Name: Mohammad Faizan

Roll No:312027
CLass: TE Experiment no: 9 Aim: To implement Apriori Algorithm

import pandas as pd

from google.colab import files

uploaded = files.upload()

import pandas as pd
import io

df = pd.read_csv(io.BytesIO(uploaded['dataset_apriori.csv']))
print(df)

Choose Files No file chosen Upload widget is only available when the cell has been executed in the
current browser session. Please rerun this cell to enable.
Saving dataset_apriori.csv to dataset_apriori.csv
tid items
0 1 Handphone,Laptop
1 2 Handphone,Charger,Laptop
2 3 Powerbank,Laptop,Charger,Handphone
3 4 Tablet,Laptop,Handphone
4 5 Handphone,Charger,Tablet
5 6 Tablet,Powerbank
6 7 Handphone,Laptop,Tablet,Charger
7 8 Charger,Handphone
8 9 Handphone,Powerbank
9 10 Laptop,Charger,Powerbank

dataset = pd.read_csv('dataset_apriori.csv')

dataset
 tid items

0 = 1
df_items dataset['items'] Handphone,Laptop
df_tid = dataset['tid']
1 2 Handphone,Charger,Laptop

2 3 Powerbank,Laptop,Charger,Handphone
df_items
3 4 Tablet,Laptop,Handphone
0 Handphone,Laptop
14 5 Handphone,Charger,Tablet
Handphone,Charger,Laptop
2 Powerbank,Laptop,Charger,Handphone
35 6 Tablet,Powerbank
Tablet,Laptop,Handphone
4 Handphone,Charger,Tablet
6 7 Handphone,Laptop,Tablet,Charger
5 Tablet,Powerbank
67 8 Handphone,Laptop,Tablet,Charger
Charger,Handphone
7 Charger,Handphone
88 9 Handphone,Powerbank
Handphone,Powerbank
9 Laptop,Charger,Powerbank
9
Name:10items, dtype:Laptop,Charger,Powerbank
object

df_tid

0 1
1 2
2 3
3 4
4 5
5 6
6 7
7 8
8 9
9 10
Name: tid, dtype: int64

Convert Items to Number

dictionaries = {'Handphone': 1, 'Laptop': 2, 'Charger': 3, 'Powerbank': 4, 'Tablet': 5 }

type(df_items)

pandas.core.series.Series

comma_splitted_df = df_items.apply(lambda x: x.split(','))

comma_splitted_df

0 [Handphone, Laptop]
1 [Handphone, Charger, Laptop]
2 [Powerbank, Laptop, Charger, Handphone]
3 [Tablet, Laptop, Handphone]
4 [Handphone, Charger, Tablet]
5 [Tablet, Powerbank]
6 [Handphone, Laptop, Tablet, Charger]
7 [Charger, Handphone]
8 [Handphone, Powerbank]
 9 [Laptop, Charger, Powerbank]
Name: items, dtype: object
numbered_col = []
for i in range(len(comma_splitted_df)):
list_numbered = list(map(lambda x: dictionaries[x], comma_splitted_df[i]))
sort_numbered = sorted(list_numbered)
numbered_col.append(sort_numbered)

numbered_col

[[1, 2],
[1, 2, 3],
[1, 2, 3, 4],
[1, 2, 5],
[1, 3, 5],
[4, 5],
[1, 2, 3, 5],
[1, 3],
[1, 4],
[2, 3, 4]]

Create Dataframe from Numbered Items

dict_data = {'items': numbered_col}

df = pd.DataFrame.from_dict(dict_data)

df

items

0 [1, 2]

1 [1, 2, 3]

2 [1, 2, 3, 4]

3 [1, 2, 5]

4 [1, 3, 5]

5 [4, 5]

6 [1, 2, 3, 5]

7 [1, 3]

8 [1, 4]

9 [2, 3, 4]

pd.concat([df_tid, df], axis=1)

 tid items

0 1 [1, 2]

1 2 [1, 2, 3]

2 3 [1, 2, 3, 4]

3 4 [1, 2, 5]

4 5 [1, 3, 5]

5 6 [4, 5]

6 7 [1, 2, 3, 5]
items = []
for i 7 8 [1, 3]
in range(len(df)):
for
8 j in
9 range(len(df['items'][i])):
[1, 4]
items.append(df['items'][i][j])
items 9 10 [2, 3, 4]

