Sentiment Analysis Multi-tool

A multi-tool for danish sentiment analysis. It is the bachelor project of Steffan Eybye Christensen, Lucas Puvis de Chavannes, Peter Due Jensen & Mads Kongsbak, under supervision of assistant professor Leon Derczynski at ITU

HOWTO

The executeable has several different classifiers and functionality that you can use.

To run a classifier use:

SAM svm -r MyData.csv

The available classifiers are:

• svm - Support vector machine
• rb - Rule based(lexical) classification
• lstm - Long term short memory
• random - Random classification

If you want to create training/test data for the classifiers, you can instead do:

SAM data -r MyData.csv

This will random order your data and take 10% as test data.

REQUIREMENTS

To run this program, you need Windows and the following software installed on your computer:

• .NET Core 2.1 or higher
• Python 3.6 or higher
• Pip3

PYTHON CLASSIFIER

To run the classifier, navigate to the first Sentimentinator folder in a shell, and execute the following command to install the required packages:

pip3 install -r requirements.txt

There are two classifiers that can be trained

• classifier_svm.py
• classifier_svm_experimental.py

To train the models, run the python scripts and exchange the filename with the desired:

python3 Sentimentinator/Classifiers/classifier_svm.py

FURTHER READING

Further information about the dataset we gathered can be read here.

If you wish to read more about both the dataset and this software, you can read our thesis.

LICENSE

This code is licenced CC-BY: you may use it for commercial and non-commercial purposes, and you MUST acknowledge the source clearly if you so. Acknowledge this reference for SAM:

Mads Guldborg Kjeldgaard Kongsbak, Steffan Eybye Christensen, Lucas Høyberg Puvis de Chavannes, Peter Due Jensen. "Sentiment Analysis Multitool, SAM". 2019. Bachelor dissertation, IT University of Copenhagen.

Bibtex:

@techreport{SAMsentiment,
  author={Mads Guldborg Kjeldgaard Kongsbak and Steffan Eybye Christensen and Lucas Høyberg Puvis~de~Chavannes and Peter Due Jensen},
  title={Sentiment Analysis Multitool, SAM},
  year={2019},
  institution={IT University of Copenhagen},
}

For details, see LICENSE.md.