Ontology Data Modeling of Indonesian Medicinal Plants and Efficacy

— People use medicinal plants for as early prevention and treating disease. Medicinal plants must be careful not to cause side effects, so knowledge is needed. Medicinal plant knowledge is stored using an ontology data model. In some ontology studies, there are still shortcomings in managing information, namely the absence of a relationship between scientific terms related to medicinal plants and phrases already known to the public. Hence, it is necessary to have this relationship. In other studies, there is no information related to disease protein, so this research also develops ontologies to enrich knowledge about medicinal plants and their efficacy. Based on the results, the developed ontology test can build a relationship between scientific terms of therapeutic pants and phrases that are known to the public. The public also knows which proteins affect a disease, so public knowledge about medicinal plants is getting wider. the testing stage, ontology


I. INTRODUCTION
National Food and Drug Agency (BPOM) 2020 [1] states that Indonesia has 30.000 plant species, and around 9600 are known to have medicinal properties. Plants are known to have efficacy, as herbal medicines have not been used optimally because of the lack of medicinal plant knowledge. Medicinal plant knowledge tends to be degraded so that, if it is not well documented, it can be lost and eroded as natural medicine [2]. Knowledge of medicinal is not only limited to their benefits. It must be viewed from a pharmacology perspective, such as side effects, contraindications, and the content of therapeutic plant compounds. Information about compound content is needed to determine the biological activity of medicinal plants because one medicinal plant can have more than one pharmacological effect [3]. Information on side effects and contraindications of medicinal plants is needed to be more careful in the use of medicinal plants.
Medicinal plant data is the form of knowledge descriptions, so it requires data modeling that can store hierarchical knowledge descriptions and perform knowledge inference. The inference is making conclusions to generate new knowledge [4]. The data modeling method used is ontology. "An Ontology is an explicit specification of a conceptualization," The ontology explains the concept of knowledge of domain such as classes, relations, and objects that humans can understand well to improve semantic interoperability [5]. Ontology can perform knowledge inference, which is suitable for managing medicinal plant data [6]. Conceptual ontology explains the meaning of a domain, and ontology is more concentrated on how semantic concepts are interrelated [7]. Ontology is a link between users and computers so that users get knowledge of a domain such as medicinal plants. Designing an ontology model such as a medicinal plant ontology allows users to obtain information about medicinal plants, such as side effects and the correlation between medicinal plants and diseases [8].
Previous research has designed an ontology based on medicinal plant ethno medicine and produced a prototype ontology of medicinal plants [9]. This research expects the addition of new domains in medicinal plant knowledge. Building an ontology from scratch is unnecessary in developing an understanding of medicinal plants. However, it can use existing ontologies to produce complete and accurate information about the domain of knowledge [10]. Fazriani's research used the ontology data modeling method to build an understanding of mobile-based medicinal plants [11]. The drawback of this study is there is no relationship between the activity of the compound and the term efficacy known to the public. This relationship is needed so that people can more easily understand scientific terms. Several studies used ontology data models for  [12], Persian herbal medication [13], Indian medicinal plants with standardized medical terms [14], and individual ontology of Thai ginger based on taxonomy [15].
This study aims to design an ontology model of medicinal plants by building a link between the activity of compounds and the efficacy of medicinal plants, as well as being able to perform knowledge inference, which has not been done in previous studies, namely the research of Wardani [9] and Fazriani [11]. This research also adds a new domain, namely the disease protein domain, to expand medicinal plant knowledge. This research will produce a medicinal plant ontology data model that is more flexible in adding relationships and entities. The ontology data model uses the Web Ontology Language (OWL) format. Testing ontology model of medicinal plants uses the Simple Protocol and RDF Query Language (SPARQL). SPARQL is a standard query language for extracting information. SPARQL is structurally similar to SQL queries, but the perfomance of SPARQL is more semantic compared to SQL, because SPARQL can extracting information from data that has an OWL format [16]. OWL has a triple data structure, namely subject, predicate, and object, so that SPARQL can produce information in the form of class hierarchies from the ontology of medicinal plants [17]. In future research, the ontology data model with OWL format is expected to assist the development of the web-based semantic system or Knowledge Management System (KMS).

II. RESEARCH METHOD
This research has several stages: the first step is data collection, the second step is data analysis, the third step is ontology design, the fourth step is ontology implementation, and the final step is ontology testing seen in Figure 1. This research uses qualitative techniques for data analysis because collected medicinal plant data describe knowledge [18] and uses ontology model data to describe the medicinal plant domain and hierarchical structure clearly [19].  Table   1.

Data Analysis
Data analysis used qualitative techniques because medicinal plant data described knowledge [18]. There are several processes at the data analysis stage: first, reviewing all data on medicinal plants that have been collected. Second, perform data reduction to retrieve core data related to medicinal plants. The Third is the presentation of data in tabular form to produce information related to medicinal plants. Fourth, draw conclusions and information from the tables presented to support knowledge in ontology design.

Ontology Design
In this research, the ontology design adopted the research that Mitsis [20]

Ontology Implementation
The ontology implementation stage, using the protégé 5.5 tools. Protégé tools to implement the ontology design into the software so that the results of the ontology model can be visualized [21].
The ontology implementation has several processes: building classes and subclasses, then building properties between classes. The property in protégé consists of object properties that connect or map between one type and another and data properties that connect each class with a value [12]. The last process of implementing the ontology is to build individuals according to their respective categories.

Ontology Testing
Ontology

C. Ontology Implementation
The ontology implementation uses protégé 5.5 tools to visualize the ontology model. The first stage of ontology implementation is to define the ontology domain of medicinal plants into classes whose representation using graphs can be seen in Figure 6.   The SPARQL query and results can be seen in Figure 9.

IV. CONCLUSION
The ontology model that has been designed can perform knowledge inference and build new