daisy-bf/create_model.py at main · BSc-learned-indexes/daisy-bf · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
import pandas as pd
import pickle
import os
import argparse
from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.neural_network import MLPClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix, accuracy_score, f1_score, ConfusionMatrixDisplay
from progress.bar import Bar

# Progress bar
bar = Bar('Creating model           ', max=5)

# Arguments
parser = argparse.ArgumentParser()
parser.add_argument("--file_name", action="store", dest="file_name", type=str, required=False,
                    help="which dataset?", default = "url_data_vectorized")
parser.add_argument("--data_path", action="store", dest="data_path", type=str, required=False,
                    help="path to vectorized data-set", default = "./data/vectorized")
parser.add_argument("--model_type", action="store", dest="model_type", type=str, required=False,
                    help="model_type", default = "random_forest")
parser.add_argument("--train_split", action="store", dest="train_split", type=float, required=False,
                    help="split of training data", default = 0.3)
parser.add_argument("--rfc_max_dept", action="store", dest="rfc_max_dept", type=int, required=False,
                    help="split of training data", default = None)
parser.add_argument("--rfc_n_estimators", action="store", dest="rfc_n_estimators", type=int, required=False,
                    help="split of training data", default = 100)
args = parser.parse_args()

# Path
path = f"{args.data_path}/{args.file_name}.csv"
data = pd.read_csv(path)

bar.next()

# -------------------------------------
# Split data into train and test set
# -------------------------------------

keys = data[data.label == "malicious"]
non_keys = data[data.label == "benign"]
non_keys = non_keys.sample(n=len(keys), random_state=42)

model_data = pd.concat([keys, non_keys])
# input vector
x = model_data[['hostname_length',
    'path_length', 'fd_length', 'tld_length', 'count-', 'count@', 'count?',
    'count%', 'count.', 'count=', 'count-http','count-https', 'count-www', 'count-digits',
    'count-letters', 'count_dir', 'use_of_ip']]
# label
y = model_data['result']

# Splitting the data into Training and Testing
x_train, x_test, y_train, y_test = train_test_split(x, y, train_size=args.train_split, random_state=42)

# -------------------------------------
# Model type 1: Random forest classifier
# -------------------------------------

if (args.model_type == "random_forest"):

    # Classify with random forest
    model = RandomForestClassifier(n_estimators=args.rfc_n_estimators, max_leaf_nodes=20, random_state=42)
    model.fit(x_train, y_train)
    model_predictions = model.predict(x_test)

    # Model performance indicators
    model_accuracy = accuracy_score(y_test, model_predictions)
    model_f1 = f1_score(y_test, model_predictions)
    model_confusion_matrix = confusion_matrix(y_test,model_predictions)

    bar.next()

# ------------------------------
# Model type 2: Regression model
# ------------------------------

if (args.model_type == "regression"):

    # Classify with LogisticRegression
    model = LogisticRegression()
    model.fit(x_train,y_train)

    model_predictions = model.predict(x_test)

    # Model performance indicators
    model_accuracy = accuracy_score(y_test, model_predictions)
    model_f1 = f1_score(y_test, model_predictions)
    model_confusion_matrix = confusion_matrix(y_test,model_predictions)


# ------------------------------
# Model type 3: Multi-layer Perceptron (MLP)
# ------------------------------
if (args.model_type == "mlp"):
    model = MLPClassifier(random_state=42, max_iter=1000)
    model.fit(x_train,y_train)

    model_predictions = model.predict(x_test)

    # Model performance indicators
    model_accuracy = accuracy_score(y_test, model_predictions)
    model_f1 = f1_score(y_test, model_predictions)
    model_confusion_matrix = confusion_matrix(y_test,model_predictions)

# ------------------------------
# Model type 4: Gaussian Naive Bayes
# ------------------------------

if (args.model_type == "naive_bayes"):
    model = GaussianNB()
    model.fit(x_train,y_train)

    model_predictions = model.predict(x_test)

    # Model performance indicators
    model_accuracy = accuracy_score(y_test, model_predictions)
    model_f1 = f1_score(y_test, model_predictions)
    model_confusion_matrix = confusion_matrix(y_test,model_predictions)

# -------
# Exports
# -------

# Export data with model scores
# index, url, label, score (prediction)
x = data[['hostname_length',
       'path_length', 'fd_length', 'tld_length', 'count-', 'count@', 'count?',
       'count%', 'count.', 'count=', 'count-http','count-https', 'count-www', 'count-digits',
       'count-letters', 'count_dir', 'use_of_ip']]
x_predictions = model.predict_proba(x)
x_predictions_df = pd.DataFrame(x_predictions, columns=["benign_score", "malicious_score"])

export = data[["url"]].copy()
export["label"] = data["result"]
export["score"] = x_predictions_df["malicious_score"].round(15)
export["label"] = export["label"].replace([0], -1)
export.to_csv('./data/scores/exported_urls.csv', index=False)

bar.next()

#Export pickled model
pickle.dump(model, open('./models/model.pickle', 'wb'))
bar.next()

# Export model metadata
# f1, accuracy, confusion_matrix, size
model_size = os.path.getsize('./models/model.pickle') * 8 # convert to bits

model_metadata = {
    "f1": model_f1,
    "accuracy": model_accuracy,
    "confusion_matrix_0_0": model_confusion_matrix[0,0],
    "confusion_matrix_0_1": model_confusion_matrix[0,1],
    "confusion_matrix_1_0": model_confusion_matrix[1,0],
    "confusion_matrix_1_1": model_confusion_matrix[1,1],
    "size_bits": model_size
}

confusion_plot = ConfusionMatrixDisplay(model_confusion_matrix)

export_meta = pd.DataFrame(data=model_metadata, index=[0])

export_meta.to_csv('./models/model_meta.csv', index=False)
bar.next()

bar.finish()