From 0ea8df7682158f82fb463292b08201ef6680993e Mon Sep 17 00:00:00 2001
From: Jeel Gajera <jeelgajera00@gmail.com>
Date: Sat, 21 Oct 2023 11:20:06 +0530
Subject: [PATCH 1/8] Add: FP Growth Algorithm
---
DIRECTORY.md | 1 +
machine_learning/fp_growth.py | 333 ++++++++++++++++++++++++++++++++++
2 files changed, 334 insertions(+)
create mode 100644 machine_learning/fp_growth.py
diff --git a/DIRECTORY.md b/DIRECTORY.md
index b92f8f877e97..df6692fdddae 100644
--- a/DIRECTORY.md
+++ b/DIRECTORY.md
@@ -541,6 +541,7 @@
* [Dimensionality Reduction](machine_learning/dimensionality_reduction.py)
* Forecasting
* [Run](machine_learning/forecasting/run.py)
+ * [FP Growth Algorithm](machine_learning/fp_growth.py)
* [Gradient Descent](machine_learning/gradient_descent.py)
* [K Means Clust](machine_learning/k_means_clust.py)
* [K Nearest Neighbours](machine_learning/k_nearest_neighbours.py)
diff --git a/machine_learning/fp_growth.py b/machine_learning/fp_growth.py
new file mode 100644
index 000000000000..eaf59693f231
--- /dev/null
+++ b/machine_learning/fp_growth.py
@@ -0,0 +1,333 @@
+"""
+The FP-Growth (Frequent Pattern Growth) algorithm is a widely used
+data mining technique for discovering frequent itemsets in
+large transaction databases.
+It overcomes some of the limitations of traditional methods like
+Apriori by efficiently constructing the FP-Tree
+
+WIKI: https://athena.ecs.csus.edu/~mei/associationcw/FpGrowth.html
+Examples: https://www.javatpoint.com/fp-growth-algorithm-in-data-mining
+"""
+
+from typing import Optional
+
+
+class TreeNode:
+ """
+ Initialize a TreeNode.
+
+ Args:
+ name_value (str): The name of the node.
+ num_occur (int): The number of occurrences of the node.
+ parent_node (TreeNode): The parent node.
+
+ Example:
+ >>> parent = TreeNode("Parent", 1, None)
+ >>> child = TreeNode("Child", 2, parent)
+ >>> child.name
+ 'Child'
+ >>> child.count
+ 2
+ """
+
+ def __init__(
+ self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
+ ) -> None:
+ self.name = name_value
+ self.count = num_occur
+ self.node_link = None # Initialize node_link to None
+ self.parent = parent_node
+ self.children: dict[str, TreeNode] = {}
+
+ def inc(self, num_occur: int) -> None:
+ self.count += num_occur
+
+ def disp(self, ind: int = 1) -> None:
+ print(" " * ind, self.name, " ", self.count)
+ for child in self.children.values():
+ child.disp(ind + 1)
+
+
+def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
+ """
+ Create FP tree
+
+ Args:
+ data_set (list): A list of transactions, where each transaction
+ is a list of items.
+ min_sup (int, optional): The minimum support threshold.
+ Items with support less than this will be pruned. Default is 1.
+
+ Returns:
+ TreeNode: The root of the FP-Tree.
+ dict: The header table.
+
+ Example:
+ >>> data_set = [
+ ... ['A', 'B', 'C'],
+ ... ['A', 'C'],
+ ... ['A', 'B', 'E'],
+ ... ['A', 'B', 'C', 'E'],
+ ... ['B', 'E']
+ ... ]
+ >>> min_sup = 2
+ >>> fp_tree, header_table = create_tree(data_set, min_sup)
+
+ >>> sorted(list(header_table.keys()))
+ ['A', 'B', 'C', 'E']
+
+ >>> fp_tree.name
+ 'Null Set'
+ >>> sorted(fp_tree.children.keys())
+ ['A', 'B']
+ >>> fp_tree.children['A'].name
+ 'A'
+ >>> sorted(fp_tree.children['A'].children.keys())
+ ['B', 'C']
+
+ """
+ header_table: dict = {}
+ for trans in data_set:
+ for item in trans:
+ header_table[item] = header_table.get(item, [0, None])
+ header_table[item][0] += 1
+
+ for k in list(header_table.keys()):
+ if header_table[k][0] < min_sup:
+ del header_table[k]
+
+ freq_item_set = set(header_table.keys())
+
+ if len(freq_item_set) == 0:
+ return TreeNode("Null Set", 1, None), {}
+
+ for k in header_table:
+ header_table[k] = [header_table[k], None]
+
+ fp_tree = TreeNode("Null Set", 1, None) # Parent is None for the root node
+ for tran_set in data_set:
+ local_d = {}
+ for item in tran_set:
+ if item in freq_item_set:
+ local_d[item] = header_table[item][0]
+ if len(local_d) > 0:
+ sorted_items = sorted(
+ local_d.items(), key=lambda item_info: item_info[1], reverse=True
+ )
+ ordered_items = [item[0] for item in sorted_items]
+ update_tree(ordered_items, fp_tree, header_table, 1)
+
+ return fp_tree, header_table
+
+
+def update_tree(items: list, in_tree: TreeNode, header_table: dict, count: int) -> None:
+ """
+ Update the FP-Tree with a transaction.
