1+ #############################
2+ # ECLAT algorithm in Python #
3+ #############################
4+
5+ #
6+ # Importing the libraries
7+ #
8+ import numpy as np
9+ import matplotlib .pyplot as plt
10+ import pandas as pd
11+ from itertools import combinations
12+
13+ #
14+ # Data Preprocessing
15+ #
16+ dataset = pd .read_csv ('Market_Basket_Optimisation.csv' , header = None )
17+
18+ #
19+ # Generating correct input format
20+ #
21+ transactions = [[str (dataset .values [j ,i ]) for i in range (0 ,20 )] for j in range (0 ,7501 )]
22+
23+ #
24+ # ECLAT function (any k-item combo)
25+ #
26+ def eclat (transactions_list , output_filename , min_support = 0.002 ):
27+
28+ """ Implements the eclat algorithm on a list of lists containing transactions
29+ in the format of the video. Data output is written to a file specified by
30+ output_filename argument.
31+
32+ transactions_list: list of lists, each list representing a transaction, e.g.
33+
34+ dataset = pd.read_csv('Market_Basket_Optimisation.csv', header = None)
35+ transactions_list = [[str(dataset.values[j,i]) for i in range(0,20)] for j in range(0,7501)]
36+
37+ output_filename: string to specify output filename, e.g. "eclat_out.tsv"
38+
39+ min_support: skips all combinations of items with supoort < min_support,
40+ e.g. 0.002
41+
42+ """
43+ import time
44+ t_start = time .time ()
45+
46+ combos_TO_counts = {}
47+ for transaction in transactions :
48+ goods = list (np .unique (transaction ))
49+ length = len (goods )
50+ for k in range (2 ,length + 1 ):
51+ k_combos = list (combinations (goods , k ))
52+ for combo in k_combos :
53+ if set (combo ).issubset (transaction ):
54+ try :
55+ combos_TO_counts [combo ] += 1
56+ except (KeyError ):
57+ combos_TO_counts [combo ] = 1
58+
59+ t_end = time .time ()
60+ t_duration = t_end - t_start
61+
62+ #
63+ # Calculate supports for combinations of goods
64+ #
65+ combo_support_vec = []
66+ for combo in combos_TO_counts .keys ():
67+ # NOTE: Support(M) = #transactions inc. M / #Total transactions,
68+ # i.e. M's popularity
69+ support = float (combos_TO_counts [combo ])/ len (transactions )
70+ combo_support_vec .append ((combo , support ))
71+ #
72+ # Sort in order of support
73+ #
74+ combo_support_vec .sort (key = lambda x : float (x [1 ]), reverse = True )
75+ #
76+ # Create tab-separated output file (skipping sets w/ < min_support)
77+ #
78+ # Note: first column is the set of goods, and the second column is the support
79+ with open ("./eclat_out.tsv" ,"w" ) as fo :
80+ for combo , support in combo_support_vec :
81+ if support < min_support :
82+ continue
83+ else :
84+ fo .write (", " .join (combo )+ "\t " + str (support )+ "\n " )
85+ print ("Completion time (seconds):" + str (t_duration ))
86+ return combo_support_vec
87+
88+ combos_vs_supports = eclat (transactions , "./eclat.tsv" , min_support = 0.002 )
0 commit comments