Hugging Face's logo

AndyJamesTurner
/
suicideDetector

Text Classification
Scikit-learn
skops
Model card Files
xet
Community
suicideDetector
File size: 4,905 Bytes 
fcbc617
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
import os 
import sklearn
import xgboost

import dill as pickle
import pandas as pd
from pathlib import Path

from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.pipeline import Pipeline
from sklearn.metrics import classification_report, f1_score, accuracy_score, roc_auc_score, RocCurveDisplay, roc_curve, auc, confusion_matrix, ConfusionMatrixDisplay

from skops import hub_utils
from skops import card

# load in in the Suicide Detection dataset
# accessible at https://www.kaggle.com/datasets/nikhileswarkomati/suicide-watch
df = pd.read_csv(
    "Suicide_Detection.csv",
    usecols=["text", "class"],
    dtype= {"text":str,"class":str}
)

# separate text and target class
X = df['text'].to_list()
y = df['class'].apply(lambda x: 1 if x == 'suicide' else 0).to_list()

# construct training and testing splits 
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)


def preprocessor(s):
    """preprocessor for the tfidf vectorizer"""

    from sklearn.feature_extraction.text import ENGLISH_STOP_WORDS

    stopwords_set = set(ENGLISH_STOP_WORDS)

    def filter(text):
        if text == None:
            return ""
        words = str(text).split()
        filtered_words = [word for word in words if word and word.lower() not in stopwords_set]
        return " ".join(filtered_words)

    return filter(s)

# construct the model pipeline
model = Pipeline([
    ('tfidf', TfidfVectorizer(preprocessor=preprocessor, ngram_range=(1, 3), min_df=100)),
    ('classifier', xgboost.XGBClassifier())
], verbose=True)

# fit the model using the training split
model.fit(X_train, y_train)

# use the trained model to make predictions on the testing set
y_pred = model.predict(X_test)
y_pred_proba = model.predict_proba(X_test)[:, 1]

# save the trained model
model_filename = "model.pkl"
with open(model_filename, mode="bw") as f:
    pickle.dump(model, file=f)

local_repo = Path("suicide-detector")

# construct the hugging face page 
hub_utils.init(
    model=model_filename,
    requirements=[f"scikit-learn={sklearn.__version__}", f"xgboost={xgboost.__version__}"],
    dst=str(local_repo),
    task="text-classification",
    data=X_test,
)

# made a header card from the metadata
model_card = card.Card(model, metadata=card.metadata_from_config(local_repo))

# add license
model_card.metadata.license = "mit"

model_description = """
Suicide Detection text classification model.

Trained on the Suicide and Depression Detection dataset (https://www.kaggle.com/datasets/nikhileswarkomati/suicide-watch)

The model vectorises each text using a trained tfidf vectorizer and then classifies using xgboost.
"""

get_started_code = """
import sklearn 
import dill as pickle

from skops import hub_utils
from pathlib import Path

suicide_detector_repo = Path("./suicide-detector")

hub_utils.download(
    repo_id="AndyJamesTurner/suicideDetector",
    dst=suicide_detector_repo
)

with open(suicide_detector_repo / "model.pkl", 'rb') as file:
    clf = pickle.load(file)

classification = clf.predict(["I want to kill myself"])[0]
"""

# add additional details to the page including 
# model description, getting started guide, and author
model_card.add(
    get_started_code=get_started_code,
    model_card_authors="Andy Turner",
    model_description=model_description,
)

# add description of how the model was evaluated
eval_descr = (
    "The model was evaluated on a 0.3 holdout split using f1 score, accuracy, confusion matrix and ROC curves."
)
model_card.add(eval_method=eval_descr)

# compute model evaluation metrics and add details to the hugging face model card
accuracy = accuracy_score(y_test, y_pred)
f1 = f1_score(y_test, y_pred, average="micro")
model_card.add_metrics(**{"accuracy": accuracy, "f1 score": f1})

cm = confusion_matrix(y_test, y_pred, labels=model.classes_)
disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=model.classes_)
disp.plot()
disp.figure_.savefig(local_repo / "confusion_matrix.png")
model_card.add_plot(**{"Confusion matrix": "confusion_matrix.png"})

fpr, tpr, thresholds = roc_curve(y_test, y_pred_proba)
roc_auc = auc(fpr, tpr)
disp = RocCurveDisplay(fpr=fpr, tpr=tpr, roc_auc=roc_auc)
disp.plot()
disp.figure_.savefig(local_repo / "roc_curve.png")
model_card.add_plot(**{"ROC Curve": "roc_curve.png"})

clf_report = classification_report(
    y_test, y_pred, output_dict=True, target_names=["not suicide", "suicide"]
)

del clf_report["accuracy"]
clf_report = pd.DataFrame(clf_report).T.reset_index()
model_card.add_table(
    folded=True,
    **{
        "Classification Report": clf_report,
    },
)

# construct a readme from the model card
model_card.save(local_repo / "README.md")

# add this file to the repo to document how it was constructed
hub_utils.add_files(
    os.path.realpath(__file__),
    dst=str(local_repo),
)