[FRAUD] 그래프 그림 그리기

Author

김보람

Published

November 1, 2023

imports

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt 
import networkx as nx
import sklearn
import xgboost as xgb

# sklearn
from sklearn import model_selection # split함수이용
from sklearn import ensemble # RF,GBM
from sklearn import metrics
from sklearn.metrics import precision_score, recall_score, f1_score
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from sklearn.naive_bayes import GaussianNB

# gnn
import torch
import torch.nn.functional as F
import torch_geometric
from torch_geometric.nn import GCNConv

def down_sample_textbook(df):
    df_majority = df[df.is_fraud==0].copy()
    df_minority = df[df.is_fraud==1].copy()
    df_maj_dowsampled = sklearn.utils.resample(df_majority, n_samples=len(df_minority), replace=False, random_state=42)
    df_downsampled = pd.concat([df_minority, df_maj_dowsampled])
    return df_downsampled

def compute_time_difference(group):
    n = len(group)
    result = []
    for i in range(n):
        for j in range(n):
            time_difference = abs(group.iloc[i].trans_date_trans_time.value - group.iloc[j].trans_date_trans_time.value)
            result.append([group.iloc[i].name, group.iloc[j].name, time_difference])
    return result

def mask(df):
    df_tr,df_test = sklearn.model_selection.train_test_split(df, random_state=42)
    N = len(df)
    train_mask = [i in df_tr.index for i in range(N)]
    test_mask = [i in df_test.index for i in range(N)]
    train_mask = np.array(train_mask)
    test_mask = np.array(test_mask)
    return train_mask, test_mask

def edge_index_selected(edge_index):
    theta = edge_index[:,2].mean()
    edge_index[:,2] = (np.exp(-edge_index[:,2]/theta) != 1)*(np.exp(-edge_index[:,2]/theta))
    edge_index = edge_index.tolist()
    mean_ = np.array(edge_index)[:,2].mean()
    selected_edges = [(int(row[0]), int(row[1])) for row in edge_index if row[2] > mean_]
    edge_index_selected = torch.tensor(selected_edges, dtype=torch.long).t()
    return edge_index_selected


fraudTrain = pd.read_csv("~/Desktop/fraudTrain.csv").iloc[:,1:]
fraudTrain = fraudTrain.assign(trans_date_trans_time= list(map(lambda x: pd.to_datetime(x), fraudTrain.trans_date_trans_time)))
fraudTrain
trans_date_trans_time cc_num merchant category amt first last gender street city ... lat long city_pop job dob trans_num unix_time merch_lat merch_long is_fraud
0 2019-01-01 00:00:00 2.703190e+15 fraud_Rippin, Kub and Mann misc_net 4.97 Jennifer Banks F 561 Perry Cove Moravian Falls ... 36.0788 -81.1781 3495 Psychologist, counselling 1988-03-09 0b242abb623afc578575680df30655b9 1325376018 36.011293 -82.048315 0
1 2019-01-01 00:00:00 6.304230e+11 fraud_Heller, Gutmann and Zieme grocery_pos 107.23 Stephanie Gill F 43039 Riley Greens Suite 393 Orient ... 48.8878 -118.2105 149 Special educational needs teacher 1978-06-21 1f76529f8574734946361c461b024d99 1325376044 49.159047 -118.186462 0
2 2019-01-01 00:00:00 3.885950e+13 fraud_Lind-Buckridge entertainment 220.11 Edward Sanchez M 594 White Dale Suite 530 Malad City ... 42.1808 -112.2620 4154 Nature conservation officer 1962-01-19 a1a22d70485983eac12b5b88dad1cf95 1325376051 43.150704 -112.154481 0
3 2019-01-01 00:01:00 3.534090e+15 fraud_Kutch, Hermiston and Farrell gas_transport 45.00 Jeremy White M 9443 Cynthia Court Apt. 038 Boulder ... 46.2306 -112.1138 1939 Patent attorney 1967-01-12 6b849c168bdad6f867558c3793159a81 1325376076 47.034331 -112.561071 0
4 2019-01-01 00:03:00 3.755340e+14 fraud_Keeling-Crist misc_pos 41.96 Tyler Garcia M 408 Bradley Rest Doe Hill ... 38.4207 -79.4629 99 Dance movement psychotherapist 1986-03-28 a41d7549acf90789359a9aa5346dcb46 1325376186 38.674999 -78.632459 0
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
1048570 2020-03-10 16:07:00 6.011980e+15 fraud_Fadel Inc health_fitness 77.00 Haley Wagner F 05561 Farrell Crescent Annapolis ... 39.0305 -76.5515 92106 Accountant, chartered certified 1943-05-28 45ecd198c65e81e597db22e8d2ef7361 1362931649 38.779464 -76.317042 0
1048571 2020-03-10 16:07:00 4.839040e+15 fraud_Cremin, Hamill and Reichel misc_pos 116.94 Meredith Campbell F 043 Hanson Turnpike Hedrick ... 41.1826 -92.3097 1583 Geochemist 1999-06-28 c00ce51c6ebb7657474a77b9e0b51f34 1362931670 41.400318 -92.726724 0
1048572 2020-03-10 16:08:00 5.718440e+11 fraud_O'Connell, Botsford and Hand home 21.27 Susan Mills F 005 Cody Estates Louisville ... 38.2507 -85.7476 736284 Engineering geologist 1952-04-02 17c9dc8b2a6449ca2473726346e58e6c 1362931711 37.293339 -84.798122 0
1048573 2020-03-10 16:08:00 4.646850e+18 fraud_Thompson-Gleason health_fitness 9.52 Julia Bell F 576 House Crossroad West Sayville ... 40.7320 -73.1000 4056 Film/video editor 1990-06-25 5ca650881b48a6a38754f841c23b77ab 1362931718 39.773077 -72.213209 0
1048574 2020-03-10 16:08:00 2.283740e+15 fraud_Buckridge PLC misc_pos 6.81 Shannon Williams F 9345 Spencer Junctions Suite 183 Alpharetta ... 34.0770 -84.3033 165556 Prison officer 1997-12-27 8d0a575fe635bbde12f1a2bffc126731 1362931730 33.601468 -83.891921 0