+
+ Args:
+ items (list): List of items in the transaction.
+ in_tree (TreeNode): The current node in the FP-Tree.
+ header_table (dict): The header table with item information.
+ count (int): The count of the transaction.
+
+ Example:
+ >>> data_set = [
+ ... ['A', 'B', 'C'],
+ ... ['A', 'C'],
+ ... ['A', 'B', 'E'],
+ ... ['A', 'B', 'C', 'E'],
+ ... ['B', 'E']
+ ... ]
+ >>> min_sup = 2
+ >>> fp_tree, header_table = create_tree(data_set, min_sup)
+
+ >>> transaction = ['A', 'B', 'E']
+ >>> update_tree(transaction, fp_tree, header_table, 1)
+
+ >>> sorted(fp_tree.children['A'].children['B'].children['E'].children.keys())
+ []
+ >>> fp_tree.children['A'].children['B'].children['E'].count
+ 2
+ >>> header_table['E'][1].name
+ 'E'
+ """
+ if items[0] in in_tree.children:
+ in_tree.children[items[0]].inc(count)
+ else:
+ in_tree.children[items[0]] = TreeNode(items[0], count, in_tree)
+ if header_table[items[0]][1] is None:
+ header_table[items[0]][1] = in_tree.children[items[0]]
+ else:
+ update_header(header_table[items[0]][1], in_tree.children[items[0]])
+ if len(items) > 1:
+ update_tree(items[1:], in_tree.children[items[0]], header_table, count)
+
+
+def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
+ """
+ Update the header table with a node link.
+
+ Args:
+ node_to_test (TreeNode): The node to be updated in the header table.
+ target_node (TreeNode): The node to link to.
+
+ Example:
+ >>> data_set = [
+ ... ['A', 'B', 'C'],
+ ... ['A', 'C'],
+ ... ['A', 'B', 'E'],
+ ... ['A', 'B', 'C', 'E'],
+ ... ['B', 'E']
+ ... ]
+ >>> min_sup = 2
+ >>> fp_tree, header_table = create_tree(data_set, min_sup)
+
+ >>> node1 = TreeNode("A", 3, None)
+ >>> node2 = TreeNode("B", 4, None)
+ >>> node1 = update_header(node1, node2)
+ >>> node1.node_link.name
+ 'B'
+ >>> node2.node_link is None
+ True
+ """
+ while node_to_test.node_link is not None:
+ node_to_test = node_to_test.node_link
+ if node_to_test.node_link is None:
+ node_to_test.node_link = TreeNode(target_node.name, target_node.count, node_to_test)
+ # Return the updated node
+ return node_to_test
+
+
+def ascend_tree(leaf_node: TreeNode, prefix_path: list) -> None:
+ """
+ Ascend the FP-Tree from a leaf node to its root,
+ adding item names to the prefix path.
+
+ Args:
+ leaf_node (TreeNode): The leaf node to start ascending from.
+ prefix_path (list): A list to store the item as they are ascended.
+
+ Example:
+ >>> data_set = [
+ ... ['A', 'B', 'C'],
+ ... ['A', 'C'],
+ ... ['A', 'B', 'E'],
+ ... ['A', 'B', 'C', 'E'],
+ ... ['B', 'E']
+ ... ]
+ >>> min_sup = 2
+ >>> fp_tree, header_table = create_tree(data_set, min_sup)
+
+ >>> path = []
+ >>> ascend_tree(fp_tree.children['A'], path)
+ >>> path # ascending from a leaf node 'A'
+ ['A']
+ """
+ if leaf_node.parent is not None:
+ prefix_path.append(leaf_node.name)
+ ascend_tree(leaf_node.parent, prefix_path)
+
+
+def find_prefix_path(base_pat: frozenset, tree_node: TreeNode | None) -> dict:
+ """
+ Find the conditional pattern base for a given base pattern.
+
+ Args:
+ base_pat (frozenset): The base pattern for which to find
+ the conditional pattern base.
+ tree_node (TreeNode): The node in the FP-Tree.
+
+ Example:
+ >>> data_set = [
+ ... ['A', 'B', 'C'],
+ ... ['A', 'C'],
+ ... ['A', 'B', 'E'],
+ ... ['A', 'B', 'C', 'E'],
+ ... ['B', 'E']
+ ... ]
+ >>> min_sup = 2
+ >>> fp_tree, header_table = create_tree(data_set, min_sup)
+ >>> base_pattern = frozenset(['A'])
+ >>> cond_pat = find_prefix_path(base_pattern, fp_tree.children['A'])
+ >>> sorted(cond_pat.keys())
+ []
+ """
+ cond_pats: dict = {}
+ while tree_node is not None:
+ prefix_path: list = []
+ ascend_tree(tree_node, prefix_path)
+ if len(prefix_path) > 1:
+ cond_pats[frozenset(prefix_path[1:])] = tree_node.count
+ tree_node = tree_node.node_link
+ return cond_pats
+
+
+def mine_tree(
+ in_tree: TreeNode,
+ header_table: dict,
+ min_sup: int,
+ pre_fix: set,
+ freq_item_list: list,
+) -> None:
+ """
+ Mine the FP-Tree recursively to discover frequent itemsets.
