TY  - JOUR
AU  - Kovačević, Ana
AU  - Devedžić, Vladan
AU  - Pocajt, Viktor
PY  - 2010
UR  - https://rhinosec.fb.bg.ac.rs/handle/123456789/85
AB  - Purpose This paper aims to propose a solution for recommending digital library services based on data mining techniques (clustering and predictive classification). Design/methodology/approach - Data mining techniques are used to recommend digital library services based on the user's profile and search history. First, similar users were clustered together, based on their profiles and search behavior. Then predictive classification for recommending appropriate services to them was used. It has been shown that users in the same cluster have a high probability of accepting similar services or their patterns. Findings - The results indicate that k-means clustering and Naive Bayes classification may be used to improve the accuracy of service recommendation. The overall accuracy is satisfying, while average accuracy depends on the specific service. The results were better for frequently occurring services. Research limitations/implications - Datasets were used from the KOBSON digital library. Only clustering and predictive classification was applied. If the correlation between the service and the institution were higher, it would have better accuracy. Originality/value - The paper applied different and efficient data mining techniques for clustering digital library users based on their profiles and their search behavior, i.e. users' interaction with library services, and obtain user patterns with respect to the library services they use. A digital library may apply this approach to offer appropriate services to new users more easily. The recommendations will be based on library items that similar users have already found useful.
PB  - Emerald Group Publishing Limited, Bingley
T2  - Electronic Library
T1  - Using data mining to improve digital library services
VL  - 28
IS  - 6
SP  - 829
EP  - 843
DO  - 10.1108/02640471011093525
ER  - 

@article{
author = "Kovačević, Ana and Devedžić, Vladan and Pocajt, Viktor",
year = "2010",
abstract = "Purpose This paper aims to propose a solution for recommending digital library services based on data mining techniques (clustering and predictive classification). Design/methodology/approach - Data mining techniques are used to recommend digital library services based on the user's profile and search history. First, similar users were clustered together, based on their profiles and search behavior. Then predictive classification for recommending appropriate services to them was used. It has been shown that users in the same cluster have a high probability of accepting similar services or their patterns. Findings - The results indicate that k-means clustering and Naive Bayes classification may be used to improve the accuracy of service recommendation. The overall accuracy is satisfying, while average accuracy depends on the specific service. The results were better for frequently occurring services. Research limitations/implications - Datasets were used from the KOBSON digital library. Only clustering and predictive classification was applied. If the correlation between the service and the institution were higher, it would have better accuracy. Originality/value - The paper applied different and efficient data mining techniques for clustering digital library users based on their profiles and their search behavior, i.e. users' interaction with library services, and obtain user patterns with respect to the library services they use. A digital library may apply this approach to offer appropriate services to new users more easily. The recommendations will be based on library items that similar users have already found useful.",
publisher = "Emerald Group Publishing Limited, Bingley",
journal = "Electronic Library",
title = "Using data mining to improve digital library services",
volume = "28",
number = "6",
pages = "829-843",
doi = "10.1108/02640471011093525"
}

Kovačević, A., Devedžić, V.,& Pocajt, V.. (2010). Using data mining to improve digital library services. in Electronic Library
Emerald Group Publishing Limited, Bingley., 28(6), 829-843.
https://doi.org/10.1108/02640471011093525

Kovačević A, Devedžić V, Pocajt V. Using data mining to improve digital library services. in Electronic Library. 2010;28(6):829-843.
doi:10.1108/02640471011093525 .

Kovačević, Ana, Devedžić, Vladan, Pocajt, Viktor, "Using data mining to improve digital library services" in Electronic Library, 28, no. 6 (2010):829-843,
https://doi.org/10.1108/02640471011093525 . .

TY  - JOUR
AU  - Kovačević, Ana
AU  - Devedžić, Vladan
AU  - Pocajt, Viktor
PY  - 2010
UR  - https://rhinosec.fb.bg.ac.rs/handle/123456789/84
AB  - Purpose - This paper aims to address the problem of enhancing the selection of titles offered by a digital library, by analysing the differences in these titles when they are cited by local authors in their publications and when they are listed in the digital library offer. Design/methodology/approach - Text mining techniques were used to identify duplicate references. Moreover, the process of identifying syntactically different data was improved with the automated discovery of thesauri from correctly matched data, and the generated thesaurus was further used in semantic clustering. The results were effectively visually represented. Findings - The paper finds that the function based on the Jaro-Winkler algorithm may be efficiently used in the de-duplication process. A generated thesaurus that utilises domain-specific knowledge can also be used in the semantic clustering of references. It was shown that semantic clustering may be most useful in partitioning data, which is particularly significant when dealing with large amounts of data, which is usually the case. Moreover, those references that have the same or similar scores may be considered as candidate matches in the further de-duplication process. Finally, it proved to be a more efficient way of visually representing the results. Originality/value - This function can be implemented to enhance the selection of titles to be offered by a digital library, in terms of making that offer more compliant with what the library users frequently cite.
PB  - Emerald Group Publishing Ltd, Bingley
T2  - Program-Electronic Library and Information Systems
T1  - Enhancing a core journal collection for digital libraries
VL  - 44
IS  - 2
SP  - 132
EP  - 148
DO  - 10.1108/00330331011039490
ER  - 