1048575 rows × 22 columns

데이터정리

_df1 = fraudTrain[fraudTrain["is_fraud"] == 0].sample(frac=0.20, random_state=42)
_df2 = fraudTrain[fraudTrain["is_fraud"] == 1]
df02 = pd.concat([_df1,_df2])
df02.shape
(214520, 22)
df50 = down_sample_textbook(df02)
df50 = df50.reset_index()
df50.shape
(12012, 23)

tr/test

mask(df50)
(array([False,  True,  True, ...,  True, False,  True]),
 array([ True, False, False, ..., False,  True, False]))
train_mask, test_mask = mask(df50)

edge_index 설정

# groups = df50.groupby('cc_num')
# edge_index_list_plus = [compute_time_difference(group) for _, group in groups]
# edge_index_list_plus_flat = [item for sublist in edge_index_list_plus for item in sublist]
# edge_index_list_plus_nparr = np.array(edge_index_list_plus_flat)
# np.save('edge_index_list_plus50.npy', edge_index_list_plus_nparr)
edge_index = np.load('edge_index_list_plus50.npy').astype(np.float64)
edge_index.shape
(200706, 3)
edge_index_selected = edge_index_selected(edge_index)
edge_index_selected
tensor([[ 1023,  1023,  1023,  ..., 11944, 11944, 11944],
        [ 1024,  1028,  1031,  ...,  4257,  9241,  9782]])
np.array(edge_index_selected)
array([[ 1023,  1023,  1023, ..., 11944, 11944, 11944],
       [ 1024,  1028,  1031, ...,  4257,  9241,  9782]])

분석 1(GCN): amt

x = torch.tensor(df50['amt'], dtype=torch.float).reshape(-1,1)
y = torch.tensor(df50['is_fraud'],dtype=torch.int64)
data = torch_geometric.data.Data(x=x, edge_index = edge_index_selected, y=y, train_mask = train_mask, test_mask = test_mask)
data



torch.manual_seed(202250926)
class GCN2(torch.nn.Module):
    def __init__(self):
        super().__init__()
        self.conv1 = GCNConv(1, 32)
        self.conv2 = GCNConv(32,2)

    def forward(self, data):
        x, edge_index = data.x, data.edge_index

        x = self.conv1(x, edge_index)
        x = F.relu(x)
        x = F.dropout(x, training=self.training)
        x = self.conv2(x, edge_index)

        return F.log_softmax(x, dim=1)

X = (data.x[data.train_mask]).numpy()
XX = (data.x[data.test_mask]).numpy()
y = (data.y[data.train_mask]).numpy()
yy = (data.y[data.test_mask]).numpy()

model = GCN2()
optimizer = torch.optim.Adam(model.parameters(), lr=0.01, weight_decay=5e-4)
model.train()
for epoch in range(400):
    optimizer.zero_grad()
    out = model(data)
    loss = F.nll_loss(out[data.train_mask], data.y[data.train_mask])
    loss.backward()
    optimizer.step()
model.eval()

pred = model(data).argmax(dim=1)
yyhat = pred[data.test_mask]

metrics = [sklearn.metrics.accuracy_score,
           sklearn.metrics.precision_score,
           sklearn.metrics.recall_score,
           sklearn.metrics.f1_score]

_results1= pd.DataFrame({m.__name__:[m(yy,yyhat).round(6)] for m in metrics},index=['분석1'])
_results1
accuracy_score precision_score recall_score f1_score
분석1 0.902098 0.862478 0.95913 0.90824 
import matplotlib.pyplot as plt
G = torch_geometric.utils.to_networkx(data, to_undirected=True)
pos = nx.spring_layout(G)
labels = {i: f"{i}\n{data.y[i].item()}" for i in range(data.num_nodes)}
nx.draw(G, pos, with_labels=True, node_color='lightblue', labels=labels)
plt.show()

anomaly_nodes = [i for i in range(data.num_nodes) if data.y[i].item() == 1]

# 중요한 노드에 대한 레이블 설정
labels = {i: f"{i}\n{data.y[i].item()}" for i in anomaly_nodes}

# 중요한 노드 강조
node_color = ['lightblue' if i not in anomaly_nodes else 'red' for i in range(data.num_nodes)]

# 그래프 시각화
nx.draw(G, pos, with_labels=True, node_color=node_color, labels=labels)
plt.show()

G
<networkx.classes.graph.Graph at 0x7f3eaffc0880>

흠…..

import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 노드 A와 B의 인덱스
node_A_index = 1023
node_B_index = 1024

