95% Accuracy TF-IDF + Logistic Regression 3,000 Tweets · 5 Topics
🐦
Dataset
—
Tweets Analyzed
3,000 labeled samples
🎯
Test Set
—
Test Accuracy
CV: — ± —
📈
OVR
—
ROC-AUC Score
Weighted one-vs-rest
🔤
TF-IDF
—
Feature Dimensions
Unigram + bigram
🧪
Split
—
Test Samples
80/20 stratified split
Model Performance
Precision · Recall · F1 per Class
Classification report breakdown
Per Class
Confusion Matrix
Actual vs predicted labels
Test Set · 600
Sentiment Distribution
Ground truth label breakdown
3,000 tweets
Data Insights
Monthly Sentiment Trend
Positive / Negative / Neutral tweet volume over time
2024
Sentiment by Topic
Distribution of sentiments across each topic
5 Topics
Top TF-IDF Features per Sentiment
Most Discriminative Words & Bigrams
Logistic Regression coefficients — highest weighted features per class
Coefficients
Sample Predictions

Prediction Results

— records
TweetTopicActualPredicted CorrectConfidenceP(pos)P(neg)P(neu)