**4. Proposed architecture of a knowledge base of natural disasters**

The proposed architecture is illustrated in **Figure 2**. The conceptual architecture is made up of four parts:

The first part is the Information gathering; this part uses GIS to map areas and so uses electronic sensors to collect information. It uses a set of sensors installed in

**Figure 2.** *Architecture of a knowledge base of natural disasters.*

#### *Decision Making in the Context of Natural Disasters Based on a Geographic Information System… DOI: http://dx.doi.org/10.5772/intechopen.98778*

well-defined places in the structure and in the Soil. These sensors can communicate with each other. The collection of information also uses a documentary warehouse for the collection of documents (expert report, opinions, etc.). These documents are annotated by the ontology in order to extract additional information. Documentary warehouse allows experts to record the events produced in order to make the right decisions. In the previous work [23, 24] we have demonstrated the importance of annotating documents handled during the process of decision making.

The second part is the knowledge base, it is used to store the information sent automatically by the electronic sensors (high water level, vibration, temperature, humidity, rotation, acceleration, inclination, force, etc.) and linked this information to graphical information mapped by GIS and information annotated by experts. **Figure 3** gives the class diagram of the knowledge base. The validation phase for the classification of knowledge according to the criteria of value and veracity of knowledge for the group is important. According to [25] there are three categories of knowledge among the knowledge to be certified (validated, not validated and rejected), we used these three categories as annotative acts between the actors of the group to certify and validate their knowledge and the data collected. In this function the following annotative acts are used: a validated annotation, a rejected annotation and an invalidated annotation.

The third part is the OSSIWAE ontology. The ontology used to conceptualize the interaction between structure and soil. The ontology is used to semantically capture the information sent by the sensors and also to annotate the documents stored in the document repository. The OSSIWAE ontology described in Section 3.1.

The fourth part is the user interface, which allows users to view the evolution in real time of the graphs for each mapped area and send them the Short Message (SMS) of notifications to mobile phones for decision makers to inform themselves

**Figure 3.** *The class diagram of the knowledge base.*

and take the necessary measures. And also allows fully understanding, analyzing and estimating this phenomenon by experts.