# 노드 A와 B 간의 경로 찾기
shortest_path = nx.shortest_path(G, source=node_A_index, target=node_B_index)

# 노드 A와 B 간의 경로로 이루어진 부분 그래프 추출
subgraph = G.subgraph(shortest_path)

# 그래프 시각화
pos = nx.spring_layout(subgraph)  # 그래프 레이아웃 설정

# 노드 그리기
nx.draw_networkx_nodes(subgraph, pos, node_size=200)

# 엣지 그리기
nx.draw_networkx_edges(subgraph, pos)

# 그래프 출력
plt.axis('off')  # 축 숨기기
plt.show()

subgraph = nx.ego_graph(G, 1023, radius=1)

# 그래프 시각화
pos = nx.spring_layout(subgraph)  # 그래프 레이아웃 설정

# 노드 그리기
nx.draw_networkx_nodes(subgraph, pos, node_size=200)

# 엣지 그리기
nx.draw_networkx_edges(subgraph, pos)

# 그래프 출력
plt.axis('off')  # 축 숨기기
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 노드 1023과 연결된 모든 노드와 엣지를 포함하는 부분 그래프 추출
subgraph = nx.ego_graph(G, 1023, radius=1)

# 그래프 시각화
pos = nx.spring_layout(subgraph)  # 그래프 레이아웃 설정

# 노드 그리기
nx.draw_networkx_nodes(subgraph, pos, node_size=200)

# 엣지 그리기
nx.draw_networkx_edges(subgraph, pos)

# 노드 인덱스 표시
labels = {node: str(node) for node in subgraph.nodes()}
nx.draw_networkx_labels(subgraph, pos, labels, font_size=10)

# 그래프 출력
plt.axis('off')  # 축 숨기기
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

subgraph = nx.ego_graph(G, 9782, radius=1)

# 그래프 시각화
pos = nx.spring_layout(subgraph)  # 그래프 레이아웃 설정

# 노드 그리기
nx.draw_networkx_nodes(subgraph, pos, node_size=200)

# 엣지 그리기
nx.draw_networkx_edges(subgraph, pos)

# 노드 인덱스 표시
labels = {node: str(node) for node in subgraph.nodes()}
nx.draw_networkx_labels(subgraph, pos, labels, font_size=10)

# 그래프 출력
plt.axis('off')  # 축 숨기기
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 노드 1023과 연결된 모든 노드와 엣지를 포함하는 부분 그래프 추출
subgraph1 = nx.ego_graph(G, 1023, radius=1)
subgraph2 = nx.ego_graph(G, 9782, radius=1)

# 그래프 시각화
pos1 = nx.spring_layout(subgraph1)  # 그래프 레이아웃 설정
pos2 = nx.spring_layout(subgraph2)

# 노드 그리기
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color='b', label='Node 1023')
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color='g', label='Node 1024')

# 엣지 그리기
nx.draw_networkx_edges(subgraph1, pos1)
nx.draw_networkx_edges(subgraph2, pos2)

# 노드 인덱스 표시
labels1 = {node: str(node) for node in subgraph1.nodes()}
labels2 = {node: str(node) for node in subgraph2.nodes()}
nx.draw_networkx_labels(subgraph1, pos1, labels1, font_size=10)
nx.draw_networkx_labels(subgraph2, pos2, labels2, font_size=10)

# 그래프 출력
plt.axis('off')  # 축 숨기기

# 레이블 표시
plt.legend(loc='best')

plt.show()

# 노드 1023와 9782 간의 연결 확인
are_connected = G.has_edge(1023, 9782)

if are_connected:
    print("Node 1023 and Node 9782 are connected.")
else:
    print("Node 1023 and Node 9782 are not connected.")
Node 1023 and Node 9782 are not connected.
import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 노드 1023과 연결된 모든 노드와 엣지를 포함하는 부분 그래프 추출
subgraph1 = nx.ego_graph(G, 1023, radius=1)
subgraph2 = nx.ego_graph(G, 9782, radius=1)

# 왼쪽 그래프 레이아웃 설정
pos1 = nx.spring_layout(subgraph1, pos=None, seed=42)  # 그래프 레이아웃 설정

# 오른쪽 그래프 레이아웃 설정
pos2 = nx.spring_layout(subgraph2, pos=None, seed=43)

# 노드 그리기
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color='b', label='Node 1023')
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color='g', label='Node 1024')

# 엣지 그리기
nx.draw_networkx_edges(subgraph1, pos1)
nx.draw_networkx_edges(subgraph2, pos2)

# 노드 인덱스 표시
labels1 = {node: str(node) for node in subgraph1.nodes()}
labels2 = {node: str(node) for node in subgraph2.nodes()}
nx.draw_networkx_labels(subgraph1, pos1, labels1, font_size=10)
nx.draw_networkx_labels(subgraph2, pos2, labels2, font_size=10)

