ESWC-15 Challenge on Concept-Level Sentiment Analysis
Check the wiki page to see more details
Results of the challenge
Task 1:
KIM - Accuracy: 0.4129
FBK - Accuracy: 0.4078
SENTILO - Accuracy: 0.3011
KIM is the winner of task 1 and is awarded with a Springer voucher of 150 euros
Task 3:
FBK - Precision: 0.3996; Recall: 0.5336; F1: 0.4570
FBK is the winner of task 3 and is awarded with a Springer voucher of 150 euros
Most Innovative Approach
SENTILO - The system builds on top of Discourse Representation Theory, relies on VerbNet for identifying and formalizing events and their associated thematic roles. SENTILO uses FRED which transforms such logical form to RDF by complying to Semantic Web and Linked Data design principles, and by extending the representation model with event- and situation- semantics as formally defined by DOLCE+DnS ontology. A new sentiment ontology is developed, OntoSentilo, that defines concepts and relations that characterize the entities composing an opinion sentence (opinion trigger events, holders, topics and subtopics, opinion features).
SENTILO is the winner of the "most innovative approach" and is awarded with a Springer voucher of 150 euros
Added task1Challenge_testGold.zip and task3Challenge_testGold.zip. These are the two testset where systems have to be run on. Results will be sent in the required format.
Please check the wiki for further information.
To ask questions and information please join our mailing list at http://groupspaces.com/ESWC2015-CLSA.
After you join the group, you can post messages to the mailing list address ESWC2015-CLSA@groupspaces.com
Added task3_CLSA.zip
It contains the xml annotated file for testing your system.
The output of your system for task3 should use the same structure of task3AnnotatedChallenge.xml. You do not need to report the ... content. You need to indicate the ... and all the other tags.
You can then use precisionTask3.py to evaluate your results using:
python precisionTask3.py task3AnnotatedChallenge.xml yourFileResults.xml
Please check the wiki for further information.
To ask questions and information please join our mailing list at http://groupspaces.com/ESWC2015-CLSA.
After you join the group, you can post messages to the mailing list address ESWC2015-CLSA@groupspaces.com
Added task2_CLSA.zip
It contains the xml annotated file for testing your system.
The dataset has been taken from SemEval2015 Task12: Aspect Based Sentiment Analysis http://alt.qcri.org/semeval2015/task12/
The task has been proposed for the first time during ESWC2014 for the Challenge on Concept-Level Sentiment Analysis 2014, the first edition of this current challenge. Then, the task has been reproposed by SemEval2015.
File precisionTask2.zip contains the instructions on how to run the precision recall analysis for the Task2.
Please check the wiki for further information.
To ask questions and information please join our mailing list at http://groupspaces.com/ESWC2015-CLSA.
After you join the group, you can post messages to the mailing list address ESWC2015-CLSA@groupspaces.com
Added task1_CLSA.zip
It contains the task1AnnotatedChallenge.xml annotated file for testing your system.
The output of your system for task1 should use the same structure of task1AnnotatedChallenge.xml. You do not need to report the ... content. You need to indicate the ... and ... tags.
You can then use precisionTask1.py to evaluate your results using:
python precisionTask1.py task1AnnotatedChallenge.xml yourFileResults.xml
Please check the wiki for further information.
To ask questions and information please join our mailing list at http://groupspaces.com/ESWC2015-CLSA.
After you join the group, you can post messages to the mailing list address ESWC2015-CLSA@groupspaces.com