def generate_frequent_items(minimum_support):
frequent_itemsets = fpg.find_frequent_itemsets(data_list, minimum_support=0.07 * len(data_list),
include_support=True)
print(type(frequent_itemsets)) # print type
result = []
for itemset, support in frequent_itemsets: # 将generator结果存入list
result.append((itemset, support / len(data_list)))
result_patterns = [i[0] for i in result]
result_support = [i[1] for i in result]
patterns_df = pd.DataFrame({"fluent_patterns": result_patterns, "support": result_support})
patterns = {}
for i in result:
patterns[frozenset(sorted(i[0]))] = i[1]
print("-------------挖掘频繁项集---------------")
print(patterns_df)
print("--------------频繁项集------------------")
return patterns
# python3
# -*- coding: utf-8 -*-
# @Author : lina
# @Time : 2018/5/13 11:40
import fp_growth_py3 as fpg
# 数据集
dataset = [['啤酒', '牛奶', '可乐'], ['尿不湿', '啤酒', '牛奶', '橙汁'], ['啤酒', '尿不湿'],
['啤酒', '可乐', '尿不湿'], ['啤酒', '牛奶', '可乐']]
if __name__ == '__main__':
'''
调用find_frequent_itemsets()生成频繁项
@:param minimum_support表示设置的最小支持度，即若支持度大于等于inimum_support，保存此频繁项，否则删除
@:param include_support表示返回结果是否包含支持度，若include_support=True，返回结果中包含itemset和support，否则只返回itemset
'''
frequent_itemsets = fpg.find_frequent_itemsets(dataset,
minimum_support=1,
include_support=True)
print(type(frequent_itemsets)) # print type
result = []
for itemset, support in frequent_itemsets: # 将generator结果存入list
result.append((itemset, support))
result = sorted(result, key=lambda i: i[0]) # 排序后输出
for itemset, support in result:
print(str(itemset) + ' ' + str(support))
f.write(k + ' ' + str(v) + '\n')
if __name__ == '__main__':
print('Reading input file...')
articles = input('台南 美食(100頁).txt')
print('Splitting...')
s = split_into_word(articles, 'str')
print('Converting to transactions...')
trans = text2trans(s)
sw_path = 'stop_words.txt'
print('Cleaning symbols...')
trans = clean(trans, sw_path) # remove stop words
trans = rm_dup(trans)
print('Finding frequent patterns...')
fp = fp.find_frequent_itemsets(trans, 2, True)
fp = sort(fp) # sort by support
food = food_dict('food_list.txt')
result = find_food(fp, food) # return's a dictionary
print('Writing output file...')
output(fp, 'fp台南(100頁).txt')
output_result(result, 'fp台南(100頁)(only food).txt')
print('Done!')