+
+ Args:
+ in_tree (TreeNode): The FP-Tree to mine.
+ header_table (dict): The header table with item information.
+ min_sup (int): The minimum support threshold.
+ pre_fix (set): A set of items as a prefix for the itemsets being mined.
+ freq_item_list (list): A list to store the frequent itemsets.
+
+ Example:
+ >>> data_set = [
+ ... ['A', 'B', 'C'],
+ ... ['A', 'C'],
+ ... ['A', 'B', 'E'],
+ ... ['A', 'B', 'C', 'E'],
+ ... ['B', 'E']
+ ... ]
+ >>> min_sup = 2
+ >>> fp_tree, header_table = create_tree(data_set, min_sup)
+
+ >>> frequent_itemsets = []
+ >>> mine_tree(fp_tree, header_table, min_sup, set([]), frequent_itemsets)
+ >>> expe_itm = [{'C'}, {'C', 'A'}, {'E'}, {'A', 'E'}, {'E', 'B'}, {'A'}, {'B'}]
+ >>> all(expected in frequent_itemsets for expected in expe_itm)
+ True
+ """
+ sorted_items = sorted(header_table.items(), key=lambda item_info: item_info[1][0])
+ big_l = [item[0] for item in sorted_items]
+ for base_pat in big_l:
+ new_freq_set = pre_fix.copy()
+ new_freq_set.add(base_pat)
+ freq_item_list.append(new_freq_set)
+ cond_patt_bases = find_prefix_path(base_pat, header_table[base_pat][1])
+ my_cond_tree, my_head = create_tree(list(cond_patt_bases.keys()), min_sup)
+ if my_head is not None:
+ # Pass header_table[base_pat][1] as node_to_test to update_header
+ header_table[base_pat][1] = update_header(
+ header_table[base_pat][1], my_cond_tree
+ )
+ mine_tree(my_cond_tree, my_head, min_sup, new_freq_set, freq_item_list)
+
+
+if __name__ == "__main__":
+ import doctest
+
+ doctest.testmod()
+
+ data_set: list = [
+ frozenset(["bread", "milk", "cheese"]),
+ frozenset(["bread", "milk"]),
+ frozenset(["bread", "diapers"]),
+ frozenset(["bread", "milk", "diapers"]),
+ frozenset(["milk", "diapers"]),
+ frozenset(["milk", "cheese"]),
+ frozenset(["diapers", "cheese"]),
+ frozenset(["bread", "milk", "cheese", "diapers"]),
+ ]
+ fp_tree, header_table = create_tree(data_set, min_sup=3)
+ freq_items: list = []
+ mine_tree(fp_tree, header_table, 3, set(), freq_items)
+ print(freq_items)
From 71776e78022d6a9699f2ced68bd28d6667e7335e Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
<66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Sat, 21 Oct 2023 05:52:58 +0000
Subject: [PATCH 2/8] [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---
machine_learning/fp_growth.py | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/machine_learning/fp_growth.py b/machine_learning/fp_growth.py
index eaf59693f231..356e95347079 100644
--- a/machine_learning/fp_growth.py
+++ b/machine_learning/fp_growth.py
@@ -193,7 +193,9 @@ def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
while node_to_test.node_link is not None:
node_to_test = node_to_test.node_link
if node_to_test.node_link is None:
- node_to_test.node_link = TreeNode(target_node.name, target_node.count, node_to_test)
+ node_to_test.node_link = TreeNode(
+ target_node.name, target_node.count, node_to_test
+ )
# Return the updated node
return node_to_test
From c0470094d01391294617df6a92734b78b470b127 Mon Sep 17 00:00:00 2001
From: Jeel Gajera <jeelgajera00@gmail.com>
Date: Sat, 21 Oct 2023 14:48:21 +0530
Subject: [PATCH 3/8] changes names
---
DIRECTORY.md | 2 +-
...p_growth.py => frequent_pattern_growth.py} | 32 +++++++++++--------
2 files changed, 20 insertions(+), 14 deletions(-)
rename machine_learning/{fp_growth.py => frequent_pattern_growth.py} (92%)
diff --git a/DIRECTORY.md b/DIRECTORY.md
index df6692fdddae..916d993c563a 100644
--- a/DIRECTORY.md
+++ b/DIRECTORY.md
@@ -541,7 +541,7 @@
* [Dimensionality Reduction](machine_learning/dimensionality_reduction.py)
* Forecasting
* [Run](machine_learning/forecasting/run.py)
- * [FP Growth Algorithm](machine_learning/fp_growth.py)
+ * [Frequent Pattern Growth Algorithm](machine_learning/frequent_pattern_growth.py)
* [Gradient Descent](machine_learning/gradient_descent.py)
* [K Means Clust](machine_learning/k_means_clust.py)
* [K Nearest Neighbours](machine_learning/k_nearest_neighbours.py)
diff --git a/machine_learning/fp_growth.py b/machine_learning/frequent_pattern_growth.py
similarity index 92%
rename from machine_learning/fp_growth.py
rename to machine_learning/frequent_pattern_growth.py
index 356e95347079..fa37313be9da 100644
--- a/machine_learning/fp_growth.py
+++ b/machine_learning/frequent_pattern_growth.py
@@ -10,8 +10,10 @@
"""
from typing import Optional
+from dataclasses import dataclass, field
+@dataclass
class TreeNode:
"""
Initialize a TreeNode.
