I am a data scientist with 3+ years of experience at Mantis Software Company. I received my Ph.D. degree from the Department of Computer Engineering at Hacettepe University. I am a member of Hacettepe University Multimedia Information Retrieval laboratory (HUMIR). During my academic career, I was immersed in the fields of Information Retrieval, Machine Learning and Text Mining. I also published my works on the imbalanced data problem in machine learning algorithms. I continued my research on sentiment analysis, task-specific word embedding, and deep learning models through national and international academic projects. To make large-scale data processing, I joined international industry-driven R&D&I projects (ITEA/Eureka) and incorporated the academic solutions into the industry problems. My main research goals in this area focus on the health and education problems, such as Health Recommendation Systems, Big Data Analytics, and Social Media Analysis.
I have accomplished studies on the term evaluation metrics, imbalanced document indexing, word sense disambiguation, semantic analytics, and lexical search engines. I am also interested in semantic search engines and concept-oriented text retrieval. In some domains, semantic information retrieval plays an important role in providing access to information, such as health and biomedical domain.
I have studied word semantics, word embeddings, neural language models, probabilistic topic models, sentiment analysis, word sense disambiguation, and text classification. I am also interested in compositional semantics and task-specific/knowledge-based word embeddings.
In the context of deep learning, I have worked with RNN, CNN, LSTM, and BiLSTM models. My research objective in this area is to exploit different deep models in multistream networks to learn complex events in the text. In addition, I am interested in seq2seq models with encoder-decoder architecture to learn compositional semantics.
In this area, I have studied several distributed Big data analytics systems. I have worked with Solr Search, Elasticsearch, MongoDB, and Cassandra database and also familiar with Hbase, Druid, OpenTSDB, and OrientDB. In addition, I am interested in modern health data monitoring systems with near real-time data processing.
full list at my Google Scholar Profile
@article{naderalvojoud2020sentiment,
title={Sentiment Aware Word Embeddings Using Refinement and Senti-Contextualized Learning Approach},
author={Naderalvojoud, Behzad and Sezer, Ebru Akcapinar},
journal={Neurocomputing},
volume={405},
pages={149--160},
year={2020},
publisher={Elsevier}
}
@article{naderalvojoud2020term,
title={Term evaluation metrics in imbalanced text categorization},
author={Naderalvojoud, Behzad and Sezer, Ebru Akcapinar},
journal={Natural Language Engineering},
volume={26},
number={1},
pages={31--47},
year={2019},
publisher={Cambridge University Press}
}
@inproceedings{naderalvojoud2018humir,
title={HUMIR at IEST-2018: Lexicon-Sensitive and Left-Right Context-Sensitive BiLSTM for Implicit Emotion Recognition},
author={Naderalvojoud, Behzad and Ucan, Alaettin and Sezer, Ebru Akcapinar},
booktitle={Proceedings of the 9th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis},
pages={182--188},
year={2018},
publisher = {Association for Computational Linguistics}
}
@inproceedings{naderalvojoud2018cross,
title={A cross-lingual approach for building multilingual sentiment lexicons},
author={Naderalvojoud, Behzad and Qasemizadeh, Behrang and Kallmeyer, Laura and Sezer, Ebru Akcapinar},
booktitle={International Conference on Text, Speech, and Dialogue},
pages={259--266},
year={2018},
organization={Springer}
}
@inproceedings{germevaltask2017,
title = {{GermEval 2017: Shared Task on Aspect-based Sentiment in Social Media Customer Feedback}},
author = {Naderalvojoud, Behzad, and Qasemizadeh, Behrang and Kallmeyer, Laura},
year = {2017},
booktitle = {Proceedings of the GermEval 2017: Shared Task on Aspect-based Sentiment in Social Media Customer Feedback},
address={Berlin, Germany},
pages={18--21}
}
@inproceedings{ucan2016sentiwordnet,
title={SentiWordNet for new language: automatic translation approach},
author={Ucan, Alaettin and Naderalvojoud, Behzad and Sezer, Ebru Akcapinar and Sever, Hayri},
booktitle={2016 12th International Conference on Signal-Image Technology & Internet-Based Systems (SITIS)},
pages={308--315},
year={2016},
organization={IEEE}
}
@inproceedings{naderalvojoud2015imbalanced,
title={Imbalanced text categorization based on positive and negative term weighting approach},
author={Naderalvojoud, Behzad and Sezer, Ebru Akcapinar and Ucan, Alaettin},
booktitle={International Conference on Text, Speech, and Dialogue},
pages={325--333},
year={2015},
organization={Springer}
}
@inproceedings{naderalvojoud2014investigation,
title={Investigation of term weighting schemes in classification of imbalanced texts},
author={Naderalvojoud, Behzad and Bozkir, Ahmet Selman and Sezer, Ebru Akcapinar},
booktitle={Proceedings of European Conference on Data Mining (ECDM)},
pages={39--46},
year={2014}
}
@mastersthesis{mastersthesis,
author={Naderalvojoud, Behzad},
title={Investigation of Imbalance Problem Effects on Text Categorization},
school={Hacettepe University, Graduate School of Science and Engineering},
year={2015},
address={Turkey},
month = 1
}
My main role in this project is to develop a graph-based representation for scheduling and controlling industrial projects.
My main role in this project is to propose a cross-lingual distributional semantic model to create a sense-based sentiment lexicon for Turkish language using bilingual movie subtitles.
My main role in this project is to incorporate the GPU-based string matching algorithms into word embedding models for non-sequential refining of particular word vectors.
This project provides a system for the analysis of news in the media.
This project provides a journal citation analysis system for TRDizin and implements all standard evaluation metrics for TRDizin journals.
In this project, I propose a machine learning based recommendation system for assigning subject headings and subject terms to academic papers to be used in search engines when indexing papers. This system is used in TRDizin.
This project determines academic collaboration networks between authors, countries, institutions, and subject areas using citation and reference information.
The main goal of this project is to empower people to monitor and improve their health using personal data and technology-assisted coaching. My role in this project is to propose a data monitoring and analytics architecture for health recommendation systems as well as a data-driven coaching engine to provide intelligent models for recognizing health problems, exacerbations and patients at risk.
This project proposes a hybrid policy-making model based on data-driven, model-driven, and societal-driven sources. My role in this project is to propose a social media monitoring and analytics system using topic-based sentiment analysis models.
Research Fellowship
Best Sentiment Model
Research Scholarship
Research Scholarship
Research Scholarship
behzadn@mantis.com.tr
n.behzad@hacettepe.edu.tr