Senti Strength Java Manual
User Manual:
Open the PDF directly: View PDF
Page Count: 12
- SentiStrength Java User Manual
- Quick start
- Sentiment classification tasks
- Classify a single text
- Classify all lines of text in a file for sentiment [includes accuracy evaluations]
- Classify texts in a column within a file or folder
- Listen at a port for texts to classify
- Run interactively from the command line
- Process stdin and send to stdout
- Import the JAR file to run within your Java program
- Improving the accuracy of SentiStrength
- Options:
- Explain the classification
- Only classify text near specified keywords
- Classify positive (1 to 5) and negative (-1 to -5) sentiment strength separately
- Use trinary classification (positive-negative-neutral)
- Use binary classification (positive-negative)
- Use a single positive-negative scale classification
- Location of linguistic data folder
- Location of sentiment term weights
- Location of output folder
- File name extension for output
- Classification algorithm parameters
- Additional considerations
- Machine learning evaluations