[Python] Error in predicting the strategy in a trained model

I am trying to run a ML Program to fetch strategy of fix based on error keys. I tried to manipulate the testsize and randomstate but no luck. Sample of the rules in the "error_strategy_data.csv"

Note: the below is formatted in rows and columns "errorMessage ", "strategy" in the csv

errorMessage : Deadlock found when trying to get lock
strategy: Retry transaction or redesign locking errorMessage :Syntax error near SELECT
strategy: Fix SQL syntax or use parameterized queries errorMessage :Connection refused to database server
strategy: Verify DB network settings and credentials errorMessage :Timeout occurred while executing query
strategy: Retry the request with exponential backoff

Below is my training model

from sklearn.model_selection import train_test_split
from sklearn.pipeline import Pipeline
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report
import joblib

data = pd.read_csv("C:\\Users\\Abhishek.mishra01\\Desktop\\error_strategy_data.csv")
X = data["error_message"]
y = data["strategy"]

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=42)

pipeline = Pipeline([
('tfidf', TfidfVectorizer(
lowercase=True,
stop_words='english',
ngram_range=(1,2),
max_features=5000,
#strip_accents='unicode'
)),
('clf', LogisticRegression(max_iter=1000, C=1))
])

pipeline.fit(X_train, y_train)

preds = pipeline.predict(X_test)
print(classification_report(y_test, preds))

joblib.dump(pipeline, "error_strategy_model.joblib") ```

My Test class

import joblib

# Load
model = joblib.load("error_strategy_model.joblib")

# Predict
error_message = "select"
predicted_strategy = model.predict([error_message])[0]
print("Suggested strategy:", predicted_strategy)



Output expected:
"Fix SQL syntax or use parameterized queries" but getting "Verify DB network settings and credentials"

More interestingly, it is only picking the 3rd row as a strategy for any of the error message entered.
Please help me out

Continue reading...