Natural Language Classification
You did a great such a great job for DeFalco's restaurant in the previous exercise that the chef has hired you for a new project.
The restaurant's menu includes an email address where visitors can give feedback about their food.
The manager wants you to create a tool that automatically sends him all the negative reviews so he can fix them, while automatically sending all the positive reviews to the owner, so the manager can ask for a raise.
You will first build a model to distinguish positive reviews from negative reviews using Yelp reviews because these reviews include a rating with each review. Your data consists of the text body of each review along with the star rating. Ratings with 1-2 stars count as "negative", and ratings with 4-5 stars are "positive". Ratings with 3 stars are "neutral" and have been dropped from the data.
Let's get started. First, run the next code cell.
import pandas as pd
# Set up code checking
from learntools.core import binder
binder.bind(globals())
from learntools.nlp.ex2 import *
print("\nSetup complete")
Setup complete
Step 1: Evaluate the Approach
Is there anything about this approach that concerns you? After you've thought about it, run the function below to see one point of view.
# Check your answer (Run this code cell to receive credit!)
step_1.solution()