CODEWITHSUNDEEP
pythonanacondadata-miningapriori
AMER ALAHMAD
AMER ALAHMAD

Reputation: 11

How Get equations linking elements from rules with apriori algorithm?

I have a data base from excel file, include on 75 item, I want get equation that calculate the 75th element depending on the 74 elements value.

I am using Anaconda and python to mining data and get the rule by Apriori algorithm, I get the rule but it is not useful, because it is in An incomprehensible formula, and I have one question: * How convert the rule to equation that calculate the 75th element depending on the 74 elements value?

#Import libraries & dataset
import pandas as pd

from apyori import apriori

number_of_raws = 4000
number_of_columns = 75

store_data = pd.read_excel('C:\\Users\\smaol\\Aro\\1.xlsx',header=1)

#convert to list
transactions  = []

for i in range(0, number_of_raws):
    for j in range(0, number_of_columns):
        transactions.append(str(store_data.values[i,j]))

#found Rule
association_rules = apriori(transactions, min_support=0.025, min_confidence=0.5, min_lift=3, min_length=2)  
association_results = list(association_rules)

The result:

[RelationRecord(items=frozenset({'a', 'n'}), support=0.043116666666666664, ordered_statistics=[OrderedStatistic(items_base=frozenset({'a'}), items_add=frozenset({'n'}), confidence=1.0, lift=23.192887514495556), OrderedStatistic(items_base=frozenset({'n'}), items_add=frozenset({'a'}), confidence=1.0, lift=23.192887514495556)])]
[RelationRecord(items=frozenset({'a', 'n'}), support=0.043116666666666664, ordered_statistics=[OrderedStatistic(items_base=frozenset({'a'}), items_add=frozenset({'n'}), confidence=1.0, lift=23.192887514495556), OrderedStatistic(items_base=frozenset({'n'}), items_add=frozenset({'a'}), confidence=1.0, lift=23.192887514495556)])]
[RelationRecord(items=frozenset({'a', 'n'}), support=0.043116666666666664, ordered_statistics=[OrderedStatistic(items_base=frozenset({'a'}), items_add=frozenset({'n'}), confidence=1.0, lift=23.192887514495556), OrderedStatistic(items_base=frozenset({'n'}), items_add=frozenset({'a'}), confidence=1.0, lift=23.192887514495556)])]
[RelationRecord(items=frozenset({'a', 'n'}), support=0.043116666666666664, ordered_statistics=[OrderedStatistic(items_base=frozenset({'a'}), items_add=frozenset({'n'}), confidence=1.0, lift=23.192887514495556), OrderedStatistic(items_base=frozenset({'n'}), items_add=frozenset({'a'}), confidence=1.0, lift=23.192887514495556)])]

enter image description here

enter image description here

Upvotes: 1

Views: 205

Answers (1)

Has QUIT--Anony-Mousse
Has QUIT--Anony-Mousse

Reputation: 77454

If you always have exactly the same number of rows, this most likely is not "market basket" data. There probably is value to the position information that is lost when you transform this to itemsets.

What about using a standard prediction algorithm. Such as decision trees and random forests?

Don't use "apyrori". It's an incorrect implementation.

Upvotes: 1

Related Questions