[1,
2,
1,
2,
3,
1,
2,
3,
4,
1,
2,
5,
1,
3,
5,
4,
5,
1,
2,
3,
5,
1,
3,
1,
4,
2,
3,
4]

Creating First Candidate (C1)

#Get unique element from list/array

unique_item = set(items)
unique_item

{1, 2, 3, 4, 5}

#Convert it to list
list_unique_item = list(unique_item)
list_unique_item

[1, 2, 3, 4, 5]

count_unique = []
for value in (list_unique_item):
count_unique.append((value, items.count(value)))
count_unique

[(1, 8), (2, 6), (3, 6), (4, 4), (5, 4)]

candidate1_df = pd.DataFrame(count_unique, columns=["itemset", "sup"])

candidate1_df

itemset sup

0 1 8

1 2 6

2 3 6

3 4 4

4 5 4

Creating first Frequent Itemset (L1)

def filter_sup(candidate):
minimum_sup = 2
filtering = candidate['sup'] > minimum_sup
freq = candidate[filtering]
return freq

# filter_sup = candidate1_df['sup'] > minimum_sup

freq_itemset1 = filter_sup(candidate1_df)

freq_itemset1

itemset sup

0 1 8

1 2 6

2 3 6

3 4 4

4 5 4

Create the Second Candidate (C2)

 Self join

import numpy
def self_join(prev_freq_itemset):
self_join_candidate = []
for i in range(len(prev_freq_itemset['itemset'])):
for j in range((i+1), len(prev_freq_itemset['itemset'])):
itemset_i = prev_freq_itemset['itemset'][i]
itemset_j = prev_freq_itemset['itemset'][j]
if(type(itemset_i) == numpy.int64 and type(itemset_j) == numpy.int64):
itemset_i = {itemset_i}
itemset_j = {itemset_j}
union_candidate = itemset_i.union(itemset_j)

if union_candidate not in self_join_candidate:

self_join_candidate.append(union_candidate)
return self_join_candidate

candidate2_list = self_join(freq_itemset1)

candidate2_list

[{1, 2},
{1, 3},
{1, 4},
{1, 5},
{2, 3},
{2, 4},
{2, 5},
{3, 4},
{3, 5},
{4, 5}]

count_candidate2 = []

#Set the Initial value of Second Count Candidate (C2)

for i in range(len(candidate2_list)):
count_candidate2.append((candidate2_list[i], 0))

count_candidate2

[({1, 2}, 0),

({1, 3}, 0),
({1, 4}, 0),
({1, 5}, 0),
({2, 3}, 0),
({2, 4}, 0),
({2, 5}, 0),
({3, 4}, 0),
({3, 5}, 0),
({4, 5}, 0)]

initial_df_candidate = pd.DataFrame(count_candidate2, columns=['itemset', 'sup'])

initial_df_candidate
 itemset sup

0 {1, 2} 0

1 {1, 3} 0

2 {1, 4} 0

3 {1, 5} 0

4 {2, 3} 0

5 {2, 4} 0

6 {2, 5} 0

7 {3, 4} 0

8 {3, 5} 0

9 {4, 5} 0

#Let's see what's inside df

df

items

0 [1, 2]

1 [1, 2, 3]

2 [1, 2, 3, 4]

3 [1, 2, 5]

4 [1, 3, 5]

5 [4, 5]

6 [1, 2, 3, 5]

7 [1, 3]

8 [1, 4]

9 [2, 3, 4]

#Let's add it with 1 whenever we found every candidate is a subset from Database D

def count_support(database_dataframe, prev_candidate_list):

initial_df_candidate['sup'] = 0 #set All value into 0 only for initial value for consistency val
count_prev_candidate = []

#Set the Initial value of Previous Candidate

for i in range(len(prev_candidate_list)):
count_prev_candidate.append((prev_candidate_list[i], 0))

df_candidate = pd.DataFrame(count_prev_candidate, columns=['itemset', 'sup'])

print('Database D dataframe\n', database_dataframe)
print('(Initial) Dataframe from Candidate with All zeros sup\n', df_candidate)
 for i in range(len(database_dataframe)):
for j in range(len(count_prev_candidate)):
#using issubset() function to check whether every itemset is a subset of Database or not
if (df_candidate['itemset'][j]).issubset(set(database_dataframe['items'][i])):
df_candidate.loc[j, 'sup'] += 1

return df_candidate

count_candidate2_df = count_support(df, candidate2_list)

