A Collaborative Learning Framework to Tag Refinement for Points of Interest

Zhou, Jingbo; Gou, Shan; Hu, Ruifeng; Zhang, Dongxiang; Xu, Jin; Jiang, Airong; Liu, Ying; Xiong, Hui

doi:10.1145/3292500.3330698

Cited by 30 publications

(20 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many times, multiple tags are available for the same data hence it is important to remove the duplicate information. The author suggested a non-negative matrix factorization method based on the maximum likelihood estimation to find the similarity between different labels assigned to the same POI [8]. Researchers work on the POI data organization using multidimensional ranking organization to arrange the POI data which can be retrieved easily during the recommendation process [9].…”

Section: A Poi Data Verificationmentioning

confidence: 99%

Fine-tuned Predictive Model for Verifying POI Data

Monika¹,

Bundele²,

Bothale³

et al. 2021

IJACSA

View full text Add to dashboard Cite

Mapping websites and geo portals are playing a vital role in daily life due to the availability of geo-tagged data. From booking a cab to search a place, getting traffic information, review of the place, searching for a doctor or best school available in the locality, we are heavily dependent on the map services and geo portals available for finding such information. There is voluminous data available on these sources and it is getting increasing every moment. These data are majorly collected through crowdsourcing methods where people are contributing. As a basic principle of Garbage in garbage out, the quality of this data impacts the quality of the services based on this data. Therefore, it is highly desired to have a model which can predict the quality/accuracy of the geotagged Point of interest data. We propose a novel Fine-Tuned Predictive Model to check the accuracy of this data using the best suitable supervised machine learning approach. This work focuses on the complete life cycle of the model building, starting from the data collection to the fine-tuning of the hyperparameters. We covered the challenges particularly to the geotagged POI data and remedies to resolve the issues to make it suitable for predictive modeling for classifying the data based on their accuracy. This is a unique work that considers multiple sources including ground truth data to verify the geotagged data using a machine learning approach. After exhaustive experiments, we obtained the best values for hyperparameters for the selected predictive model built on the real data set prepared specifically to target the proposed solution. This work provides a way to develop a robust pipeline for predicting the accuracy of crowdsourced geotagged data.

show abstract

Section: A Poi Data Verificationmentioning

confidence: 99%

Fine-tuned Predictive Model for Verifying POI Data

Monika¹,

Bundele²,

Bothale³

et al. 2021

IJACSA

View full text Add to dashboard Cite

show abstract

“…Category-aware POI Recommendation. Categories of POIs visited by a user often capture preferred activities, thus they are important indicators to model user preferences [16,27,41]. Liu et al [17] exploited the transition patterns of user preferences over location categories to enhance recommendation performance.…”

Section: Related Workmentioning

confidence: 99%

Spatial Object Recommendation with Hints: When Spatial Granularity Matters

Luo,

Zhou,

Bao

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Existing spatial object recommendation algorithms generally treat objects identically when ranking them. However, spatial objects often cover different levels of spatial granularity and thereby are heterogeneous. For example, one user may prefer to be recommended a region (say Manhattan), while another user might prefer a venue (say a restaurant). Even for the same user, preferences can change at different stages of data exploration. In this paper, we study how to support top-𝑘 spatial object recommendations at varying levels of spatial granularity, enabling spatial objects at varying granularity, such as a city, suburb, or building, as a Point of Interest (POI). To solve this problem, we propose the use of a POI tree, which captures spatial containment relationships between POIs. We design a novel multi-task learning model called MPR (short for Multi-level POI Recommendation), where each task aims to return the top-𝑘 POIs at a certain spatial granularity level. Each task consists of two subtasks: (i) attribute-based representation learning; (ii) interaction-based representation learning. The first subtask learns the feature representations for both users and POIs, capturing attributes directly from their profiles. The second subtask incorporates user-POI interactions into the model. Additionally, MPR can provide insights into why certain recommendations are being made to a user based on three types of hints: user-aspect, POIaspect, and interaction-aspect. We empirically validate our approach using two real-life datasets, and show promising performance improvements over several state-of-the-art methods.

show abstract

“…Geographical data. We use large-scale geographical datasets to build Polestar, including: (1) the transportation station data, (2) the transportation line data, (3) the road network data, and (4) the point of interest (POI) data [31]. All geographical data are collected from (i) professional surveyors employed by Baidu Maps, (ii) the crowdsourcing platform in Baidu.…”

Section: Data Description and Analysismentioning

confidence: 99%

Polestar: An Intelligent, Efficient and National-Wide Public Transportation Routing Engine

Liu,

Li,

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Public transportation plays a critical role in people's daily life. It has been proven that public transportation is more environmentally sustainable, e cient, and economical than any other forms of travel [3,15]. However, due to the increasing expansion of transportation networks and more complex travel situations, people are having di culties in e ciently nding the most preferred route from one place to another through public transportation systems. To this end, in this paper, we present Polestar, a data-driven engine for intelligent and e cient public transportation routing. Specically, we rst propose a novel Public Transportation Graph (PTG) to model public transportation system in terms of various travel costs, such as time or distance. en, we introduce a general route search algorithm coupled with an e cient station binding method for e cient route candidate generation. A er that, we propose a two-pass route candidate ranking module to capture user preferences under dynamic travel situations. Finally, experiments on two real-world data sets demonstrate the advantages of Polestar in terms of both e ciency and e ectiveness. Indeed, in early 2019, Polestar has been deployed on Baidu Maps, one of the world's largest map services. To date, Polestar is servicing over 330 cities, answers over a hundred millions of queries each day, and achieves substantial improvement of user click ratio.

show abstract

A Collaborative Learning Framework to Tag Refinement for Points of Interest

Cited by 30 publications

References 30 publications

Fine-tuned Predictive Model for Verifying POI Data

Fine-tuned Predictive Model for Verifying POI Data

Spatial Object Recommendation with Hints: When Spatial Granularity Matters

Polestar: An Intelligent, Efficient and National-Wide Public Transportation Routing Engine

Contact Info

Product

Resources

About