Explaining sentiment analysis predictions
You are provided with a model that classifies product reviews as expressing positive or negative sentiment. Your task is to use LIME to identify which words in a given text_instance most influence the model's predictions.
The model_predict function for processing input texts is pre-loaded for you.
This exercise is part of the course
Explainable AI in Python
Exercise instructions
- Create a LIME text explainer named
explainer. - Generate an explanation for the model's prediction on the given
text_instancefeaturing the top five features. - Display the top contributing words and their weights that influence the model's decision.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
from lime.lime_text import LimeTextExplainer
text_instance = "Amazing battery life and the camera quality is perfect! I highly recommend this smartphone."
# Create a LIME text explainer
explainer = ____
# Generate the explanation
exp = ____
# Display the explanation
____
plt.show()