@article{
author = "Kovačević, Ana and Devedžić, Vladan and Pocajt, Viktor",
year = "2010",
abstract = "Purpose - This paper aims to address the problem of enhancing the selection of titles offered by a digital library, by analysing the differences in these titles when they are cited by local authors in their publications and when they are listed in the digital library offer. Design/methodology/approach - Text mining techniques were used to identify duplicate references. Moreover, the process of identifying syntactically different data was improved with the automated discovery of thesauri from correctly matched data, and the generated thesaurus was further used in semantic clustering. The results were effectively visually represented. Findings - The paper finds that the function based on the Jaro-Winkler algorithm may be efficiently used in the de-duplication process. A generated thesaurus that utilises domain-specific knowledge can also be used in the semantic clustering of references. It was shown that semantic clustering may be most useful in partitioning data, which is particularly significant when dealing with large amounts of data, which is usually the case. Moreover, those references that have the same or similar scores may be considered as candidate matches in the further de-duplication process. Finally, it proved to be a more efficient way of visually representing the results. Originality/value - This function can be implemented to enhance the selection of titles to be offered by a digital library, in terms of making that offer more compliant with what the library users frequently cite.",
publisher = "Emerald Group Publishing Ltd, Bingley",
journal = "Program-Electronic Library and Information Systems",
title = "Enhancing a core journal collection for digital libraries",
volume = "44",
number = "2",
pages = "132-148",
doi = "10.1108/00330331011039490"
}

Kovačević, A., Devedžić, V.,& Pocajt, V.. (2010). Enhancing a core journal collection for digital libraries. in Program-Electronic Library and Information Systems
Emerald Group Publishing Ltd, Bingley., 44(2), 132-148.
https://doi.org/10.1108/00330331011039490

Kovačević A, Devedžić V, Pocajt V. Enhancing a core journal collection for digital libraries. in Program-Electronic Library and Information Systems. 2010;44(2):132-148.
doi:10.1108/00330331011039490 .

Kovačević, Ana, Devedžić, Vladan, Pocajt, Viktor, "Enhancing a core journal collection for digital libraries" in Program-Electronic Library and Information Systems, 44, no. 2 (2010):132-148,
https://doi.org/10.1108/00330331011039490 . .

TY  - JOUR
AU  - Pocajt, Viktor
AU  - Ševarac, Zoran
AU  - Kovačević, Ana
PY  - 2009
UR  - https://rhinosec.fb.bg.ac.rs/handle/123456789/38
AB  - This paper presents a method of searching, identifying and cross-referencing metal alloys based on their chemical composition and/or mechanical properties, typically obtained by analysis and tests. The method uses a general pattern similar to the approach of a human expert, and relies on a classification of metals based on metallurgical expertise and fuzzy logic for identifying metals and comparing their chemical and mechanical properties. The algorithm has been tested and deployed in real applications for fast metal identification and finding of unknown equivalents, by the leading companies in the field. The same principles can also be used in other domains for similar problems, such as organic and inorganic materials identification and generic drugs comparison. Copyright (C) 2009 John Wiley & Sons, Ltd.
PB  - John Wiley & Sons Ltd, Chichester
T2  - Journal of Chemometrics
T1  - SmartMetals: a new method for metal identification based on fuzzy logic
VL  - 23
IS  - 11-12
SP  - 555
EP  - 561
DO  - 10.1002/cem.1251
ER  - 

@article{
author = "Pocajt, Viktor and Ševarac, Zoran and Kovačević, Ana",
year = "2009",
abstract = "This paper presents a method of searching, identifying and cross-referencing metal alloys based on their chemical composition and/or mechanical properties, typically obtained by analysis and tests. The method uses a general pattern similar to the approach of a human expert, and relies on a classification of metals based on metallurgical expertise and fuzzy logic for identifying metals and comparing their chemical and mechanical properties. The algorithm has been tested and deployed in real applications for fast metal identification and finding of unknown equivalents, by the leading companies in the field. The same principles can also be used in other domains for similar problems, such as organic and inorganic materials identification and generic drugs comparison. Copyright (C) 2009 John Wiley & Sons, Ltd.",
publisher = "John Wiley & Sons Ltd, Chichester",
journal = "Journal of Chemometrics",
title = "SmartMetals: a new method for metal identification based on fuzzy logic",
volume = "23",
number = "11-12",
pages = "555-561",
doi = "10.1002/cem.1251"
}

Pocajt, V., Ševarac, Z.,& Kovačević, A.. (2009). SmartMetals: a new method for metal identification based on fuzzy logic. in Journal of Chemometrics
John Wiley & Sons Ltd, Chichester., 23(11-12), 555-561.
https://doi.org/10.1002/cem.1251

Pocajt V, Ševarac Z, Kovačević A. SmartMetals: a new method for metal identification based on fuzzy logic. in Journal of Chemometrics. 2009;23(11-12):555-561.
doi:10.1002/cem.1251 .

Pocajt, Viktor, Ševarac, Zoran, Kovačević, Ana, "SmartMetals: a new method for metal identification based on fuzzy logic" in Journal of Chemometrics, 23, no. 11-12 (2009):555-561,
https://doi.org/10.1002/cem.1251 . .