Personalizacija usluga zasnovana na obradi velike količine podataka iz telekomunikacijske mreže

master's thesis

• Title (english): Service personalization based on processing big data from telecommunication network
• Author: Antonija Mušćet
• Mentor: Vedran Podobnik (mentor)
• Committee member: Vedran Podobnik (predsjednik povjerenstva)
• Committee member: Ignac Lovrek (član povjerenstva)
• Committee member: Marina Bagić Babac (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2016-07-08, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

Budući da se danas količina podataka sve više povećava, javlja se potreba za njihovom efikasnom analizom, alatima i tehnikama s kojima bi se ta analiza provela, kao i načinima vizualizacije podataka. Ovaj diplomski rad daje jedno od mogućih rješenja izvedbe tog procesa na temelju slučaja korištenja koji uključuje personaliziran prijedlog kategorije iz koje bi se osobi mogao pokloniti poklon na temelju obrade velike količine podataka iz telekomunikacijske mreže. Poseban naglasak stavlja na odabir programskog okvira za provedbu analize te razvoja algoritma prikladnog za dobivanje relevantnog znanja. Informacije potrebne za vršenje analize dobivene su iz telekomunikacijske domene, a njihovim detaljnim proučavanjem i postavljanjem u kontekst dobiveni su vrijedni podaci iz kojih se izvuklo znanje. Proučavanjem danog skupa podataka pokazano je da ako ne postoji vremenska longitudinalnosti velikih količina podataka, tada je potrebna sadržajna vertikalnost podataka. Uz to, dokazano je da prosječna potrošnja korisnika kao i njegova prosječna vremenska duljina zapisa imaju long tail distribuciju. Na kraju, je implementiran sustav kojeg je kao komponentu moguće koristiti kao dodatan ulaz personaliziranim uslugama.

Abstract (english)

Since nowadays the amount of data increases more and more, there is a need for its efficient analysis, tools and techniques with which such analysis would be conducted, as well as the way results, should be visualized. The thesis outlines one possible solution to these challenges on a use case that involves implementing a personalized system that would display possible categories from which to choose a category of gift for each user based on the data from the telecommunication network. The main emphasis is placed on the selection of programming framework for the implementation of the analysis and development of the algorithm suitable for providing the relevant knowledge. Information required for the analysis were obtained from the telecommunication domain, and because of its detailed study as well as building the context on top of it, valuable data was obtained from which knowledge could be extracted. By analyzing given dataset, it has been shown that if there is not a wide data time longitude, then a wide data content vertical is needed. In addition, the research showed that user consumption of traffic, as well as the user connection duration, have a long tail distribution. At the end, a system was implemented which can be a component that can be used as an additional input to the personalized services.

• Keywords: obrada velikih količina podataka; Hadoop; Spark; Scala; OpenCellID API; vizualizacija velikih podataka; CDR; personalizirana usluga
• Keywords (english): Big Data analysis; Hadoop framework; Spark framework; Scala programming language; OpenCellID API; CDR dataset; Big Data visualization
• Language: english
• URN:NBN: urn:nbn:hr:168:585489
• Study programme: Title: Information and Communication Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije (magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 10:51:53