@@ -30,14 +32,19 @@ class TreeNode:
2
"""
- def __init__(
- self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
- ) -> None:
- self.name = name_value
- self.count = num_occur
- self.node_link = None # Initialize node_link to None
- self.parent = parent_node
- self.children: dict[str, TreeNode] = {}
+ # def __init__(
+ # self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
+ # ) -> None:
+ # self.name = name_value
+ # self.count = num_occur
+ # self.node_link = TreeNode | None # Initialize node_link to None
+ # self.parent = parent_node
+ # self.children: dict[str, TreeNode] = {}
+ name: str
+ count: int
+ node_link: Optional['TreeNode'] = None # Initialize node_link to None
+ parent: Optional["TreeNode"] = None
+ children: dict[str, "TreeNode"] = field(default_factory=dict)
def inc(self, num_occur: int) -> None:
self.count += num_occur
@@ -50,7 +57,7 @@ def disp(self, ind: int = 1) -> None:
def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
"""
- Create FP tree
+ Create Frequent Pattern tree
Args:
data_set (list): A list of transactions, where each transaction
@@ -193,10 +200,7 @@ def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
while node_to_test.node_link is not None:
node_to_test = node_to_test.node_link
if node_to_test.node_link is None:
- node_to_test.node_link = TreeNode(
- target_node.name, target_node.count, node_to_test
- )
- # Return the updated node
+ node_to_test.node_link = target_node
return node_to_test
@@ -298,6 +302,7 @@ def mine_tree(
>>> all(expected in frequent_itemsets for expected in expe_itm)
True
"""
+ new_head: Optional['TreeNode'] = None
sorted_items = sorted(header_table.items(), key=lambda item_info: item_info[1][0])
big_l = [item[0] for item in sorted_items]
for base_pat in big_l:
@@ -311,6 +316,7 @@ def mine_tree(
header_table[base_pat][1] = update_header(
header_table[base_pat][1], my_cond_tree
)
+ my_head = new_head
mine_tree(my_cond_tree, my_head, min_sup, new_freq_set, freq_item_list)
From c014b1c006e1a5a7e3ba41856ad736e843df8c34 Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
<66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Sat, 21 Oct 2023 09:19:07 +0000
Subject: [PATCH 4/8] [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---
machine_learning/frequent_pattern_growth.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/machine_learning/frequent_pattern_growth.py b/machine_learning/frequent_pattern_growth.py
index fa37313be9da..3d5d6609fc53 100644
--- a/machine_learning/frequent_pattern_growth.py
+++ b/machine_learning/frequent_pattern_growth.py
@@ -42,7 +42,7 @@ class TreeNode:
# self.children: dict[str, TreeNode] = {}
name: str
count: int
- node_link: Optional['TreeNode'] = None # Initialize node_link to None
+ node_link: Optional["TreeNode"] = None # Initialize node_link to None
parent: Optional["TreeNode"] = None
children: dict[str, "TreeNode"] = field(default_factory=dict)
@@ -302,7 +302,7 @@ def mine_tree(
>>> all(expected in frequent_itemsets for expected in expe_itm)
True
"""
- new_head: Optional['TreeNode'] = None
+ new_head: Optional["TreeNode"] = None
sorted_items = sorted(header_table.items(), key=lambda item_info: item_info[1][0])
big_l = [item[0] for item in sorted_items]
for base_pat in big_l:
From 3ae692d129e5b60df7baafbd09f2005a4dc305ef Mon Sep 17 00:00:00 2001
From: Jeel Gajera <jeelgajera00@gmail.com>
Date: Sat, 21 Oct 2023 14:53:49 +0530
Subject: [PATCH 5/8] Revert "changes names"
This reverts commit c0470094d01391294617df6a92734b78b470b127.