Database D dataframe
items
0 [1, 2]
1 [1, 2, 3]
2 [1, 2, 3, 4]
3 [1, 2, 5]
4 [1, 3, 5]
5 [4, 5]
6 [1, 2, 3, 5]
7 [1, 3]
8 [1, 4]
9 [2, 3, 4]
(Initial) Dataframe from Candidate with All zeros sup
itemset sup
0 {1, 2} 0
1 {1, 3} 0
2 {1, 4} 0
3 {1, 5} 0
4 {2, 3} 0
5 {2, 4} 0
6 {2, 5} 0
7 {3, 4} 0
8 {3, 5} 0
9 {4, 5} 0

count_candidate2_df

itemset sup

0 {1, 2} 5

1 {1, 3} 5

2 {1, 4} 2

3 {1, 5} 3

4 {2, 3} 4

5 {2, 4} 2

6 {2, 5} 2

7 {3, 4} 2

8 {3, 5} 2

9 {4, 5} 1

Double-click (or enter) to edit

Creating Second Frequent Itemset (L2)

#Filter the itemset based on minimum support (occurences of items)

freq_itemset2 = filter_sup(count_candidate2_df)

freq_itemset2

itemset sup

0 {1, 2} 5

1 {1, 3} 5

3 {1, 5} 3

4 {2, 3} 4

freq_itemset2_reset = freq_itemset2.reset_index(drop=True)

#We need to reset the index, because need to access the index later.
freq_itemset2_reset

itemset sup

0 {1, 2} 5

1 {1, 3} 5

2 {1, 5} 3

3 {2, 3} 4

Creating the Third Candidate (C3) - Using the Candidate Forming

Technique
Self join
Pruning

Self Join

print(freq_itemset2_reset)
self_join_result = self_join(freq_itemset2_reset)
print('self join result')
print(self_join_result)

itemset sup
0 {1, 2} 5
1 {1, 3} 5
2 {1, 5} 3
3 {2, 3} 4
 self join result
[{1, 2, 3}, {1, 2, 5}, {1, 3, 5}, {1, 2, 3, 5}]

Pruning

def get_subset(candidate):
temp = []
final = []
for i in range(len(candidate)):
for j in range(len(candidate)):
if i != j:
temp.append(candidate[j])
temp_set = set(temp)
final.append(temp_set)
temp.clear()
print('Subset from {} : {}'.format(candidate, final))
return final

def pruning(candidate_set, prev_freq_itemset):

print('Candidate set', candidate_set)
temp = []

for idx, value in enumerate(candidate_set):

list_candidate = list(value)
temp_candidate = (get_subset(list_candidate))

for temp_item in temp_candidate:

print('Temp item', temp_item)
check = temp_item == prev_freq_itemset['itemset']
print('\nCheck candidate from Previous Frequent Itemset\n', check)

if any(check) == False:
print(any(check))
print('Val', value)
else:
print('\nAll of {} subset contained in \n{}'.format(candidate_set, prev_freq_itemset
if value not in temp:
temp.append(value)

return temp

freq_itemset2_reset

itemset sup

0 {1, 2} 5

1 {1, 3} 5

2 {1, 5} 3

3 {2, 3} 4

subset = [{2, 3}, {1, 3}, {1, 2}]

self_join_result

[{1, 2, 3}, {1, 2, 5}, {1, 3, 5}, {1, 2, 3, 5}]

for i in range(len(self_join_result)):
get_subset(list(self_join_result[i]))

Subset from [1, 2, 3] : [{2, 3}, {1, 3}, {1, 2}]

Subset from [1, 2, 5] : [{2, 5}, {1, 5}, {1, 2}]
Subset from [1, 3, 5] : [{3, 5}, {1, 5}, {1, 3}]
Subset from [1, 2, 3, 5] : [{2, 3, 5}, {1, 3, 5}, {1, 2, 5}, {1, 2, 3}]

freq_itemset2_reset

itemset sup

0 {1, 2} 5

1 {1, 3} 5

2 {1, 5} 3

3 {2, 3} 4

for item in subset:

print(item)
check = item == freq_itemset2_reset['itemset']
print('Check', any(check))

