College Project · Machine Learning

Comparative Analysis of Machine Learning Models
for Resume–Job Matching

TF-IDF + SVM vs BERT Embeddings

Upload a resume PDF and paste a job description to see both models compare.

Resume PDF
📄

Click to browse or drag & drop your PDF resume

Max 5 MB

Job Description

Analysis Results

Both models analysed the resume against the job description

⭐ WINNER
Baseline Model
TF-IDF + SVM
0%
Score
Combined TF-IDF + SVM score
TF-IDF Cosine Sim
0%
SVM Confidence
0%
Matched Keywords
Accuracy
F1 Score
70
Train Pairs
Fast inference
Interpretable weights
Keyword matching only
Misses synonyms
⭐ WINNER
Advanced Model
BERT Embeddings
0%
Score
Semantic similarity score
Semantic Similarity
0%
Contextual Depth
Matched Keywords
1B+
Pre-train pairs
384
Embedding dim
MiniLM
Architecture
Understands synonyms
Semantic context
No training data needed
Deeper understanding

📊 Side-by-Side Comparison

Metric / Property TF-IDF + SVM BERT Embeddings
Match Score
Verdict
Approach Keyword frequency (TF-IDF) + SVM classifier Dense semantic embeddings (Transformer)
Synonym Handling ❌ No ✅ Yes
Context Understanding ❌ Limited ✅ Deep
Training Required Yes (labelled pairs) No (pre-trained)
Inference Speed ⚡ Very fast Moderate (CPU)
Model Size ~few KB ~90 MB
Interpretability High (feature weights) Low (black box)
Winner Baseline 🏆 Better Overall