ayseirmak
Teaching and Research Assistant in Computer Science with specialization in Machine Learning & Artificial Intelligence.
Pinned Repositories
Chatter-Detection
Utilized a VAE (Variational Autoencoder) and CGAN (Conditional Generative Adversarial Network) models to generate synthetic chatter signals, addressing the challenge of imbalanced data in turning operations. Compared othe performance of synthetic chatter signals.
DistributedFraudDetection
In this study, we propose to use a distributed storage and computation system in order to track money transfers instantly. In particular, we keep our transaction history in a distributed file system as a graph data structure. We try to detect illegal activities by using Graph Neural Networks (GNN) in distributed manner.
EBUG-Entity_Based_User_Geolocation_on_Twitter
In this study, we propose a geolocation method based on named entities that are present in user tweets. We try to predict the US states where Twitter users reside in. We present a base method, compare it with a common approach in literature, and also examine some design decisions about our method.
ImageSimilarityEngine
An image similarity search engine has been developed that finds images similar to the image selected by the user from the dataset provided by the user. The latent vectors of the images are generated by a CNN based autoencoder model. KNN is used to find images similar to the selected image. An image similarity search engine platform has been created
ModelExplainability_TurkishSentimentAnalysis
An explainability model that can be applied to BERT-based Turkish sentiment analysis models has been developed and its performance has been compared with model spesific Layer-wise relevance propogation expailanbility model of Hila Chefer.
Survey-TextMining-for-SpamDetection
This study investigates the use of text mining techniques in spam detection. The study summarises the objectives, text pre-processing, feature extraction and classification methods, the datasets used, the results obtained and the findings of the studies conducted in the detection of comment , email & SMS spams on social network platforms.
TUBITAK-1711-CANBus_OptimalFS
This project includes sub-studies conducted as part of the ongoing 'Predictive Maintenance Prediction of Vehicle Data with Artificial Intelligence and Machine Learning Methods and Smart Spare Parts Warehouse Management Project', which is supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK).
TUBITAK-1711-PredictiveMaintanance
This project includes sub-studies conducted as part of the ongoing 'Predictive Maintenance Prediction of Vehicle Data with Artificial Intelligence and Machine Learning Methods and Smart Spare Parts Warehouse Management Project', which is supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK).
ayseirmak's Repositories
ayseirmak/Chatter-Detection
Utilized a VAE (Variational Autoencoder) and CGAN (Conditional Generative Adversarial Network) models to generate synthetic chatter signals, addressing the challenge of imbalanced data in turning operations. Compared othe performance of synthetic chatter signals.
ayseirmak/DistributedFraudDetection
In this study, we propose to use a distributed storage and computation system in order to track money transfers instantly. In particular, we keep our transaction history in a distributed file system as a graph data structure. We try to detect illegal activities by using Graph Neural Networks (GNN) in distributed manner.
ayseirmak/TUBITAK-1711-PredictiveMaintanance
This project includes sub-studies conducted as part of the ongoing 'Predictive Maintenance Prediction of Vehicle Data with Artificial Intelligence and Machine Learning Methods and Smart Spare Parts Warehouse Management Project', which is supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK).
ayseirmak/EBUG-Entity_Based_User_Geolocation_on_Twitter
In this study, we propose a geolocation method based on named entities that are present in user tweets. We try to predict the US states where Twitter users reside in. We present a base method, compare it with a common approach in literature, and also examine some design decisions about our method.
ayseirmak/ImageSimilarityEngine
An image similarity search engine has been developed that finds images similar to the image selected by the user from the dataset provided by the user. The latent vectors of the images are generated by a CNN based autoencoder model. KNN is used to find images similar to the selected image. An image similarity search engine platform has been created
ayseirmak/ModelExplainability_TurkishSentimentAnalysis
An explainability model that can be applied to BERT-based Turkish sentiment analysis models has been developed and its performance has been compared with model spesific Layer-wise relevance propogation expailanbility model of Hila Chefer.
ayseirmak/Survey-TextMining-for-SpamDetection
This study investigates the use of text mining techniques in spam detection. The study summarises the objectives, text pre-processing, feature extraction and classification methods, the datasets used, the results obtained and the findings of the studies conducted in the detection of comment , email & SMS spams on social network platforms.
ayseirmak/TUBITAK-1711-CANBus_OptimalFS
This project includes sub-studies conducted as part of the ongoing 'Predictive Maintenance Prediction of Vehicle Data with Artificial Intelligence and Machine Learning Methods and Smart Spare Parts Warehouse Management Project', which is supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK).