Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Forecasting Trajectory Data: A study by Experimentation
Blekinge Institute of Technology, Faculty of Computing, Department of Communication Systems.
2017 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Context. The advances in location-acquisition and mobile computing techniques have generated massive spatial trajectory data. Such spatial trajectory data accumulated by telecommunication operators is huge, analyzing the data with a right tool or method can uncover patterns and connections which can be used for improving telecom services. Forecasting trajectory data or predicting next location of users is one of such analysis. It can be used for producing synthetic data and also to determine the network capacity needed for a cell tower in future.

Objectives. The objectives of this thesis is, Firstly, to have a new application for CWT (Collapsed Weighted Tensor) method. Secondly, to modify the CWT method to predict the location of a user. Thirdly, to provide a suitable method for the given Telenor dataset to predict the user’s location over a period of time.

 

Methods. The thesis work has been carried out by implementing the modified CWT method. The predicted location obtained by modified CWT cannot be determined to which time stamp it belongs as the given Telenor dataset contains missing time stamps. So, the modified CWT method is implemented in two different methods.

  1. Replacing missing values with first value in dataset.
  2. Replacing missing values with second value in dataset.

These two methods are implemented and determined which method can predict the location of users with minimal error.

 

Results. The results are carried by assuming that the given Telenor dataset for one week will be same as that for the next week. Users are selected in a random sample and above mentioned methods are performed. Furthermore, RMSD values and computational time are calculated for each method and selected users.

 

Conclusion. Based on the analysis of the results, Firstly, it can be concluded that CWT method have been modified and used for predicting the user’s location for next time stamp. Secondly, the method can be extended to predict over a period of time. Finally, modified CWT method predicts location of the user with minimal error when missing values are replaced by first value in the dataset.

Place, publisher, year, edition, pages
2017. , 32 p.
Keyword [en]
Collapsed weighted tensor method, periodic temporal link prediction, Trajectory data.
National Category
Telecommunications
Identifiers
URN: urn:nbn:se:bth-13976OAI: oai:DiVA.org:bth-13976DiVA: diva2:1078462
External cooperation
Telenor, Sweden
Subject / course
ET2580 Master's Thesis (120 credits) in Electrical Engineering with emphasis on Telecommunication Systems
Educational program
ETATX Master of Science Programme in Electrical Engineering with emphasis on Telecommunication Systems
Presentation
2017-01-25, Aristoteles J3423, Blekinge Tekniska Högskola, Karlskrona, 14:45 (English)
Supervisors
Examiners
Available from: 2017-03-17 Created: 2017-03-03 Last updated: 2017-03-17Bibliographically approved

Open Access in DiVA

fulltext(636 kB)40 downloads
File information
File name FULLTEXT02.pdfFile size 636 kBChecksum SHA-512
cdd0aac2804e13f4ba1492c843633d3bbf9ce3dde0e19e10e5cc556de2e6c28431d7e0850cfe4f6c02d0f5666d29efd9b0f3536c844bc84275252e64b38c6dc0
Type fulltextMimetype application/pdf

By organisation
Department of Communication Systems
Telecommunications

Search outside of DiVA

GoogleGoogle Scholar
Total: 40 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

Total: 38 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf