JarCastingixa-pipe-opinion
ixa-pipe-opinion is a multilingual Aspect Based Opinion tagger consisting of Opinion Target Extraction (OTE), Aspect detection and polarity tagging.
ixa-pipe-opinion is part of IXA pipes, a multilingual set of NLP tools developed by the IXA NLP Group [http://ixa2.si.ehu.es/ixa-pipes].
Please go to [http://ixa2.si.ehu.es/ixa-pipes] for general information about the IXA pipes tools but also for official releases, including source code and binary packages for all the tools in the IXA pipes toolkit.
This document is intended to be the usage guide of ixa-pipe-opinion. If you really need to clone and install this repository instead of using the releases provided in [http://ixa2.si.ehu.es/ixa-pipes], please scroll down to the end of the document for the installation instructions.
NOTICE!!: ixa-pipe-opinion is in Maven Central.
TABLE OF CONTENTS
- Aspect Based Sentiment Analysis(ABSA)
- Opinion Target Extraction (OTE)
- Aspect detection
- Polarity tagging
- Server mode
- Training your own models
- Evaluation
OVERVIEW
ixa-pipe-opinion provides Aspect Based Sentiment Analysis (ABSA) using sequence labeling and document classification trained on SemEval ABSA 2014-2016 datasets.
- Opinion Target Extraction (OTE): Sequence labeler to detect the opinion targets.
- Aspect detection: Sequence labeler and Document Classification to detect aspects of opinions.
- Polarity tagging: For a document and/or sentence the polarity is tagged.
We provide competitive models based on robust local features and exploiting unlabeled data via clustering features. The clustering features are based on Brown, Clark (2003) and Word2Vec clustering. To avoid duplication of efforts, we use and contribute to the API provided by the Apache OpenNLP project with our own custom developed features for each of the three tasks.
Features
A description of every feature is provided in the sequenceTrainer.properties and docTrainer.properties file distributed with ixa-pipe-ml. As the training functionality is configured in properties files, please do check this document.
ABSA-Models
We distribute Opinion Target Extraction models for the restaurant domain trained with the SemEval ABSA 2014-2016 datasets for four languages: English, Spanish, French and Dutch. We obtain state-of-the-art results for every evaluation setting and language.
- English 2016 ABSA Models [21MB]
- Spanish ABSA Models [14MB]
- French ABSA Models [12MB]
- Dutch ABSA Models [5MB]
OTE-Models
We distribute Opinion Target Extraction models for the restaurant domain trained with the SemEval ABSA 2014-2016 datasets for six languages: English, Spanish, French, Dutch, Turkish and Russian. The system was the task winner of the ote substask in the ABSA 2015 edition. We also obtain state-of-the-art results for the rest of evaluation settings and languages.
- English 2014-2016 ABSA OTE Models [119MB]
- Spanish ABSA OTE Models [29MB]
- French ABSA OTE Models [34MB]
- Dutch ABSA OTE Models [27MB]
- Turkish ABSA OTE Models [12MB]
- Russian ABSA OTE Models [61MB]
We also distribute a set of scripts to reproduce the results using the ABSA test set and evaluation scripts.
Usage
ixa-pipe-opinion provides a runable jar with the following command-line basic functionalities:
- absa: reads a NAF document containing wf and term elements and performs Aspect Based Sentiment Analysis.
- ote: reads a NAF document containing wf and term elements and performs opinion target extraction (OTE).
- aspect: reads a NAF document containing wf and term elements and detects aspects.
- pol: reads a NAF document containing wf and term elements and tags polarity.
- server: starts a TCP service loading the model and required resources.
- client: sends a NAF document to a running TCP server.
Each of these functionalities are accessible by adding (absa|ote|aspect|pol|server|client) as a subcommand to ixa-pipe-opinion-${version}-exec.jar. Please read below and check the -help parameter:
java -jar target/ixa-pipe-opinion-${version}-exec.jar (absa|ote|aspect|pol|server|client) -help
ABSA
Aspect Based Sentiment Analysis requires an input NAF with wf and term elements:
cat file.txt | java -jar ixa-pipe-tok-$version-exec.jar tok -l $lang | java -jar ixa-pipe-pos-$version-exec.jar tag -m posmodel.bin -lm lemma-model.bin | java -jar ixa-pipe-opinion-${version}-exec.jar absa -t targetModel.bin -p polarityModel.bin
ixa-pipe-opinion reads NAF documents (with wf and term elements) via standard input and outputs opinion elements containing targets, aspects and their polarities in NAF through standard output. The NAF format specification is here:
(http://wordpress.let.vupr.nl/naf/)
You can get the necessary input for ixa-pipe-nerc by piping ixa-pipe-tok and ixa-pipe-pos as shown in the example.
There are several options to tag with ixa-pipe-opinion (check the -help parameter for more info).
- targetModel: model for target and aspect detection.
- polarityModel: model for polarity classification
- language: pass the language as a parameter.
- outputFormat: Output annotation in a format: available OpenNLP native format and NAF. It defaults to NAF.
OTE
Opinion Target Extraction requires an input NAF with wf and term elements:
cat file.txt | java -jar ixa-pipe-tok-$version-exec.jar tok -l $lang | java -jar ixa-pipe-pos-$version-exec.jar tag -m posmodel.bin -lm lemma-model.bin | java -jar ixa-pipe-opinion-${version}-exec.jar ote -m model.bin
ixa-pipe-opinion reads NAF documents (with wf and term elements) via standard input and outputs opinion targets in NAF through standard output. The NAF format specification is here:
(http://wordpress.let.vupr.nl/naf/)
You can get the necessary input for ixa-pipe-nerc by piping ixa-pipe-tok and ixa-pipe-pos as shown in the example.
