The Worksshop was hled by PP Algebra University College from Croatia on 9.6.2021. 

At the very beginning of the event, Mr. Leo Mršić Ph.D. from Algebra University College welcomed all the participants and gave a short overview of the manual annotation process taking place at the moment. He explained how this phase will help the outcome and why it is important. Also, Mr. Mršić stressed out the importance of the inclusion of the partners who are not annotating at the moment as their understanding of the process is also of big importance for the project and its deliverables.

Mr. Mršić continued by shortly explaining the logic of the annotation process, he also gave a few examples and finished off by giving the word to Mr. Artur Šilić Ph.D. from the Presscut.

Further on, he gave an explanation of what sentiment analysis is and what its goal is (finding emotional context and opinions towards certain entity). Solutions used for sentiment analysis are sentiment dictionaries, regression methods and the number one solution, neural networks. After giving a short description of neural networks, Mr. Šilić continued with explaining how annotation is done, firstly explaining positive, negative and neutral classification and leading the participants through 4 examples.

There were a lot of doubts with annotating certain words as they were positive but used in a negative context or vice versa, so the thorough analysis took place and Mr. Šilić gave an answer to all the questions considering the examples participants were annotating. 

Mr. Mršić took the word to give some final thoughts considering the overall project as well as the timeline.