---
DIRECTORY.md | 2 +-
...requent_pattern_growth.py => fp_growth.py} | 32 ++++++++-----------
2 files changed, 14 insertions(+), 20 deletions(-)
rename machine_learning/{frequent_pattern_growth.py => fp_growth.py} (92%)
diff --git a/DIRECTORY.md b/DIRECTORY.md
index 916d993c563a..df6692fdddae 100644
--- a/DIRECTORY.md
+++ b/DIRECTORY.md
@@ -541,7 +541,7 @@
* [Dimensionality Reduction](machine_learning/dimensionality_reduction.py)
* Forecasting
* [Run](machine_learning/forecasting/run.py)
- * [Frequent Pattern Growth Algorithm](machine_learning/frequent_pattern_growth.py)
+ * [FP Growth Algorithm](machine_learning/fp_growth.py)
* [Gradient Descent](machine_learning/gradient_descent.py)
* [K Means Clust](machine_learning/k_means_clust.py)
* [K Nearest Neighbours](machine_learning/k_nearest_neighbours.py)
diff --git a/machine_learning/frequent_pattern_growth.py b/machine_learning/fp_growth.py
similarity index 92%
rename from machine_learning/frequent_pattern_growth.py
rename to machine_learning/fp_growth.py
index fa37313be9da..356e95347079 100644
--- a/machine_learning/frequent_pattern_growth.py
+++ b/machine_learning/fp_growth.py
@@ -10,10 +10,8 @@
"""
from typing import Optional
-from dataclasses import dataclass, field
-@dataclass
class TreeNode:
"""
Initialize a TreeNode.
@@ -32,19 +30,14 @@ class TreeNode:
2
"""
- # def __init__(
- # self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
- # ) -> None:
- # self.name = name_value
- # self.count = num_occur
- # self.node_link = TreeNode | None # Initialize node_link to None
- # self.parent = parent_node
- # self.children: dict[str, TreeNode] = {}
- name: str
- count: int
- node_link: Optional['TreeNode'] = None # Initialize node_link to None
- parent: Optional["TreeNode"] = None
- children: dict[str, "TreeNode"] = field(default_factory=dict)
+ def __init__(
+ self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
+ ) -> None:
+ self.name = name_value
+ self.count = num_occur
+ self.node_link = None # Initialize node_link to None
+ self.parent = parent_node
+ self.children: dict[str, TreeNode] = {}
def inc(self, num_occur: int) -> None:
self.count += num_occur
@@ -57,7 +50,7 @@ def disp(self, ind: int = 1) -> None:
def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
"""
- Create Frequent Pattern tree
+ Create FP tree
Args:
data_set (list): A list of transactions, where each transaction
@@ -200,7 +193,10 @@ def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
while node_to_test.node_link is not None:
node_to_test = node_to_test.node_link
if node_to_test.node_link is None:
- node_to_test.node_link = target_node
+ node_to_test.node_link = TreeNode(
+ target_node.name, target_node.count, node_to_test
+ )
+ # Return the updated node
return node_to_test
@@ -302,7 +298,6 @@ def mine_tree(
>>> all(expected in frequent_itemsets for expected in expe_itm)
True
"""
- new_head: Optional['TreeNode'] = None
sorted_items = sorted(header_table.items(), key=lambda item_info: item_info[1][0])
big_l = [item[0] for item in sorted_items]
for base_pat in big_l:
@@ -316,7 +311,6 @@ def mine_tree(
header_table[base_pat][1] = update_header(
header_table[base_pat][1], my_cond_tree
)
- my_head = new_head
mine_tree(my_cond_tree, my_head, min_sup, new_freq_set, freq_item_list)
From 973ae0200870f88ea0b2b17e40abc0890582df57 Mon Sep 17 00:00:00 2001
From: Jeel Gajera <jeelgajera00@gmail.com>
Date: Sat, 21 Oct 2023 15:59:24 +0530
Subject: [PATCH 6/8] refactore code
---
DIRECTORY.md | 2 +-
...p_growth.py => frequent_pattern_growth.py} | 23 +++++++++++--------
2 files changed, 15 insertions(+), 10 deletions(-)
rename machine_learning/{fp_growth.py => frequent_pattern_growth.py} (95%)
diff --git a/DIRECTORY.md b/DIRECTORY.md
index df6692fdddae..916d993c563a 100644
--- a/DIRECTORY.md
+++ b/DIRECTORY.md
@@ -541,7 +541,7 @@
* [Dimensionality Reduction](machine_learning/dimensionality_reduction.py)
* Forecasting
* [Run](machine_learning/forecasting/run.py)
- * [FP Growth Algorithm](machine_learning/fp_growth.py)
+ * [Frequent Pattern Growth Algorithm](machine_learning/frequent_pattern_growth.py)
* [Gradient Descent](machine_learning/gradient_descent.py)
* [K Means Clust](machine_learning/k_means_clust.py)
* [K Nearest Neighbours](machine_learning/k_nearest_neighbours.py)
diff --git a/machine_learning/fp_growth.py b/machine_learning/frequent_pattern_growth.py
similarity index 95%
rename from machine_learning/fp_growth.py
rename to machine_learning/frequent_pattern_growth.py
index 356e95347079..df37eeee2f08 100644
--- a/machine_learning/fp_growth.py
+++ b/machine_learning/frequent_pattern_growth.py
@@ -9,9 +9,10 @@
Examples: https://www.javatpoint.com/fp-growth-algorithm-in-data-mining
"""
+# from dataclasses import dataclass, field
from typing import Optional
-
+# @dataclass
class TreeNode:
"""
Initialize a TreeNode.