# 그래프 출력
plt.axis('off')  # 축 숨기기

# 레이블 표시
plt.legend(loc='best')

plt.show()

len(df50['cc_num'].unique())
932
df50['cc_num'].value_counts()
4.302480e+15    43
1.800650e+14    36
2.131740e+14    35
2.720430e+15    34
2.242540e+15    33
                ..
3.885950e+13     1
4.026220e+12     1
6.526450e+15     1
4.972230e+15     1
6.535330e+15     1
Name: cc_num, Length: 932, dtype: int64
cc_num_counts = df50['cc_num'].value_counts()
mean_counts = cc_num_counts.mean()
mean_counts
12.888412017167383
counts = df50['cc_num'].value_counts()
cc_num_with_13_counts = counts[counts == 13].index
cc_num_with_13_counts
Float64Index([3585740000000000.0, 4839040000000000.0,  370349000000000.0,
              3583090000000000.0, 3534720000000000.0,  342351000000000.0,
               213161000000000.0, 6011860000000000.0, 2288810000000000.0,
                  560881000000.0,  346273000000000.0,    4760120000000.0,
              2356280000000000.0,     501803000000.0, 2297450000000000.0,
              4933460000000000.0, 3565940000000000.0, 3597980000000000.0,
                36913600000000.0,     630423000000.0,    4746000000000.0,
              3536820000000000.0, 2720890000000000.0,          4.537e+18,
                 4755700000000.0,   38057500000000.0,         4.2929e+18,
              6011230000000000.0, 4060580000000000.0, 3546900000000000.0,
               345060000000000.0,   30596500000000.0, 3597340000000000.0,
              4124540000000000.0, 4204240000000000.0,  375237000000000.0,
                36153900000000.0, 4334230000000000.0, 6012000000000000.0,
              4134460000000000.0, 4653880000000000.0, 4509140000000000.0,
              3565420000000000.0],
             dtype='float64')
df50[df50['cc_num'] ==  4.537e+18].index
Int64Index([ 6298,  6714,  7704,  7823,  7910,  7990,  9145,  9806, 10338,
            10910, 11621, 11847, 11926],
           dtype='int64')
import networkx as nx
import matplotlib.pyplot as plt

selected_indices = df50[df50['cc_num'] ==  3585740000000000.0].index
is_fraud_values = df50.loc[selected_indices, 'is_fraud'].tolist()
G = torch_geometric.utils.to_networkx(data, to_undirected=True)
subgraph = G.subgraph(selected_indices)
node_colors = ['r' if node in selected_indices and is_fraud_values[selected_indices.get_loc(node)] == 1 else 'k' for node in subgraph.nodes()]
pos = nx.spring_layout(subgraph, seed=42)
plt.figure(figsize=(10, 5))
nx.draw_networkx_nodes(subgraph, pos, node_size=200, node_color=node_colors)
nx.draw_networkx_edges(subgraph, pos, edge_color='gray')
nx.draw_networkx_labels(subgraph, pos, font_size=10)
plt.axis('off')
plt.show()

index = df50[df50['cc_num'] == 3585740000000000.0].index
is_fraud_value = df50.loc[index, 'is_fraud']
is_fraud_value
4437     1
4443     1
4446     1
4449     1
4451     1
4452     1
4455     1
4465     1
6914     0
6974     0
8624     0
8830     0
11375    0
Name: is_fraud, dtype: int64
df50[df50['cc_num'] == 4.302480e+15].index
Int64Index([ 2293,  2294,  2295,  2296,  2297,  2298,  2299,  2300,  4709,
             4711,  4718,  4720,  4728,  4729,  4730,  4733,  4734,  4736,
             4739,  4742,  6081,  6751,  6855,  7025,  7554,  7817,  7946,
             8173,  9152,  9166,  9548,  9708,  9773,  9807, 10198, 10828,
            10909, 10953, 10962, 11273, 11330, 11562, 11946],
           dtype='int64')
index = df50[df50['cc_num'] == 4.302480e+15].index
is_fraud_value = df50.loc[index, 'is_fraud']
is_fraud_value
2293     1
2294     1
2295     1
2296     1
2297     1
2298     1
2299     1
2300     1
4709     1
4711     1
4718     1
4720     1
4728     1
4729     1
4730     1
4733     1
4734     1
4736     1
4739     1
4742     1
6081     0
6751     0
6855     0
7025     0
7554     0
7817     0
7946     0
8173     0
9152     0
9166     0
9548     0
9708     0
9773     0
9807     0
10198    0
10828    0
10909    0
10953    0
10962    0
11273    0
11330    0
11562    0
11946    0
Name: is_fraud, dtype: int64

- cc_num=4.302480e+15 인 그래프

해당 노드의 is_fraud=1 이면 빨간색으로 표시했다

import networkx as nx
import matplotlib.pyplot as plt

# 4.50~에 해당하는 행의 인덱스를 찾음
selected_indices = df50[df50['cc_num'] == 4.302480e+15].index

# 선택된 인덱스에 대한 is_fraud 값을 가져옴
is_fraud_values = df50.loc[selected_indices, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 서브그래프를 추출
subgraph = G.subgraph(selected_indices)

# 노드와 엣지 색상을 설정
node_colors = ['r' if node in selected_indices and is_fraud_values[selected_indices.get_loc(node)] == 1 else 'k' for node in subgraph.nodes()]

# 그래프 레이아웃 설정
pos = nx.spring_layout(subgraph, seed=42)

# 그래프 그리기
plt.figure(figsize=(10, 5))
nx.draw_networkx_nodes(subgraph, pos, node_size=200, node_color=node_colors)
nx.draw_networkx_edges(subgraph, pos, edge_color='gray')
nx.draw_networkx_labels(subgraph, pos, font_size=10)

