jentegeo/jennifer_email_classifier

Email Classifier – RandomForest + TF-IDF

A machine learning model that classifies IT service desk emails into four categories: Incident, Problem, Change, and Request.

Model Details

• Vectorizer: TF-IDF (1–3 n-grams, 15k features)
• Classifier: RandomForestClassifier
• Preprocessing: Lowercasing, lemmatization, stopword removal (retains negations)
• Oversampling: RandomOverSampler to handle class imbalance

Overall accuracy: 76%
Macro avg F1: 0.73

How to Use

Load Pretrained Model

import joblib

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

# Predict
sample = ["My Outlook stopped working after update."]
prediction = model.predict(sample)
print(prediction)