@@ -31,7 +32,8 @@ class TreeNode:
"""
def __init__(
- self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
+ self, name_value: str, num_occur: int,
+ parent_node: Optional["TreeNode"] = None
) -> None:
self.name = name_value
self.count = num_occur
@@ -39,6 +41,13 @@ def __init__(
self.parent = parent_node
self.children: dict[str, TreeNode] = {}
+ # name: str
+ # count: int
+ # node_link: Optional["TreeNode"] = None
+ # parent: Optional["TreeNode"] = None
+ # children: dict[str, "TreeNode"] = field(default_factory=dict)
+
+
def inc(self, num_occur: int) -> None:
self.count += num_occur
@@ -50,7 +59,7 @@ def disp(self, ind: int = 1) -> None:
def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
"""
- Create FP tree
+ Create Frequent Pattern tree
Args:
data_set (list): A list of transactions, where each transaction
@@ -96,9 +105,7 @@ def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
if header_table[k][0] < min_sup:
del header_table[k]
- freq_item_set = set(header_table.keys())
-
- if len(freq_item_set) == 0:
+ if not (freq_item_set := set(header_table)):
return TreeNode("Null Set", 1, None), {}
for k in header_table:
@@ -193,9 +200,7 @@ def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
while node_to_test.node_link is not None:
node_to_test = node_to_test.node_link
if node_to_test.node_link is None:
- node_to_test.node_link = TreeNode(
- target_node.name, target_node.count, node_to_test
- )
+ node_to_test.node_link = target_node
# Return the updated node
return node_to_test
From 8a1f71b812b678097cb6e962c01aa04cd6c53ee4 Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
<66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Sat, 21 Oct 2023 10:30:26 +0000
Subject: [PATCH 7/8] [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---
machine_learning/frequent_pattern_growth.py | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/machine_learning/frequent_pattern_growth.py b/machine_learning/frequent_pattern_growth.py
index df37eeee2f08..b58ef5d45746 100644
--- a/machine_learning/frequent_pattern_growth.py
+++ b/machine_learning/frequent_pattern_growth.py
@@ -12,6 +12,7 @@
# from dataclasses import dataclass, field
from typing import Optional
+
# @dataclass
class TreeNode:
"""
@@ -32,8 +33,7 @@ class TreeNode:
"""
def __init__(
- self, name_value: str, num_occur: int,
- parent_node: Optional["TreeNode"] = None
+ self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
) -> None:
self.name = name_value
self.count = num_occur
@@ -47,7 +47,6 @@ def __init__(
# parent: Optional["TreeNode"] = None
# children: dict[str, "TreeNode"] = field(default_factory=dict)
-
def inc(self, num_occur: int) -> None:
self.count += num_occur
From d4ded62a0d217749b430ebac926ba1a4dc1a81ec Mon Sep 17 00:00:00 2001
From: Christian Clauss <cclauss@me.com>
Date: Sat, 21 Oct 2023 16:47:43 +0200
Subject: [PATCH 8/8] Update frequent_pattern_growth.py
---
machine_learning/frequent_pattern_growth.py | 168 +++++++++++---------
1 file changed, 89 insertions(+), 79 deletions(-)
diff --git a/machine_learning/frequent_pattern_growth.py b/machine_learning/frequent_pattern_growth.py
index b58ef5d45746..205d598464a1 100644
--- a/machine_learning/frequent_pattern_growth.py
+++ b/machine_learning/frequent_pattern_growth.py
@@ -1,27 +1,28 @@
"""
-The FP-Growth (Frequent Pattern Growth) algorithm is a widely used
-data mining technique for discovering frequent itemsets in
-large transaction databases.
-It overcomes some of the limitations of traditional methods like
-Apriori by efficiently constructing the FP-Tree
+The Frequent Pattern Growth algorithm (FP-Growth) is a widely used data mining
+technique for discovering frequent itemsets in large transaction databases.
+
+It overcomes some of the limitations of traditional methods such as Apriori by
+efficiently constructing the FP-Tree
WIKI: https://athena.ecs.csus.edu/~mei/associationcw/FpGrowth.html
+
Examples: https://www.javatpoint.com/fp-growth-algorithm-in-data-mining
"""
+from __future__ import annotations
-# from dataclasses import dataclass, field
-from typing import Optional
+from dataclasses import dataclass, field
-# @dataclass
+@dataclass
class TreeNode:
"""
- Initialize a TreeNode.
+ A node in a Frequent Pattern tree.
Args:
- name_value (str): The name of the node.
- num_occur (int): The number of occurrences of the node.
- parent_node (TreeNode): The parent node.
+ name: The name of this node.
+ num_occur: The number of occurrences of the node.
+ parent_node: The parent node.