# 축 숨기기
plt.axis('off')

# 그래프 출력
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# 4.30~에 해당하는 행의 인덱스를 찾음
selected_indices1 = df50[df50['cc_num'] == 4.302480e+15].index
selected_indices2 = df50[df50['cc_num'] == 2.242540e+15].index

# 선택된 인덱스에 대한 is_fraud 값을 가져옴
is_fraud_values1 = df50.loc[selected_indices1, 'is_fraud'].tolist()
is_fraud_values2 = df50.loc[selected_indices2, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 첫 번째 서브그래프 추출
subgraph1 = G.subgraph(selected_indices1)

# 노드와 엣지 색상 설정 (서브그래프 1)
node_colors1 = ['r' if node in selected_indices1 and is_fraud_values1[selected_indices1.get_loc(node)] == 1 else 'k' for node in subgraph1.nodes()]

# 그래프 레이아웃 설정 (서브그래프 1)
pos1 = nx.spring_layout(subgraph1, seed=42)

# 두 번째 서브그래프 추출
subgraph2 = G.subgraph(selected_indices2)

# 노드와 엣지 색상 설정 (서브그래프 2)
node_colors2 = ['r' if node in selected_indices2 and is_fraud_values2[selected_indices2.get_loc(node)] == 1 else 'k' for node in subgraph2.nodes()]

# 그래프 레이아웃 설정 (서브그래프 2)
pos2 = nx.spring_layout(subgraph2, seed=43)

# 그래프 그리기
plt.figure(figsize=(15, 5))

# 첫 번째 하위 그래프
plt.subplot(1, 2, 1)
plt.title('cc_num=4.302480e+15')
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color=node_colors1)
nx.draw_networkx_edges(subgraph1, pos1, edge_color='gray')
nx.draw_networkx_labels(subgraph1, pos1, font_size=10)
plt.axis('off')

# 두 번째 하위 그래프
plt.subplot(1, 2, 2)
plt.title('cc_num=2.242540e+15')
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color=node_colors2)
nx.draw_networkx_edges(subgraph2, pos2, edge_color='gray')
nx.draw_networkx_labels(subgraph2, pos2, font_size=10)
plt.axis('off')

# 그래프 출력
plt.tight_layout()
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# 4.30~에 해당하는 행의 인덱스를 찾음
selected_indices1 = df50[df50['cc_num'] == 4.302480e+15].index
selected_indices2 = df50[df50['cc_num'] == 2.242540e+15].index

# 선택된 인덱스에 대한 is_fraud 값을 가져옴
is_fraud_values1 = df50.loc[selected_indices1, 'is_fraud'].tolist()
is_fraud_values2 = df50.loc[selected_indices2, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 첫 번째 서브그래프 추출
subgraph1 = G.subgraph(selected_indices1)

# 노드와 엣지 색상 설정 (서브그래프 1)
node_colors1 = ['r' if node in selected_indices1 and is_fraud_values1[selected_indices1.get_loc(node)] == 1 else 'k' for node in subgraph1.nodes()]

# 엣지 색상 설정 (파란색)
edge_colors1 = ['b' if (u, v) in subgraph1.edges() else 'gray' for u, v in subgraph1.edges()]

# 그래프 레이아웃 설정 (서브그래프 1)
pos1 = nx.spring_layout(subgraph1, seed=42)

# 두 번째 서브그래프 추출
subgraph2 = G.subgraph(selected_indices2)

# 노드와 엣지 색상 설정 (서브그래프 2)
node_colors2 = ['r' if node in selected_indices2 and is_fraud_values2[selected_indices2.get_loc(node)] == 1 else 'k' for node in subgraph2.nodes()]

# 엣지 색상 설정 (초록색)
edge_colors2 = ['g' if (u, v) in subgraph2.edges() else 'gray' for u, v in subgraph2.edges()]

# 그래프 레이아웃 설정 (서브그래프 2)
pos2 = nx.spring_layout(subgraph2, seed=43)

# 그래프 그리기
plt.figure(figsize=(10, 5))

# 첫 번째 하위 그래프
plt.title('cc_num=4.302480e+15')
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color=node_colors1)
nx.draw_networkx_edges(subgraph1, pos1, edge_color=edge_colors1)
nx.draw_networkx_labels(subgraph1, pos1, font_size=10)

# 두 번째 하위 그래프
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color=node_colors2)
nx.draw_networkx_edges(subgraph2, pos2, edge_color=edge_colors2)
nx.draw_networkx_labels(subgraph2, pos2, font_size=10)

# 축 숨기기
plt.axis('off')

# 그래프 출력
plt.show()

df50[df50['cc_num'] == 6.535330e+15].index
Int64Index([11740], dtype='int64')
df50[df50['cc_num'] == 4.503100e+18].index
Int64Index([5381, 5382, 7996, 8217, 9176, 10248, 10517, 10613, 10950, 11321], dtype='int64')
df50[df50['is_fraud'] == 1]['cc_num'].value_counts()
2.131740e+14    24
3.506040e+15    23
1.800680e+14    22
6.011380e+15    22
1.800850e+14    21
                ..
6.011110e+15     2
4.809700e+12     2
3.576430e+15     2
4.503100e+18     2
3.546670e+15     2
Name: cc_num, Length: 596, dtype: int64
import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

subgraph = nx.ego_graph(G, 2293, radius=1)