{2, 3}
Check True
{1, 3}
Check True
{1, 2}
Check True

self_join_result

[{1, 2, 3}, {1, 2, 5}, {1, 3, 5}, {1, 2, 3, 5}]

candidate3_list = pruning(self_join_result, freq_itemset2_reset)

2 {1, 5} 3
3 {2, 3} 4
Temp item {1, 3}

Check candidate from Previous Frequent Itemset

0 False
1 True
2 False
3 False
Name: itemset, dtype: bool

All of [{1, 2, 3}, {1, 2, 5}, {1, 3, 5}, {1, 2, 3, 5}] subset contained in
itemset sup
0 {1, 2} 5
1 {1, 3} 5
2 {1, 5} 3
 3 {2, 3} 4
Subset from [1, 2, 3, 5] : [{2, 3, 5}, {1, 3, 5}, {1, 2, 5}, {1, 2, 3}]
Temp item {2, 3, 5}

Check candidate from Previous Frequent Itemset

0 False
1 False
2 False
3 False
Name: itemset, dtype: bool
False
Val {1, 2, 3, 5}
Temp item {1, 3, 5}

Check candidate from Previous Frequent Itemset

0 False
1 False
2 False
3 False
Name: itemset, dtype: bool
False
Val {1, 2, 3, 5}
Temp item {1, 2, 5}

Check candidate from Previous Frequent Itemset

0 False
1 False
2 False
3 False
Name: itemset, dtype: bool
False
Val {1, 2, 3, 5}
Temp item {1, 2, 3}

Check candidate from Previous Frequent Itemset

0 False
1 False
2 False
3 False
Name: itemset, dtype: bool
False
Val {1, 2, 3, 5}

candidate3_list

[{1, 2, 3}, {1, 2, 5}, {1, 3, 5}]

Creating the Third Frequent Itemset (L3)

#Let's see the database again

df
 items

0 [1, 2]

1 [1, 2, 3]

2 [1, 2, 3, 4]

3 [1, 2, 5]

4 [1, 3, 5]

5 [4, 5]

#Then 6 [1, 2,
check 3, 5]
the newest candidate value
candidate3_list
7 [1, 3]
[{1,
8 2, [1,
3},4]{1, 2, 5}, {1, 3, 5}]

9 [2, 3, 4]
count_candidate3_df = count_support(df, candidate3_list)

Database D dataframe
items
0 [1, 2]
1 [1, 2, 3]
2 [1, 2, 3, 4]
3 [1, 2, 5]
4 [1, 3, 5]
5 [4, 5]
6 [1, 2, 3, 5]
7 [1, 3]
8 [1, 4]
9 [2, 3, 4]
(Initial) Dataframe from Candidate with All zeros sup
itemset sup
0 {1, 2, 3} 0
1 {1, 2, 5} 0
2 {1, 3, 5} 0

count_candidate3_df

itemset sup

0 {1, 2, 3} 3

1 {1, 2, 5} 2

2 {1, 3, 5} 2

freq_itemset3 = filter_sup(count_candidate3_df)

freq_itemset3

itemset sup

0 {1, 2, 3} 3
All Frequent Itemset
#Let'see each frequent itemset (L)
freq_itemset1

itemset sup

0 1 8

1 2 6

2 3 6

3 4 4

4 5 4

freq_itemset2

itemset sup

0 {1, 2} 5

1 {1, 3} 5

3 {1, 5} 3

4 {2, 3} 4

freq_itemset3

itemset sup

0 {1, 2, 3} 3

frequent_itemset = pd.concat([freq_itemset1, freq_itemset2, freq_itemset3], axis=0)

frequent_itemset

itemset sup

0 1 8

1 2 6

2 3 6

3 4 4

4 5 4

0 {1, 2} 5

1 {1, 3} 5

3 {1, 5} 3

4 {2, 3} 4

0 {1, 2, 3} 3
#Reset the index
frequent_itemset_final = frequent_itemset.reset_index(drop=True)

Final Output of Freq. Itemset (L1-L3)

frequent_itemset_final

itemset sup

0 1 8

1 2 6

2 3 6

3 4 4

4 5 4

5 {1, 2} 5

6 {1, 3} 5

7 {1, 5} 3

8 {2, 3} 4

9 {1, 2, 3} 3

Conclusion : Thus, we have successfully learned and implemented Apriori Algorithm using Python from
Scratch.

Colab paid products - Cancel contracts here

check 0s completed at 13:55