|Title:||Research of POI data ecosystem and quality control models based on crowdsourced feedback|
Mobile geographic information systems.
Electronic data processing -- Quality control.
Hong Kong Polytechnic University -- Dissertations
|Department:||Faculty of Construction and Environment|
|Pages:||xi, 202 pages : color illustrations|
|Abstract:||With the development of big data application and smart phones, hundreds of millions of customers use POI to simplify and facilitate their lives. So POI production and quality control have been changed correspondingly. Considering these issues, the purpose of this dissertation is to improve POI data quality by combing POI production flow and introducing quality control models.Firstly, actual POIs were grabbed to find POI uncertainties and errors.After multisource comparison, POI errors were identified.Combining with ISO/TS16949 and ISO19113, POI errors were classified and analyzed by five quality elements. Then significant quality elements and error types were selected for further study by 80-20 theory. Currently, main POI quality research lays emphasize on similarity calculation. In this study, three kinds of similarity models were introduced to analyze, including edit distance related models, token-based models and semantic models. According to the features of POI address field, instance variants were put forward to enhance models. Thirdly, for monitoring each POI quality elements and data flow comprehensively, POI data ecosystem was established, which was also a conceptual model. After that, POI WordNet and location related models were built to deal with thematic and positional errors separately, which were introduced to improve the accuracy of current quality control methods. However, to cover and locate all significant POI errors, WordNet and Address Matching (WNAM) model was established that was the core of ecoenvironment in POI ecosystem. WNAM is a combination of specific rules, POI WordNet model and location related model, which is used to tackle with all error identification and error-type locating. Finally, after empirical analysis, it was proved that WNAM and POI WordNet model were the most optimal models in most cases. During this process, evaluation consisted of different dimensions, including based on data located region, data types and data error types. The first two aspects were studied by two kinds of analysis: to find real POI errors as many as possible; to simulate actual POI error rate. The last one focused on identifying most POI errors. In addition, owing to less manual intervention involved,WNAM model can be applied to improve current work efficiency of POI quality control and drop capital investment for POI data providers.|
Files in This Item:
|b28315431.pdf||For All Users (off-campus access for PolyU Staff & Students only)||5.04 MB||Adobe PDF||View/Open|
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: