Projects / Programmes
Large networks in business system analytics
| Code |
Science |
Field |
Subfield |
| 7.00.00 |
Interdisciplinary research |
|
|
| Code |
Science |
Field |
| S189 |
Social sciences |
Organizational science |
| Code |
Science |
Field |
| 5.06 |
Social Sciences |
Political science |
Network analysis, two-mode network, multiplication of networks, generalized two-mode cores, semiring, business system, big data, statistical analysis.
Researchers (1)
| no. |
Code |
Name and surname |
Research area |
Role |
Period |
No. of publicationsNo. of publications |
| 1. |
33987 |
PhD Monika Cerinšek |
Computer intensive methods and applications |
Head |
2016 - 2019 |
21 |
Organisations (1)
Abstract
The research topic of the postdoctoral project is a part of a currently scientifically very important topic on a world scale - data analysis. Out of ten Nobel prizes for economy seven were awarded for work closely connected to data analysis. Trends that we can follow every day are also results of data analysis. Big data analysis, especially in the form of network analysis, has shown its potential when dealing with the Internet or the Internet of Things since both data sets are in a network format. We cannot comprehend such data by itself, as it is too big. However, with the help of network analysis we can track and change what is happening around us and consequently change the world.
During my PhD study I noticed that individual analysis approaches exclude each other as a consequence of too narrow expertise of the analysts. Analysts in the industry usually prefer data mining and classical statistical methods. The power of the classical approach to analysis is greatly increased when using methods and visualizations from network analysis. In my postdoctoral project I will focus my research on bringing together both worlds, which will lead to fundamental innovations and new standards in the field of big data analysis in the industry.
I will also continue developing new methods for direct two-mode network analysis. With this I will include data storage in the form of two-mode networks and new techniques for its analysis in big data. By introducing two-mode networks we also gain a special structure for data storage. This enables a direct introspection of two types of data that are related.
In my postdoctoral project I will expand the list of techniques for business system analysis with direct two-mode network analysis methods with emphasis on generalized two-mode cores and network multiplication with semirings. I will develop these two techniques because they show great promise for business system data analysis already in the theoretical stage. There are currently no comparable solutions. Because I have access to real world data in the company I will test the correctness and soundness of the methods along the way. The soundness will be controlled with modifications of the parameters of the methods; for instance with the selection of the semiring when multiplying. With these techniques I will immensely expand the spectre of techniques for data analysis since both methods allow parameter adjustment to adjust to current problems and also enable adding new parameters.
I will also develop a system for analysing business systems. This tool will enable analysts in our company to easily perform data analyses, which will lead to better company leadership and organisation for individual companies (the clients of analysis services). There are many tools available for data and network analysis (Pajek, SNAP, Orange, Weka, etc.), which mainly target researchers and do not allow the addition of own techniques. The main advantage of my system will be its focus on business system analysis. This system will consist of techniques that will show a lot of promise through my research work. I will join techniques from the fields of big data, predictive analytics, data mining, statistics, and network analysis. Individual techniques are often used in a sequence to obtain the desired results. These sequences will be joined in macros for the ease of use. The developed system will be built in a way that enables its quick adjustment to be used on data of a specific company and will thus be useful to any analyst who wants to use techniques from different fields of data analysis.
Significance for science
At the apex of network analysis research we can find many Slovenian researchers; the most visible are prof. dr. Vladimir Batagelj, prof. dr. Anuška Ferligoj in doc. dr. Jure Leskovec. Jure Leskovec was an external examiner in a commission for assessment of my doctoral thesis and works a lot with the industry at Stanford University in the USA (Twitter, LinkedIn, Volkswagen, Facebook, Boeing, Orange, Amazon, Samsung, etc.). Jure already showed the right path for transferring knowledge to the industry and this path is one I intend to follow.
My research subfield (two-mode network analysis) is on the rise (Special Issue on Advances in Two-mode Social Networks, Social Networks 35(2), 2013). Right now there are very few methods for the direct analysis of two-mode networks thus the development of every new technique greatly contributes to this scientific field. I will enrich these techniques with applications on real world data. During my doctoral studies I have established links with many bibliographic database researchers (for databases Web of Science and Zentralblatt MATH). Because we were researching similar problems we shared experiences among us, which means I already have access to foreign/outside knowledge. I will collaborate with these researchers in the future and with that I will help confirm Slovenia’s top place among network analysts. I will also promote Slovenia by attending international meetings and conferences from the fields of network analysis and big data, and by publishing in internationally renowned scientific journals.
Significance for the country
My research will have a direct effect on the company Abelium and its clients since the results of my work enriched with new techniques will be used to solve real world problems from the industry. The combination of techniques I will use for data analysis and knowledge of the analysed data will create a unique service for the analysis of business systems. With this the company will gain an important place among providers of data analysis services in the world and with it greater recognition.
Abelium is cooperating with the company TM Vista that owns the trademark GoOpti whose main product is low-cost transfer to airports. With help of my research we will transform it onto a data driven company and give its leadership the ability to make business decisions based on data from past performance and on prediction for the future and not on (often false) intuition. Since this company is well known in the field of low cost transfer the influence will spread to the entire industry since competitors will have to adjust. My work will contribute to better recognition of data analysis as an important branch of industry since companies will seek specific solutions to improve their performance. This will open possibilities for niche solutions from young start-ups.
I already use developed techniques for the direct analysis of two-mode networks in the combination with other techniques for the data analysis in project in which we research management of wood (CaReWood: http://carewood.eu/) in effort for better timber recycling or usage of waste wood. We use these techniques also in the analysis of European startups and investors ecosystem for the risk assessment and hybrid techniques that implement properties of egocentric networks of people and their companies.