Example:
>>> parent = TreeNode("Parent", 1, None)
@@ -32,26 +33,20 @@ class TreeNode:
2
"""
- def __init__(
- self, name_value: str, num_occur: int, parent_node: Optional["TreeNode"] = None
- ) -> None:
- self.name = name_value
- self.count = num_occur
- self.node_link = None # Initialize node_link to None
- self.parent = parent_node
- self.children: dict[str, TreeNode] = {}
-
- # name: str
- # count: int
- # node_link: Optional["TreeNode"] = None
- # parent: Optional["TreeNode"] = None
- # children: dict[str, "TreeNode"] = field(default_factory=dict)
+ name: str
+ count: int
+ parent: TreeNode | None = None
+ children: dict[str, TreeNode] = field(default_factory=dict)
+ node_link: TreeNode | None = None
+
+ def __repr__(self) -> str:
+ return f"TreeNode({self.name!r}, {self.count!r}, {self.parent!r})"
def inc(self, num_occur: int) -> None:
self.count += num_occur
def disp(self, ind: int = 1) -> None:
- print(" " * ind, self.name, " ", self.count)
+ print(f"{' ' * ind} {self.name} {self.count}")
for child in self.children.values():
child.disp(ind + 1)
@@ -61,14 +56,13 @@ def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
Create Frequent Pattern tree
Args:
- data_set (list): A list of transactions, where each transaction
- is a list of items.
- min_sup (int, optional): The minimum support threshold.
+ data_set: A list of transactions, where each transaction is a list of items.
+ min_sup: The minimum support threshold.
Items with support less than this will be pruned. Default is 1.
Returns:
- TreeNode: The root of the FP-Tree.
- dict: The header table.
+ The root of the FP-Tree.
+ header_table: The header table dictionary with item information.
Example:
>>> data_set = [
@@ -80,19 +74,24 @@ def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
... ]
>>> min_sup = 2
>>> fp_tree, header_table = create_tree(data_set, min_sup)
-
- >>> sorted(list(header_table.keys()))
+ >>> fp_tree
+ TreeNode('Null Set', 1, None)
+ >>> len(header_table)
+ 4
+ >>> header_table["A"]
+ [[4, None], TreeNode('A', 4, TreeNode('Null Set', 1, None))]
+ >>> header_table["E"][1] # doctest: +NORMALIZE_WHITESPACE
+ TreeNode('E', 1, TreeNode('B', 3, TreeNode('A', 4, TreeNode('Null Set', 1, None))))
+ >>> sorted(header_table)
['A', 'B', 'C', 'E']
-
>>> fp_tree.name
'Null Set'
- >>> sorted(fp_tree.children.keys())
+ >>> sorted(fp_tree.children)
['A', 'B']
>>> fp_tree.children['A'].name
'A'
- >>> sorted(fp_tree.children['A'].children.keys())
+ >>> sorted(fp_tree.children['A'].children)
['B', 'C']
-
"""
header_table: dict = {}
for trans in data_set:
@@ -100,7 +99,7 @@ def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
header_table[item] = header_table.get(item, [0, None])
header_table[item][0] += 1
- for k in list(header_table.keys()):
+ for k in list(header_table):
if header_table[k][0] < min_sup:
del header_table[k]
@@ -112,11 +111,10 @@ def create_tree(data_set: list, min_sup: int = 1) -> tuple[TreeNode, dict]:
fp_tree = TreeNode("Null Set", 1, None) # Parent is None for the root node
for tran_set in data_set:
- local_d = {}
- for item in tran_set:
- if item in freq_item_set:
- local_d[item] = header_table[item][0]
- if len(local_d) > 0:
+ local_d = {
+ item: header_table[item][0] for item in tran_set if item in freq_item_set
+ }
+ if local_d:
sorted_items = sorted(
local_d.items(), key=lambda item_info: item_info[1], reverse=True
)
@@ -131,10 +129,10 @@ def update_tree(items: list, in_tree: TreeNode, header_table: dict, count: int)
Update the FP-Tree with a transaction.
Args:
- items (list): List of items in the transaction.
- in_tree (TreeNode): The current node in the FP-Tree.
- header_table (dict): The header table with item information.
- count (int): The count of the transaction.
+ items: List of items in the transaction.
+ in_tree: The current node in the FP-Tree.
+ header_table: The header table dictionary with item information.
+ count: The count of the transaction.
Example:
>>> data_set = [
@@ -146,12 +144,14 @@ def update_tree(items: list, in_tree: TreeNode, header_table: dict, count: int)
... ]
>>> min_sup = 2
>>> fp_tree, header_table = create_tree(data_set, min_sup)
-
+ >>> fp_tree
+ TreeNode('Null Set', 1, None)
>>> transaction = ['A', 'B', 'E']
>>> update_tree(transaction, fp_tree, header_table, 1)
-
- >>> sorted(fp_tree.children['A'].children['B'].children['E'].children.keys())
- []
+ >>> fp_tree
+ TreeNode('Null Set', 1, None)
+ >>> fp_tree.children['A'].children['B'].children['E'].children
+ {}
>>> fp_tree.children['A'].children['B'].children['E'].count
2
>>> header_table['E'][1].name
@@ -174,8 +174,8 @@ def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
Update the header table with a node link.
Args:
- node_to_test (TreeNode): The node to be updated in the header table.
- target_node (TreeNode): The node to link to.
+ node_to_test: The node to be updated in the header table.
+ target_node: The node to link to.
