Browsing by Author "Aygun I."

Now showing 1 - 6 of 6
  • No Thumbnail Available
    Item
    Automatic Term Extraction on Turkish Scientific Texts
    (Institute of Electrical and Electronics Engineers Inc., 2020) Aygun I.; Kaya M.
    In order for a text or collection to be understood, it is very important to understand the terms contained in it. In this study, it is aimed to detect terms in a domain-specific (Cyber Security) corpus. A two-layer method is suggested for the determination of the terms used in single words or phrases. Term candidate words are determined by statistical methods in the first layer. In the second layer, the possibility of using these words in phrases with semantic approaches is checked. In the study, Word2Vec approach was used to determine semantic affinity and 3 different datasets were used. The results show that the terms used in singular or binary patterns were successfully determined using the proposed method. © 2020 IEEE.
  • No Thumbnail Available
    Item
    A Deep Learning Based Prediction Model for Diagnosing Diseases with Similar Symptoms
    (Institute of Electrical and Electronics Engineers Inc., 2021) Aygun I.; Kaya B.
    Diagnosis of diseases with similar symptoms may cause medical errors depending on the transfer of patient complaints. In this study, diseases that are similar to each other in terms of symptoms are primarily examined. In conducted experiments Diabetis Mellitus was the focus of the study and most similar disaeses to Diabetis Mellitus were determined by using statistical data and deep learning methods. Within the scope of the study, a data set containing the symptoms of patients with this disease was created. In experiments using the data of 205 patients, it was seen that the deep learning model produced the same diagnosis with physicians with a rate of over 84%. For nearly 10% of the patients used in the experiment, it was concluded that an alternative disease should also be checked. © 2021 IEEE.
  • No Thumbnail Available
    Item
    Aspect Based Twitter Sentiment Analysis on Vaccination and Vaccine Types in COVID-19 Pandemic With Deep Learning
    (Institute of Electrical and Electronics Engineers Inc., 2022) Aygun I.; Kaya B.; Kaya M.
    Due to the COVID-19 pandemic, vaccine development and community vaccination studies are carried out all over the world. At this stage, the opposition to the vaccine seen in the society or the lack of trust in the developed vaccine is an important factor hampering vaccination activities. In this study, aspect-base sentiment analysis was conducted for USA, U.K., Canada, Turkey, France, Germany, Spain and Italy showing the approach of twitter users to vaccination and vaccine types during the COVID-19 period. Within the scope of this study, two datasets in English and Turkish were prepared with 928,402 different vaccine-focused tweets collected by country. In the classification of tweets, 4 different aspects (policy, health, media and other) and 4 different BERT models (mBERT-base, BioBERT, ClinicalBERT and BERTurk) were used. 6 different COVID-19 vaccines with the highest frequency among the datasets were selected and sentiment analysis was made by using Twitter posts regarding these vaccines. To the best of our knowledge, this paper is the first attempt to understand people's views about vaccination and types of vaccines. With the experiments conducted, the results of the views of the people on vaccination and vaccine types were presented according to the countries. The success of the method proposed in this study in the F1 Score was between 84% and 88% in datasets divided by country, while the total accuracy value was 87%. © 2013 IEEE.
  • No Thumbnail Available
    Item
    Detection of Customer Opinions with Deep Learning Method for Metaverse Collaborating Brands
    (Institute of Electrical and Electronics Engineers Inc., 2022) Aygun I.; Kaya B.; Kaya M.
    In recent years, metaverse projects have been developed that both increase the number of users and bring a new concept to the use of the internet. With this development, collaborations are frequently established within the business world with metaverse projects that attract the attention of companies. In the study, the gains of the companies operating in the metaverse after these activities were examined. Thanks to the tweets collected before and after the companies participated in the metaverse, it was analyzed how potential users interpreted their participation in the metaverse. In this context, sentiment analysis experiments were conducted for five different clothing, sportswear, and retail companies (Adidas, Balenciaga, H&M, Nike, and Zara) serving in similar fields of activity. The BERT architecture, which is a language representation model, was used in the experiments, and it was seen that the positive shares on Twitter for companies increased greatly. After the companies transitioned to Metaverse, the biggest change in positive Twitter posts was seen in Nike, with 47%, and in second place, positive Twitter posts about Balenciaga increased by 42%. Experiments show that firms' assets in the metaverse create a positive perception within one month. © 2022 IEEE.
  • No Thumbnail Available
    Item
    Prediction of the Most Common Symptoms in Psychological Illnesses with Language Representation Models
    (Institute of Electrical and Electronics Engineers Inc., 2024) Aygun I.; Kaya M.
    It is a known fact as a result of researches that psychological disorders are seen more frequently in society day by day and early diagnosis of these disorders is very important. To detect psychological disorders, it is an important achievement to identify the symptoms in the sentences of potential patients. In the present study, the most frequently used symptoms in the sentences of past psychiatric patients were investigated. The deep learning supported BERT model was used to analyze the texts and the Named Entity Recognition (NER) method was used for symptom detection. Thus, a model is proposed that enables the detection of symptoms even when they are expressed in different ways. The success of the proposed model in detecting the symptoms is between 83.6 and 86.2% and the most common symptoms are shortness of breath, loss of attention and loss of appetite. © 2024 IEEE.
  • No Thumbnail Available
    Item
    Using aspect-based sentiment analysis to evaluate the global effects of the food security crisis during the Russia-Ukraine war
    (Elsevier B.V., 2025) Aygun O.; Aygun I.; Kaya M.
    The war that broke out between Russia and Ukraine in February 2022 has left many countries facing difficult challenges, one of which has been the escalation of global food security issues. This study examines how the conditions of war negatively reflect on food security and how social media posts collected from different countries around the world were used to categorize and examine societal concerns following the war. Concerns and reasons were analyzed based on countries using an aspect-based sentiment analysis using a sentiment classifier with the BERT architecture. In Africa and the Middle East, societal concerns are centered around the possibility of not accessing food and potential famine; other analyses conducted in these countries indicate that access to products such as bread, baby food, animal products (milk, cheese, etc.), pasta, and many other categories have become significantly challenging. A large portion of the world views the food security crisis caused by the war as potentially leading to economic problems and famine. Social media posts from Poland, especially, in Central and Eastern European countries are intensely focused on economic issues. The sentiment analysis studies show that the Black Sea Grain Initiative, aimed at securing food security between the two countries, increased positive sentiments globally by up to 24%. To the best of our knowledge, this study is the first to identify concerns stated at the societal level worldwide on food security and measure objectives by country. This study presents findings that can develop projections for assessing concerns reflected in society regarding food security and formulating alternative plans, contributing to many countries in this regard. © 2025 Elsevier B.V.