Repository files navigation

A Demo application usingCoreML framework for sentiment polarity analysis.

CoreML modelwas converted fromScikit-learn Pipelineusingcoremltools python package.

The model is based onLinearSVC classifier and is able to distinguish betweenpositive and negative sentences withbest CV score = 0.801013024602. It wastrained usingEpinions.com dataset with reviews ofproducts and services. Accuracy can be improved by usingTfidfVectorizer for feature extraction, but it's not supported bycoremltoolsat the moment.

• Xcode 9
• iOS 11

git clone https://github.com/cocoa-ai/SentimentCoreMLDemo.gitcd SentimentCoreMLDemoopen SentimentPolarity.xcodeproj/

Build the project and run it on a simulator or a device with iOS 11.

cd Convertpython sentiment.py

Vadym Markov,markov.vadym@gmail.com

• Dataset withEpinions.com reviews was taken fromHW3: Sentiment Analysis

• scikit-learn
• Apple Machine Learning
• CoreML Framework
• coremltools