Creating a model of the document representation allowing to improve the accuracy of the automated document classification
April 2021 – May 2022
The research is aimed at further improvement of the document classification framework developed earlier (see below the project “Creating a model for automated document classification, indexing and routing”). The research will be carried out in three directions:
- The document feature model will be upgraded to include the representation of the document structure
- The framework will be expanded to incorporate the means of cluster analysis of the historic documents aimed at creating a stable document classes
- The framework will be expanded to deal with both digital and scanned documents
The project is successfully completed.
- Rats, J. and Pede, I. (2022). Supporting Trainset Annotation for Text Classification of Incoming Enterprise Documents. In Proceedings of the 11th International Conference on Data Science, Technology and Applications, ISBN 978-989-758-583-8, ISSN 2184-285X, pages 211-218. DOI: 10.5220/0011113000003269
A smart assistant for a user support
September 2019 – May 2021
The aim of the research project is to create a smart assistant service for the software product support team to deal with the frequent problems and questions. The smart assistant would act as a flexible mediator between a user and a support specialist. The user support process thus is represented as an interaction between three involved actors – the User, the (support) Specialist and the Assistant (software) .
- J. Rāts, I. Pede. Structuring and Controlling the Knowledge for the Software User Support. Baltic J. Modern Computing, Vol. 9 (2021), No. 2.
- J. Rāts, I. Pede. Using a Context Based Knowledge for Software Product User Support. 2020 61st International Scientific Conference on Information Technology and Management Science of Riga Technical University, ITMS 2020 – Proceedings, 2020, 9259307.
Creating a model for automated document classification, indexing and routing
April 2019 – April 2020
The aim of the research project is to create a model for smart capturing of the digital documents. Smart capturing comprises document automated categorization according to one or several categories (document type, customer etc.), document indexing with appropriate metadata (document number, creation date, person names involved etc.) and routing to a proper assignee, role or workflow.
- J. Rāts, I. Pede, T. Rubina, G. Vītols. A Machine Learning Based Framework for Enterprise Document Classification. Springer LNBIB, ICEIS 2020: Enterprise Information Systems, vol. 417, p. 87-99.
- J. Rāts, I. Pede. Document Capturing Automation: Case Study. Baltic J. Modern Computing, Vol. 8 (2020), No. 2, 259-274
- J. Rāts, I. Pede, T. Rubina, G. Vītols. A Flexible Model for Enterprise Document Capturing Automation. International Conference on Enterprise Information Systems (ICEIS) 5-7 May 2020, Vol. 1, p. 297-304.
Analysis of hybrid model of relational/NoSQL data persistence to enable effective and flexible access to both current and history data
November 2017 – October 2018
The research targets a creation of hybrid model for data persistence spanning a relational data model to handle current data and NoSQL data model for handling of history data. The options for usage of the model in Enterprise Content Management systems are to be analysed and the prototype is to be created and evaluated.
- J. Rats. Developing and Evaluating ECM Data Persistence Architecture. Computing and Informatics, Vol 38 (2019 ), p. 454-472, doi: 10.31577/ cai 2019 2 454
- J. Rats. Polyglot Persistence Architecture for Enterprise Content Management. Baltic Journal of Modern Computing, Vol. 6 (2018), No. 3, p. 304-319 https://doi.org/10.22364/bjmc.2018.6.3.06
Defining a dynamic user access rights data model and creating and analyzing an effective and safe full-text search algorithm based on the model
August 2016 – October 2017
The research aims to explore the options provided by NoSQL technologies to create data search algorithms based on paradigm of flexible user access rights model and to develop safe and effective algorithm for information search, tuned for best use with Enterprise Content managent systems.
- J. Rats. Use of NoSQL Technology for Secure and Fast Search of Enterprise Data. Baltic Journal of Modern Computing, vol. 5 (2017) No. 2, p. 147-163, DI 10.22364/bjmc.2017.5.2.01, UT WOS:000405914500001.
- J. Rats. Optimizing the Enterprise Search. Third International Conference on Mathematics and Computers in Sciences and in Industry (MCSI), Corfu, Greece, August 27-29, 2017.
The research of the options for the automated analysis and compliance monitoring of regulatory documents and business processes
May 2014 – December 2015
Options of the automated analysis of the regulatory documents are researched to create a methodology and a prototype for the automated analysis of the regulatory documents and compliance management of the business process descriptions against the requirements of the regulatory documents.
- I. Buksa. Business processes and regulations compliance management technology. IEEE 22nd International Requirements Engineering Conference (RE), 2014, Karlskrona, 25-29.08.2014, p. 489-493. DOI: 10.1109/RE.2014.6912304.
- Survilo, L., Bukša, I., Gaidukovs, A., Shendryk, V. Analysis of Questionnaire Results in the Use of BP and ND in Public Administration. Applied Computer Systems. Vol.16, 2014, pp.97-104. ISSN 2255-8683. e-ISSN 2255-8691. Available from: doi:10.1515/acss-2014-0019, skat.: https://ortus.rtu.lv/science/en/publications/19857/fulltext.pdf
The research of Advanced Visualization methods for the analysis of linked Big data
October 2013 – December 2014
The research intents to:
- create visualization models effectively usable for data analysis;
- define architecture for the components of advanced visualization system – clustering technology, cloud computing and schemaless NoSQL data store;
- benchmark the usage of the architecture and models created for a range of data volumes and data flow velocities.
- J. Rats. NoSQL document datastore as a backend of the visualization platform for ECM system. WSEAS 9th International Conference on Computer Engineering and Applications (CEA ’15), Dubai, United Arab Emirates, February 22-24, 2015, p. 354-359
The research of the automation problems of the business processes for effective management of Big Data
January 2013 – September 2013
The research deals with creation of the architecture and methodology for development of the mentioned architecture and user interface for Big data Enterprise content management systems, featuring:
- clustering technologies, cloud computing and NoSQL data stores, to provide interactive use of document-oriented software systems;
- multi-agent systems for parallel processing of the data;
- methods of ranked grouping of data.
- J. Rats, G. Ernestsons. Using of Cloud Computing, Clustering and Document-oriented Database for Enterprise Content Management. The Second International Conference on Informatics & Applications (ICIA2013), Lodz University of Technology, Poland, September 23-25, 2013, p. 72-76.
- Juris Rats, Gints Ernestsons. Clustering and Ranked Search for Enterprise Content Management. International Journal of E-Entrepreneurship and Innovation, Volume 4, Issue 4. Copyright © 2013. p. 20-31
Research is a part of the IT Competence Centre project.