Please use this identifier to cite or link to this item:
https://rfos.fon.bg.ac.rs/handle/123456789/1499| Title: | Otkrivanje prevara u kasko osiguranju pomoću grafovskih baza podataka Fraud detection in car insurance using graph databases |
Authors: | Mitrović, Olivera Babarogić, Slađan Makajić-Nikolić, Dragana |
Keywords: | osiguranje;NoSQL;Neo4j;grafovske baze;NoSQL;Neo4j;insurance;graph database | Issue Date: | 2016 | Publisher: | Univerzitet u Beogradu - Fakultet organizacionih nauka, Beograd | Abstract: | Predmet istraživanja ovog rada predstavlja razvoj rešenja realnog problema pomoću grafovskih baza podataka. Ukratko su opisani osnovni koncepti grafovskih baza podataka. Ono što ovaj tip baza izdvaja, u odnosu na ostale NoSQL baze podataka, je osobina da podatke i veze među njima smatra ravnopravnim. Takođe, njihov način skladištenja podataka je najpribližniji ljudskom razmišljanju što ih čini veoma intuitivnim. U centralnom delu rada je dat predlog za otkrivanje pravara u kasko osiguranju korišćenjem Neo4j grafovske baze za skladištenje podataka. Pomenuti tip baza podataka za dati problem nudi jednostavno rešenje i bolje rezultate prostim prolaskom kroz graf. The main subject of this paper is the development of the solution to the real problem, using graph database. Elementary concepts of the graph databases are briefly explained. Unlike the other NoSQL databases, Graph database treats data and connections between data equally. Also, the way they store data is closest to human thinking, which makes them very intuitive. Central part of the paper presents the solution for the detection of frauds in car insurance by using Neo4j graph database for storing data. Graph database offers a simple solution and better results for this problem by simply going through the graph. |
URI: | https://rfos.fon.bg.ac.rs/handle/123456789/1499 | ISSN: | 1451-4397 |
| Appears in Collections: | Radovi istraživača / Researchers’ publications |
Show full item record
Google ScholarTM
Check
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.