# 그래프 시각화
pos = nx.spring_layout(subgraph)  # 그래프 레이아웃 설정

# 노드 그리기
nx.draw_networkx_nodes(subgraph, pos, node_size=200)

# 엣지 그리기
nx.draw_networkx_edges(subgraph, pos)

# 노드 인덱스 표시
labels = {node: str(node) for node in subgraph.nodes()}
nx.draw_networkx_labels(subgraph, pos, labels, font_size=10)

# 그래프 출력
plt.axis('off')  # 축 숨기기
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 노드 1023과 연결된 모든 노드와 엣지를 포함하는 부분 그래프 추출
subgraph1 = nx.ego_graph(G, 2293, radius=1)
subgraph2 = nx.ego_graph(G, 11740, radius=1)

# 왼쪽 그래프 레이아웃 설정
pos1 = nx.spring_layout(subgraph1, pos=None, seed=42)  # 그래프 레이아웃 설정

# 오른쪽 그래프 레이아웃 설정
pos2 = nx.spring_layout(subgraph2, pos=None, seed=43)

# 노드 그리기
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color='b', label='cc_num=4.302480e+15')
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color='g', label='cc_num=6.535330e+15')

# 엣지 그리기
nx.draw_networkx_edges(subgraph1, pos1)
nx.draw_networkx_edges(subgraph2, pos2)

# 노드 인덱스 표시
labels1 = {node: str(node) for node in subgraph1.nodes()}
labels2 = {node: str(node) for node in subgraph2.nodes()}
nx.draw_networkx_labels(subgraph1, pos1, labels1, font_size=10)
nx.draw_networkx_labels(subgraph2, pos2, labels2, font_size=10)

# 그래프 출력
plt.axis('off')  # 축 숨기기

# 레이블 표시
plt.legend(loc='best')

plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 노드 1023과 연결된 모든 노드와 엣지를 포함하는 부분 그래프 추출
subgraph1 = nx.ego_graph(G, 2293, radius=1)
subgraph2 = nx.ego_graph(G, 8217, radius=1)

# 왼쪽 그래프 레이아웃 설정
pos1 = nx.spring_layout(subgraph1, pos=None, seed=42)  # 그래프 레이아웃 설정

# 오른쪽 그래프 레이아웃 설정
pos2 = nx.spring_layout(subgraph2, pos=None, seed=43)

# 두 개의 하위 그래프 생성
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))

# 첫 번째 하위 그래프에 그래프 그리기
ax1.set_title('cc_num=4.302480e+15')
nx.draw_networkx_nodes(subgraph1, pos1, ax=ax1, node_size=200, node_color='b')
nx.draw_networkx_edges(subgraph1, pos1, ax=ax1)
labels1 = {node: str(node) for node in subgraph1.nodes()}
nx.draw_networkx_labels(subgraph1, pos1, labels1, font_size=10, ax=ax1)

# 두 번째 하위 그래프에 그래프 그리기
ax2.set_title('cc_num=4.503100e+18')
nx.draw_networkx_nodes(subgraph2, pos2, ax=ax2, node_size=200, node_color='g')
nx.draw_networkx_edges(subgraph2, pos2, ax=ax2)
labels2 = {node: str(node) for node in subgraph2.nodes()}
nx.draw_networkx_labels(subgraph2, pos2, labels2, font_size=10, ax=ax2)

# 축 숨기기
ax1.axis('off')
ax2.axis('off')

# 그래프 출력
plt.tight_layout()
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 노드 1023과 연결된 모든 노드와 엣지를 포함하는 부분 그래프 추출
subgraph1 = nx.ego_graph(G, 2293, radius=1)
subgraph2 = nx.ego_graph(G, 8217, radius=1)

# 왼쪽 그래프 레이아웃 설정
pos1 = nx.spring_layout(subgraph1, pos=None, seed=42)  # 그래프 레이아웃 설정

# 오른쪽 그래프 레이아웃 설정
pos2 = nx.spring_layout(subgraph2, pos=None, seed=43)

# 두 개의 하위 그래프 생성
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))

# 첫 번째 하위 그래프에 그래프 그리기
ax1.set_title('cc_num=4.302480e+15')
nx.draw_networkx_nodes(subgraph1, pos1, ax=ax1, node_size=200, node_color='b')
labels1 = {node: str(node) for node in subgraph1.nodes()}

# 엣지 색상 설정 (is_fraud가 1일 때 빨간색, 그 외에는 검정색)
edge_colors1 = ['r' if subgraph1[u][v].get('is_fraud', 0) == 1 else 'k' for u, v in subgraph1.edges()]
nx.draw_networkx_edges(subgraph1, pos1, edgelist=subgraph1.edges(), edge_color=edge_colors1, ax=ax1)
nx.draw_networkx_labels(subgraph1, pos1, labels1, font_size=10, ax=ax1)

# 두 번째 하위 그래프에 그래프 그리기
ax2.set_title('cc_num=4.503100e+18')
nx.draw_networkx_nodes(subgraph2, pos2, ax=ax2, node_size=200, node_color='g')
labels2 = {node: str(node) for node in subgraph2.nodes()}

# 엣지 색상 설정 (is_fraud가 1일 때 빨간색, 그 외에는 검정색)
edge_colors2 = ['r' if subgraph2[u][v].get('is_fraud', 0) == 1 else 'k' for u, v in subgraph2.edges()]
nx.draw_networkx_edges(subgraph2, pos2, edgelist=subgraph2.edges(), edge_color=edge_colors2, ax=ax2)
nx.draw_networkx_labels(subgraph2, pos2, labels2, font_size=10, ax=ax2)

