Week8 exercise
This commit is contained in:
@@ -171,14 +171,7 @@ sample_preds = clf.predict(sample_features)
|
||||
for text, pred in zip(sample_texts, sample_preds):
|
||||
print(f"\nReview: {text}\nPredicted Sentiment: {pred}")
|
||||
|
||||
"""-->Positive reviews dominate,
|
||||
|
||||
-->The model basically learned “always say positive”,
|
||||
|
||||
-->Hence, 84% accuracy but 0 recall for negative/neutral — a fake good score.
|
||||
|
||||
#Improving Model Balance & Realism
|
||||
"""
|
||||
"""#Improving Model Balance & Realism"""
|
||||
|
||||
# Separate by sentiment
|
||||
pos = df[df["sentiment"] == "positive"]
|
||||
@@ -217,14 +210,7 @@ print("Classification Report:\n", classification_report(y_test, y_pred))
|
||||
|
||||
print("\nConfusion Matrix:\n", confusion_matrix(y_test, y_pred))
|
||||
|
||||
"""-->It now distinguishes between negative, neutral, and positive.
|
||||
|
||||
-->Has a macro average F1 ≈ 0.57, which is fair for such a small, noisy sample.
|
||||
|
||||
-->Shows that balancing worked — negatives are now detected correctly (recall 0.83).
|
||||
|
||||
#Agents
|
||||
"""
|
||||
"""#Agents"""
|
||||
|
||||
# Base class for all agents
|
||||
class BaseAgent:
|
||||
@@ -301,7 +287,7 @@ class ReviewerAgent(BaseAgent):
|
||||
"""
|
||||
|
||||
response = self.client.messages.create(
|
||||
model="claude-3-5-haiku-20241022", # updated to latest available model
|
||||
model="claude-3-5-haiku-20241022",
|
||||
max_tokens=250,
|
||||
temperature=0.6,
|
||||
messages=[{"role": "user", "content": prompt}]
|
||||
Reference in New Issue
Block a user