Figures & data
Figure 1. Overall framework.
Figure 2. Knowledge graph construction for virtual landslide disaster environments.
Figure 3. Conceptual hierarchy of virtual landslide disaster environment ontology.
Table 1. User characteristic analysis.
Table 2. Scene characteristic analysis.
Table 3. Data characteristic analysis.
Table 4. Semantic relationship analysis.
Figure 4. Personalized landslide disaster scene data recommendation process.
Figure 5. Principle of the RippleNet algorithm for personalized recommendations.
Figure 6. Case area: a landslide disaster in XiaoGangJian, Sichuan Province, China.
Figure 7. Framework of the prototype system.
Figure 8. The knowledge graph constructed for the experiment.
Figure 9. The prototype system interface.
Figure 10. Recommendation and selection results for an ordinary people user type: (a) recommendation results, (b) selected results.
Figure 11. Recommendation and selection results for a disaster victims user type: (a) recommendation results, (b) selected results.
Figure 12. Recommendation and selection results for a rescue teams member user type: (a) recommendation results, (b) selected results.
Figure 13. Recommendation and selection results for a guidance experts user type: (a) recommendation results, (b) selected results.
Figure 14. Recommendation accuracy trends during the experiment.
