Sentiment Analysis Console App with ML.NET

Project roadmap

1. 01

   Load and prepare the training dataset

   ~1.5h

   Load a labeled CSV of text/sentiment pairs into an ML.NET IDataView, splitting into train/test sets.

2. 02

   Build the text featurization pipeline

   ~2h

   Configure ML.NET's text featurization transforms (tokenization, n-grams, normalization).

3. 03

   Train a binary classification model

   ~2h

   Configure and train a trainer such as SdcaLogisticRegression on the featurized data.

4. 04

   Evaluate model performance

   ~1.5h

   Evaluate accuracy, F1 score, and AUC on the held-out test set and print a metrics report.

5. 05

   Persist and reload the trained model

   ~1.5h

   Save the trained model to disk and implement loading it back for prediction without retraining.

6. 06

   Build the interactive prediction console

   ~1.5h

   Let users type sentences and see live sentiment predictions with confidence scores.

Resources

• DocsC# Documentation