Past two weeks, we all have discussed our projects with the professors. To be honest, I don't have anything else to add at this point. My project is Sentiment Analysis in the Azerbaijani language and based on the suggestions, me and my teammate did a little research on the possibility of including emotions in the analysis. Currently, there are several challenges on the sentiment analysis part of our project as well. For instance, sentiment analysis can vary with the domain for a given set of words; simple aggregation of sentiment based on individual words may not be useful in some domains; generally, managing data; building a model, and many more. We will work on all those challenges. That's really all for now.
data is all around us.
Data is everywhere and it keeps growing as time goes on. Today it plays an important role in our life even if we don't feel it right away. There are billions of data, mostly unstructured about different topics from various sources. So, working with the necessary data for a specific project might not be that easy in terms of collecting it and processing it.
To start working with data, first, it needs to be defined a project and to be understood what kind of data is necessary for that specific project and also what are the ways to access them? If such kind of data does not exist or is not accessible, then it might affect the success of the project directly.
As of my case, me and my teammate, Leyla, plan to do sentiment analysis as a thesis project. There are several options to collect the data from. The one we found is called Subjectivity Lexicon. Basically, it has a dictionary that includes thousands of words assigned to be positive or negative. This type of data exists in English, but not in Azerbaijani. The ones that exist in Azerbaijani have really low accuracy. Of course, there are numerous phrases that have different meanings in English and Azerbaijani. So, we plan to go through this dictionary as initial data and also make some additions and deletions based on the principles of the Azerbaijani language. Besides, we also found the data of news in Azerbaijani, which shows the positiveness and negativeness of some statements rather than words. We think it might also be a help for us to develop our project. So far, that's pretty much it on where we are now about the data part of our project.
Until next week.
one week has passed since the class started.
One week has passed since the class started. So far, we have had a discussion on how to choose a research topic in a general aspect. It is essential to define common requirements thoroughly before diving into a research topic and discovering as you go. The key point is defining a problem, answering its solution, and understanding how likely the idea is to be realized, whose interest in this project is, how the project will be measured to make sure it is going to succeed, and many more. After the problem is defined, it is needed to collect necessary data, choose methods and tools, and start developing. It is also important to make tons of tests and write documentation while the development is on. Another point is again to really make sure to consider the risks, the payoffs, cost, duration, etc. in detail.
As for my impressions of the class, so far so good. I am looking forward to exploring possible research topics starting the next week.
Mammadnaghiyev Kheybar,
Hey, there. I am Kheybar Mammadnaghiyev. Yes, I know, it is a very long surname, but life still goes on. I would like to briefly introduce myself here.
I am a 21-years-old student. I was born in Baku, Azerbaijan, and grew up there. I have only lived outside of Baku when I was an exchange student in the UK for a half year and now in the US.
As of Computer Science, I have an interest in Data Analytics, Machine Learning, and also Cyber Security. I am more willing to explore different topics in Computer Science in order to discover more.
As of the past projects, I have experience as a mobile application developer.
Besides Computer Science, I would love to say that I'm drawing or playing guitar, but I'm not :). I have tried them but did not work for me yet. I still love art though. I love listening to songs in the indie or alternative genre, and love to visit art galleries.