = [‘b’,’c’] diffelems(list1, list2) # regardless of order it will return same result returns -> [‘a’]

Parameters:

• list1 – First list to be compared
• list2 – Second list to be compared

Returns:

The difference elements of first and second lists

export(path)

Export rules in PMML format to visualize later on using R-packages.

Parameters:path – path to write the PMML file Returns:True on on successful operation

extract()

Association rule extraction method

Returns:

freq1_itemsets(f1itemsets)

Extracts the 1 length frequent itemsets

Parameters:f1itemsets – 1-length frequent itemsets (dict) Returns:1-length frequent itemsets (list) greater that self.min_sup

lift(left, right, right_precalculated=None)

Calculates the lift of the rule

Parameters:

• left – left handside of the rule (list)
• right – right handside of the rule (list)
• right_precalculated – Existing suppport count value

Returns:

Confidence value (float)

save_freqis(path='frequent_itemsets.csv')

Save the frequent itemsets into a file

Parameters:path – Path to be saved Returns:Returns true on successful save

save_rules(path='arules.csv')

Saves the association rules into a file

Parameters:path – Path to be saved Returns:Returns true on successful action

support(itemset, itemset_precalculated=None)

Counts support

Parameters:

• itemset – itemset (list)
• itemset_precalculated – Existing suppport count value

Returns:

support of the itemset (float)

support_count(itemset)

Computes the support count of the itemset

Parameters:itemset – itemset to be counted (list) Returns:support count of the itemset (int)