# 축 숨기기
ax1.axis('off')
ax2.axis('off')

# 그래프 출력
plt.tight_layout()
plt.show()

  • fraud값이 1인게 표시가 안되네??
data
Data(x=[12012, 1], edge_index=<function edge_index_selected at 0x7f3f8d358ee0>, y=[12012], train_mask=[12012], test_mask=[12012])
data.edge_index
<function __main__.edge_index_selected(edge_index)>
x = torch.tensor(df50['amt'], dtype=torch.float).reshape(-1,1)
y = torch.tensor(df50['is_fraud'],dtype=torch.int64)
data = torch_geometric.data.Data(x=x, edge_index = edge_index_selected, y=y, train_mask = train_mask, test_mask = test_mask)
data
Data(x=[12012, 1], edge_index=[2, 93730], y=[12012], train_mask=[12012], test_mask=[12012])

- ppt 삽입할 그래프.

- cc_num

  • 거래량이 가장 많은 cc_num:4.302480e+15

  • 거래량 평균인cc_num: 4.2929e+18

import networkx as nx
import matplotlib.pyplot as plt

# 4.50~에 해당하는 행의 인덱스를 찾음
selected_indices = df50[df50['cc_num'] == 4.302480e+15].index

# 선택된 인덱스에 대한 is_fraud 값을 가져옴
is_fraud_values = df50.loc[selected_indices, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 서브그래프를 추출
subgraph = G.subgraph(selected_indices)

# 노드와 엣지 색상을 설정
node_colors = ['r' if node in selected_indices and is_fraud_values[selected_indices.get_loc(node)] == 1 else 'k' for node in subgraph.nodes()]

# 그래프 레이아웃 설정
pos = nx.spring_layout(subgraph, seed=42)

# 그래프 그리기
plt.figure(figsize=(10, 5))
nx.draw_networkx_nodes(subgraph, pos, node_size=200, node_color=node_colors)
nx.draw_networkx_edges(subgraph, pos, edge_color='gray')
nx.draw_networkx_labels(subgraph, pos, font_size=10)

# 축 숨기기
plt.axis('off')

# 그래프 출력
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# 4.50~에 해당하는 행의 인덱스를 찾음
selected_indices = df50[df50['cc_num'] == 4.2929e+18].index

# 선택된 인덱스에 대한 is_fraud 값을 가져옴
is_fraud_values = df50.loc[selected_indices, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 서브그래프를 추출
subgraph = G.subgraph(selected_indices)

# 노드와 엣지 색상을 설정
node_colors = ['r' if node in selected_indices and is_fraud_values[selected_indices.get_loc(node)] == 1 else 'k' for node in subgraph.nodes()]

# 그래프 레이아웃 설정
pos = nx.spring_layout(subgraph, seed=42)

# 그래프 그리기
plt.figure(figsize=(10, 5))
nx.draw_networkx_nodes(subgraph, pos, node_size=200, node_color=node_colors)
nx.draw_networkx_edges(subgraph, pos, edge_color='gray')
nx.draw_networkx_labels(subgraph, pos, font_size=10)

# 축 숨기기
plt.axis('off')

# 그래프 출력
plt.show()

import networkx as nx
import matplotlib.pyplot as plt

# 첫 번째 데이터 선택
selected_indices1 = df50[df50['cc_num'] == 4.302480e+15].index
is_fraud_values1 = df50.loc[selected_indices1, 'is_fraud'].tolist()

# 두 번째 데이터 선택
selected_indices2 = df50[df50['cc_num'] == 4.2929e+18].index
is_fraud_values2 = df50.loc[selected_indices2, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 서브그래프 추출
subgraph1 = G.subgraph(selected_indices1)
subgraph2 = G.subgraph(selected_indices2)

# 노드 색상 설정
node_colors1 = ['r' if node in selected_indices1 and is_fraud_values1[selected_indices1.get_loc(node)] == 1 else 'k' for node in subgraph1.nodes()]
node_colors2 = ['r' if node in selected_indices2 and is_fraud_values2[selected_indices2.get_loc(node)] == 1 else 'k' for node in subgraph2.nodes()]

# 엣지 색상 설정
edge_colors1 = ['g' if edge in subgraph1.edges() else 'k' for edge in subgraph1.edges()]
edge_colors2 = ['b' if edge in subgraph2.edges() else 'k' for edge in subgraph2.edges()]

# 그래프 레이아웃 설정
pos1 = nx.spring_layout(subgraph1, seed=42)
pos2 = nx.spring_layout(subgraph2, seed=42)

# 그래프 그리기
plt.figure(figsize=(12, 6))

# 그래프 1 그리기
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color=node_colors1)
nx.draw_networkx_edges(subgraph1, pos1, edge_color=edge_colors1)

# 그래프 2 그리기
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color=node_colors2)
nx.draw_networkx_edges(subgraph2, pos2, edge_color=edge_colors2)

# 축 숨기기
plt.axis('off')

# 그래프 출력
plt.show()

import networkx as nx
import matplotlib.pyplot as plt
import matplotlib.style as style

# 스타일 설정 (ggplot 스타일 사용)
style.use('ggplot')

# 첫 번째 데이터 선택
selected_indices1 = df50[df50['cc_num'] == 4.302480e+15].index
is_fraud_values1 = df50.loc[selected_indices1, 'is_fraud'].tolist()

