Yeran Sun

Road-based travel recommendation using geo-tagged images

Abstract Geotagged photos on social media like Flickr explicitly indicate the trajectories of tourists. They can be employed to reveal the tourists’ preference on landmarks and routings of tourism. Most of existing works on routing searches are based on the trajectories of GPS-enabled devices’ users. From a distinct point of view, we attempt to propose a novel approach in which the basic unit of routing is separate road segment instead of GPS trajectory segment. In this paper, we build a recommendation system that provides users with the most popular landmarks as well as the best travel routings between the landmarks. By using Flickr geotaggged photos, the top ranking travel destinations in a city can be identified and then the best travel routes between the popular travel destinations are recommended. We apply a spatial clustering method to identify the main travel landmarks and subsequently rank these landmarks. Using machine learning method, we calculate the tourism popularity of the road in terms of relevant parameters, e.g., the number of users and the number of Point-of-Interests. These popularity assessments are integrated into the routing recommendation system. The routing recommendation system takes into consideration both the popularity assessment and the length of the road. The best route recommended to the user minimizes the distance while including maximal tourism popularity. Experiments were conducted in two different scenarios. The empirical results show that the recommendation system is able to provide the user good travel planning including both top ranking landmarks and suitable routings in a city. Besides, the system offers user-generated semantic information for the recommended routes.

Analyzing Human Activities Through Volunteered Geographic Information: Using Flickr to Analyze Spatial and Temporal Pattern of Tourist Accommodation

Volunteered Geographic Information (VGI) provides valuable information to analyze human activities in space and time. In this chapter, we use Flickr photos as an example to explore the possibilities of VGI to analyze spatiotemporal patterns of tourists’ accommodation in Vienna, Austria as study site. Kernel density estimations and spatial scan statistics are used to explore the distribution of photos, while seasonality is considered additionally. The results show seasonal tendency of tourists for accommodation. It has been discovered that Flickr photos have, in general, the capability to improve tourism-related researches. In particular, they are useful to investigate spatiotemporal human activities, which open new possibilities for further location and event based analysis.

Identifying the city center using human travel flows generated from location-based social networking data:

Since cities have become more complex and some large cities are likely to be polycentric, a better understanding of cities requires a clear topology that reveals how city centers are spatially distributed and interacted with. The identification of a city center that aims to find the accurate location of the city center or delineate the city center with a precise boundary becomes vital. This work attempts to achieve this by using a new type of movement data generated from location-based social networks, whereby three different methods are deployed for clustering and compared regarding identification of city centers and delineation of their boundaries. Experiments show that city centers with precise boundaries can be identified by using the proposed approach with location-based social network data. Furthermore, the results show that the three methods for clustering have different advantages and disadvantages during the process of city center identification, and thus seem to be suitable for cities with different urban structures.

Comparison between CBR and CA methods for estimating land use change in Dongguan, China

Many studies on land use change (LUC), using different approaches and models, have yielded good results. Applications of these methods have revealed both advantages and limitations. However, LUC is a complex problem due to influences of many factors, and variations in policy and natural conditions. Hence, the characteristics and regional suitability of different methods require further research, and comparison of typical approaches is required. Since the late 1980s, CA has been used to simulate urban growth, urban sprawl and land use evolution successfully. Nowadays it is very popular in resolving the LUC estimating problem. Case-based reasoning (CBR), as an artificial intelligence technology, has also been employed to study LUC by some researchers since the 2000s. More and more researchers used the CBR method in the study of LUC. The CA approach is a mathematical system constructed from many typical simple components, which together are capable of simulating complex behavior, while CBR is a problem-oriented analysis method to solve geographic problems, particularly when the driving mechanisms of geographic processes are not yet understood fully. These two methods were completely different in the LUC research. Thus, in this paper, based on the enhanced CBR model, which is proposed in our previous research (Du et al. 2009), a comparison between the CBR and CA approaches to assessing LUC is presented. LUC in Dongguan coastal region, China is investigated. Applications of the improved CBR and the cellular automata (CA) to the study area, produce results demonstrating a similarity estimation accuracy of 89% from the improved CBR, and 70.7% accuracy from the CA. From the results, we can see that the accuracies of the CA and CBR approaches are both >70%. Although CA method has the distinct advantage in predicting the urban type, CBR method has the obvious tendency in predicting non-urban type. Considering the entire analytical process, the preprocessing workload in CBR is less than that of the CA approach. As such, it could be concluded that the CBR approach is more flexible and practically useful than the CA approach for estimating land use change.

Comparison of CBR and SVM Method Used in the Prediction of Land Use Change in Pearl River Delta, China

Many methods have been employed to study Land Use Change (LUC) in different areas, including some new algorithms from Artificial Intelligence (AI) field, such as Case-Based Reasoning (CBR), Artificial Neural Network (ANN), Bayesian Network (BN) and Support Vector Machine (SVM). Applications of some new methods have indicated both advantages and limitations. This paper presents a comparison between CBR and SVM methods, both of them are used to predict the LUC in Pearl River Delta, China in this study. The comparison is made in three respects: estimation accuracy, flexibility and efficiency. The experimental results demonstrate that CBR and SVM are both effective approaches to predict the LUC with the overall accuracy of 80% and 84% respectively. According to the statistical results, when considering all the changed land use categories, CBR is more stable than SVM for LUC estimation in this study. In addition, a complementary experiment for CBR approach is carried out to compare the estimation accuracies of these two methods in the absence of character data. The results indicate that SVM performances much better than CBR method without character data. Considering the running time, the choice of CBR or SVM method for LUC estimation should be based on the number of samples and variables.