Duygu analizi ve sosyal medya alanında uygulama
Files
Date
2018-07-01
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Bursa Uludağ Üniversitesi
Abstract
Bilgisayar teknolojilerinde meydana gelen değişiklik hayatın her aşamasında yenilikleri de beraberinde getirmiştir. Bunlardan en önemlileri farklı tiplerde ki verilerin derlenip saklanabilmesine, işlenmesine imkan tanıması ve geleneksel medya araçlarına alternatif olarak sosyal medyanın kullanılmaya başlanmasıdır. Geçmişte bir verinin işlenebilmesi için o verinin yapılandırılmış olma ön koşulu varken, günümüzde yapılandırılmamış verilerde son derece kolay bir şekilde işlenebilmektedir. Metin madenciliği tam da burada karşımıza çıkmaktadır. Yapılandırılmamış halde bulunan metinler artık rahatlıkla işlenebilmektedir. Duygu analizi son yıllarda hızla popülaritesi artan bir araştırma alanı olmuştur. Duygu analizinde temel amaç, metinde yer alan duygunun ortaya çıkarılabilmesidir. Sosyal medya özellikle Web 2.0 teknolojisini getirmiş olduğu ve insanların yaşamlarını tamamen değiştiren bir kavramdır. Sosyal medya sayesinde, bireyler artık kendilerine sunulan içerikleri tüketen pasif kullanıcılar değil, kendileri de içerik üretip bunu paylaşabilen aktif bireyler haline gelmiştir. Günümüzde akıllı telefona sahip olan herkes aynı zamanda birer içerik üreticisidir. R yazılımı dünya da son yıllarda en sık kullanılan ve pek çok işlemi kolaylıkla yerine getirebilen önemli bir açık kaynak kodlu programdır. Ancak ülkemizde henüz değeri çok iyi algılanamamış ve kullanım düzeyi istenilen seviyeye ulaşamamıştır. Bu tezde R yazılımı ile duygu analizi yapılmıştır. Duygu analizi için Twitter’dan, 24 Haziran 2018 de Türkiye’de yapılan Cumhurbaşkanlığı ve 27. Dönem Millet Vekilli Genel Seçimi’ne ilişkin en yüksek oy alması beklenen aday ve partilere ilişkin bir ay boyunca atılan tweetlerden ele alınmıştır. Tweetlerden hareketle duygu analizi yapılmış, bu amaçla duygu skoru hesaplanmış ve makine öğrenmesi ile hesaplamalar yapılmıştır. Elde edilen veriler ile seçim sonuçları karşılaştırılımıştır.
The change that took place in computer technology brought innovations with every step of life. The most important of these is the compilation of data in different types, the possibility of processing them, and the introduction of social media as an alternative to traditional media tools. In the past, when a given data is processed, it can be processed very easily in the unstructured data of today, while the data is pre-configured. Text mining is exactly here. Unstructured texts can now be easily processed. Emotional analysis has been a rapidly growing field of research in recent years. The main purpose of emotion analysis is to be able to reveal the sensation in the text. Social media is a concept that especially brings Web 2.0 technology and changes people's lives completely. With social media, individuals are no longer passive users who consume content that is presented to them, but become active individuals who can produce and share content themselves. Today, everyone who owns a smartphone is also a content producer. R software is one of the most frequently used open source programs in the world in recent years and can easily handle many operations. However, in our country, the value has not been perceived very well and the level of use has not reached the desired level. In this thesis, emotion analysis was done with R software. Essential texts for emotion analysis were obtained from Twitter, an important social sharing site. For sentiment analysis on Twitter, June 24, 2018 also receive the highest number of votes related to the presidential and 27. Period Vekille Nations General elections in Turkey were taken from discarded tweets related to the expected candidates and parties. Sentiment analysis was performed by tweets, the emotion score was calculated for this purpose and calculations were made by machine learning. We compared the selection results with the obtained data.
The change that took place in computer technology brought innovations with every step of life. The most important of these is the compilation of data in different types, the possibility of processing them, and the introduction of social media as an alternative to traditional media tools. In the past, when a given data is processed, it can be processed very easily in the unstructured data of today, while the data is pre-configured. Text mining is exactly here. Unstructured texts can now be easily processed. Emotional analysis has been a rapidly growing field of research in recent years. The main purpose of emotion analysis is to be able to reveal the sensation in the text. Social media is a concept that especially brings Web 2.0 technology and changes people's lives completely. With social media, individuals are no longer passive users who consume content that is presented to them, but become active individuals who can produce and share content themselves. Today, everyone who owns a smartphone is also a content producer. R software is one of the most frequently used open source programs in the world in recent years and can easily handle many operations. However, in our country, the value has not been perceived very well and the level of use has not reached the desired level. In this thesis, emotion analysis was done with R software. Essential texts for emotion analysis were obtained from Twitter, an important social sharing site. For sentiment analysis on Twitter, June 24, 2018 also receive the highest number of votes related to the presidential and 27. Period Vekille Nations General elections in Turkey were taken from discarded tweets related to the expected candidates and parties. Sentiment analysis was performed by tweets, the emotion score was calculated for this purpose and calculations were made by machine learning. We compared the selection results with the obtained data.
Description
Keywords
Duygu analizi, Sosyal medya, Twitter, Makine öğrenmesi
Citation
Kızılkaya Y. F. (2018). Duygu analizi ve sosyal medya alanında uygulama. Yayınlanmamış doktora tezi. Bursa Uludağ Üniversitesi Sosyal Bilimler Enstitüsü.