# 두 번째 데이터 선택
selected_indices2 = df50[df50['cc_num'] == 4.2929e+18].index
is_fraud_values2 = df50.loc[selected_indices2, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 서브그래프 추출
subgraph1 = G.subgraph(selected_indices1)
subgraph2 = G.subgraph(selected_indices2)

# 노드 색상 설정
node_colors1 = ['r' if node in selected_indices1 and is_fraud_values1[selected_indices1.get_loc(node)] == 1 else 'k' for node in subgraph1.nodes()]
node_colors2 = ['r' if node in selected_indices2 and is_fraud_values2[selected_indices2.get_loc(node)] == 1 else 'k' for node in subgraph2.nodes()]

# 엣지 색상 설정
edge_colors1 = ['g' if edge in subgraph1.edges() else 'k' for edge in subgraph1.edges()]
edge_colors2 = ['b' if edge in subgraph2.edges() else 'k' for edge in subgraph2.edges()]

# 그래프 레이아웃 설정 (kamada_kawai 레이아웃 사용)
pos1 = nx.kamada_kawai_layout(subgraph1)
pos2 = nx.kamada_kawai_layout(subgraph2)

# 그래프 그리기
plt.figure(figsize=(12, 6))

# 그래프 1 그리기
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color=node_colors1, node_shape='o')  # 노드 모양: 원 (circle)
nx.draw_networkx_edges(subgraph1, pos1, edge_color=edge_colors1)

# 그래프 2 그리기
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color=node_colors2, node_shape='D')  # 노드 모양: 다이아몬드 (diamond)
nx.draw_networkx_edges(subgraph2, pos2, edge_color=edge_colors2)

# 범례 추가
legend_elements = [
    plt.Line2D([0], [0], marker='o', color='g', markerfacecolor='black', markersize=10, label='cc_num 1'),
    plt.Line2D([0], [0], marker='D', color='b', markerfacecolor='black', markersize=10, label='cc_num 2'),
    plt.Line2D([0], [0], marker='o', color='black', markerfacecolor='r', markersize=10, label='fraud=1'),
]
plt.legend(handles=legend_elements, loc='lower center', ncol=3, frameon=True)

# 그래프 출력
plt.axis('off')
plt.show()

import networkx as nx
import matplotlib.pyplot as plt
import matplotlib.style as style

# 스타일 설정 (ggplot 스타일 사용)
style.use('ggplot')

# 첫 번째 데이터 선택
selected_indices1 = df50[df50['cc_num'] == 4.302480e+15].index
is_fraud_values1 = df50.loc[selected_indices1, 'is_fraud'].tolist()

# 두 번째 데이터 선택
selected_indices2 = df50[df50['cc_num'] == 4.2929e+18].index
is_fraud_values2 = df50.loc[selected_indices2, 'is_fraud'].tolist()

# PyTorch Geometric 그래프를 NetworkX 그래프로 변환
G = torch_geometric.utils.to_networkx(data, to_undirected=True)

# 서브그래프 추출
subgraph1 = G.subgraph(selected_indices1)
subgraph2 = G.subgraph(selected_indices2)

# 노드 색상 설정
node_colors1 = ['r' if node in selected_indices1 and is_fraud_values1[selected_indices1.get_loc(node)] == 1 else 'k' for node in subgraph1.nodes()]
node_colors2 = ['r' if node in selected_indices2 and is_fraud_values2[selected_indices2.get_loc(node)] == 1 else 'k' for node in subgraph2.nodes()]

# 엣지 색상 설정
edge_colors1 = ['g' if edge in subgraph1.edges() else 'k' for edge in subgraph1.edges()]
edge_colors2 = ['b' if edge in subgraph2.edges() else 'k' for edge in subgraph2.edges()]

# 그래프 레이아웃 설정 (kamada_kawai 레이아웃 사용)
pos1 = nx.kamada_kawai_layout(subgraph1)
pos2 = nx.kamada_kawai_layout(subgraph2)

# 그래프 그리기
plt.figure(figsize=(12, 6))

# 그래프 1 그리기
nx.draw_networkx_nodes(subgraph1, pos1, node_size=200, node_color=node_colors1, node_shape='o')  # 노드 모양: 원 (circle)
nx.draw_networkx_edges(subgraph1, pos1, edge_color=edge_colors1)

# 그래프 2 그리기
nx.draw_networkx_nodes(subgraph2, pos2, node_size=200, node_color=node_colors2, node_shape='D')  # 노드 모양: 다이아몬드 (diamond)
nx.draw_networkx_edges(subgraph2, pos2, edge_color=edge_colors2)

# 범례 추가
legend_elements = [
    plt.Line2D([0], [0], marker='o', color='g', markerfacecolor='black', markersize=10, label='cc_num 1'),
    plt.Line2D([0], [0], marker='D', color='b', markerfacecolor='black', markersize=10, label='cc_num 2'),
    plt.Line2D([0], [0], marker='o', color='black', markerfacecolor='r', markersize=10, label='fraud=1'),
]
legend = plt.legend(handles=legend_elements, loc='lower center', ncol=3, frameon=True)
legend.set_bbox_to_anchor((0.5, -0.1))  # 범례 위치 조정

# 그래프 출력
plt.axis('off')
plt.show()