There are several options to tag with ixa-pipe-opinion (check the -help parameter for more info).
- model: pass the model as a parameter.
- language: pass the language as a parameter.
- outputFormat: Output annotation in a format: available OpenNLP native format and NAF. It defaults to NAF.
Aspects
Aspect detection requires an input NAF with wf and term elements. It is also required to specify the tagger type: Document Classification (doc) or Sequence labeling (seq).
cat file.txt | java -jar ixa-pipe-tok-$version-exec.jar tok -l $lang | java -jar ixa-pipe-pos-$version-exec.jar tag -m posmodel.bin -lm lemma-model.bin | java -jar ixa-pipe-opinion-${version}-exec.jar aspect -t seq -m model.bin
ixa-pipe-opinion reads NAF documents (with wf and term elements) via standard input and outputs opinion expressions containing the aspects for each sentence. The NAF format specification is here:
(http://wordpress.let.vupr.nl/naf/)
You can get the necessary input for ixa-pipe-nerc by piping ixa-pipe-tok and ixa-pipe-pos as shown in the example.
There are several other options to tag with ixa-pipe-opinion (check the -help parameter for more info).
- model: pass the model as a parameter.
- tagger: choose between doc (document classification) or seq (Sequence labeling).
- language: pass the language as a parameter.
- outputFormat: Output annotation in a format: available OpenNLP native format and NAF. It defaults to NAF.
Polarity
Polarity tagging requires an input NAF with wf and term elements.
cat file.txt | java -jar ixa-pipe-tok-$version-exec.jar tok -l $lang | java -jar ixa-pipe-pos-$version-exec.jar tag -m posmodel.bin -lm lemma-model.bin | java -jar ixa-pipe-opinion-${version}-exec.jar pol -m model.bin
The polarity parameter of ixa-pipe-opinion reads NAF documents (with wf and term elements) via standard input and outputs opinion expressions containing the polarity for each sentence. The NAF format specification is here:
(http://wordpress.let.vupr.nl/naf/)
You can get the necessary input for ixa-pipe-nerc by piping ixa-pipe-tok and ixa-pipe-pos as shown in the example.
There are several other options to tag with ixa-pipe-opinion (check the -help parameter for more info).
- model: pass the model as a parameter.
- language: pass the language as a parameter.
- outputFormat: Output annotation in a format: available OpenNLP native format and NAF. It defaults to NAF.
- dict: Tag tokens with a polarity lexicon.
Server
We can start the TCP server as follows:
java -jar target/ixa-pipe-opinion-${version}-exec.jar server -l en --port 2030 -t aspect -c seq -m model.bin
Once the server is running we can send NAF documents containing (at least) the term layer like this:
cat file.pos.naf | java -jar target/ixa-pipe-opinion-${version}-exec.jar client -p 2060
API
The easiest way to use ixa-pipe-opinion programatically is via Apache Maven. Add this dependency to your pom.xml:
<dependency>
<groupId>eus.ixa</groupId>
<artifactId>ixa-pipe-opinion</artifactId>
<version>1.0.0</version>
</dependency>
JAVADOC
The javadoc of the module is located here:
ixa-pipe-opinion/target/ixa-pipe-opinion-$version-javadoc.jar
Module contents
The contents of the module are the following:
+ formatter.xml Apache OpenNLP code formatter for Eclipse SDK
+ pom.xml maven pom file which deals with everything related to compilation and execution of the module
+ src/ java source code of the module and required resources
+ Furthermore, the installation process, as described in the README.md, will generate another directory:
target/ it contains binary executable and other directories
+ trainParams.properties A template properties file containing documention
for every available option
INSTALLATION
Installing the ixa-pipe-nerc requires the following steps:
If you already have installed in your machine the Java 1.8+ and MAVEN 3, please go to step 3 directly. Otherwise, follow these steps:
1. Install JDK 1.8
If you do not install JDK 1.8+ in a default location, you will probably need to configure the PATH in .bashrc or .bash_profile:
export JAVA_HOME=/yourpath/local/java8
export PATH=${JAVA_HOME}/bin:${PATH}
If you use tcsh you will need to specify it in your .login as follows:
setenv JAVA_HOME /usr/java/java8
setenv PATH ${JAVA_HOME}/bin:${PATH}
If you re-login into your shell and run the command
java -version
You should now see that your JDK is 1.8.
2. Install MAVEN 3
Download MAVEN from
https://maven.apache.org/download.cgi
Now you need to configure the PATH. For Bash Shell:
export MAVEN_HOME=/home/ragerri/local/apache-maven-3.3.9
export PATH=${MAVEN_HOME}/bin:${PATH}
For tcsh shell:
setenv MAVEN3_HOME ~/local/apache-maven-3.3.5
setenv PATH ${MAVEN3}/bin:{PATH}
If you re-login into your shell and run the command
mvn -version
You should see reference to the MAVEN version you have just installed plus the JDK that is using.
3. Get module source code
If you must get the module source code from here do this:
git clone https://github.com/ixa-ehu/ixa-pipe-opinion
4. Compile
Execute this command to compile ixa-pipe-opinion:
cd ixa-pipe-opinion
mvn clean package
This step will create a directory called target/ which contains various directories and files. Most importantly, there you will find the module executable:
ixa-pipe-opinion-${version}-exec.jar
This executable contains every dependency the module needs, so it is completely portable as long as you have a JVM 1.8 installed.
To install the module in the local maven repository, usually located in ~/.m2/, execute:
mvn clean install
Contact information
Rodrigo Agerri
IXA NLP Group
University of the Basque Country (UPV/EHU)
E-20018 Donostia-San Sebastián
[email protected]