Example:
>>> data_set = [
@@ -187,12 +187,17 @@ def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
... ]
>>> min_sup = 2
>>> fp_tree, header_table = create_tree(data_set, min_sup)
-
+ >>> fp_tree
+ TreeNode('Null Set', 1, None)
>>> node1 = TreeNode("A", 3, None)
>>> node2 = TreeNode("B", 4, None)
+ >>> node1
+ TreeNode('A', 3, None)
>>> node1 = update_header(node1, node2)
- >>> node1.node_link.name
- 'B'
+ >>> node1
+ TreeNode('A', 3, None)
+ >>> node1.node_link
+ TreeNode('B', 4, None)
>>> node2.node_link is None
True
"""
@@ -204,14 +209,14 @@ def update_header(node_to_test: TreeNode, target_node: TreeNode) -> TreeNode:
return node_to_test
-def ascend_tree(leaf_node: TreeNode, prefix_path: list) -> None:
+def ascend_tree(leaf_node: TreeNode, prefix_path: list[str]) -> None:
"""
- Ascend the FP-Tree from a leaf node to its root,
- adding item names to the prefix path.
+ Ascend the FP-Tree from a leaf node to its root, adding item names to the prefix
+ path.
Args:
- leaf_node (TreeNode): The leaf node to start ascending from.
- prefix_path (list): A list to store the item as they are ascended.
+ leaf_node: The leaf node to start ascending from.
+ prefix_path: A list to store the item as they are ascended.
Example:
>>> data_set = [
@@ -239,9 +244,8 @@ def find_prefix_path(base_pat: frozenset, tree_node: TreeNode | None) -> dict:
Find the conditional pattern base for a given base pattern.
Args:
- base_pat (frozenset): The base pattern for which to find
- the conditional pattern base.
- tree_node (TreeNode): The node in the FP-Tree.
+ base_pat: The base pattern for which to find the conditional pattern base.
+ tree_node: The node in the FP-Tree.
Example:
>>> data_set = [
@@ -253,9 +257,12 @@ def find_prefix_path(base_pat: frozenset, tree_node: TreeNode | None) -> dict:
... ]
>>> min_sup = 2
>>> fp_tree, header_table = create_tree(data_set, min_sup)
+ >>> fp_tree
+ TreeNode('Null Set', 1, None)
+ >>> len(header_table)
+ 4
>>> base_pattern = frozenset(['A'])
- >>> cond_pat = find_prefix_path(base_pattern, fp_tree.children['A'])
- >>> sorted(cond_pat.keys())
+ >>> sorted(find_prefix_path(base_pattern, fp_tree.children['A']))
[]
"""
cond_pats: dict = {}
@@ -279,11 +286,11 @@ def mine_tree(
Mine the FP-Tree recursively to discover frequent itemsets.
Args:
- in_tree (TreeNode): The FP-Tree to mine.
- header_table (dict): The header table with item information.
- min_sup (int): The minimum support threshold.
- pre_fix (set): A set of items as a prefix for the itemsets being mined.
- freq_item_list (list): A list to store the frequent itemsets.
+ in_tree: The FP-Tree to mine.
+ header_table: The header table dictionary with item information.
+ min_sup: The minimum support threshold.
+ pre_fix: A set of items as a prefix for the itemsets being mined.
+ freq_item_list: A list to store the frequent itemsets.
Example:
>>> data_set = [
@@ -295,7 +302,8 @@ def mine_tree(
... ]
>>> min_sup = 2
>>> fp_tree, header_table = create_tree(data_set, min_sup)
-
+ >>> fp_tree
+ TreeNode('Null Set', 1, None)
>>> frequent_itemsets = []
>>> mine_tree(fp_tree, header_table, min_sup, set([]), frequent_itemsets)
>>> expe_itm = [{'C'}, {'C', 'A'}, {'E'}, {'A', 'E'}, {'E', 'B'}, {'A'}, {'B'}]
@@ -309,7 +317,7 @@ def mine_tree(
new_freq_set.add(base_pat)
freq_item_list.append(new_freq_set)
cond_patt_bases = find_prefix_path(base_pat, header_table[base_pat][1])
- my_cond_tree, my_head = create_tree(list(cond_patt_bases.keys()), min_sup)
+ my_cond_tree, my_head = create_tree(list(cond_patt_bases), min_sup)
if my_head is not None:
# Pass header_table[base_pat][1] as node_to_test to update_header
header_table[base_pat][1] = update_header(
@@ -319,11 +327,10 @@ def mine_tree(
if __name__ == "__main__":
- import doctest
-
- doctest.testmod()
+ from doctest import testmod
- data_set: list = [
+ testmod()
+ data_set: list[frozenset] = [
frozenset(["bread", "milk", "cheese"]),
frozenset(["bread", "milk"]),
frozenset(["bread", "diapers"]),
@@ -333,7 +340,10 @@ def mine_tree(
frozenset(["diapers", "cheese"]),
frozenset(["bread", "milk", "cheese", "diapers"]),
]
+ print(f"{len(data_set) = }")
fp_tree, header_table = create_tree(data_set, min_sup=3)
+ print(f"{fp_tree = }")
+ print(f"{len(header_table) = }")
freq_items: list = []
mine_tree(fp_tree, header_table, 3, set(), freq_items)
- print(freq_items)
+ print(f"{freq_items = }")