Development of an Ontology-Based Solution to Reduce the Spread of Viruses

Abstract

With the sudden emergence of many dangerous viruses in recent years and with their rapid transmission and danger to individuals, most countries have adopted several strategies, such as closure and social distancing, to control the spread of the virus in the population. In parallel with all these precautions, scientific laboratories are working on developing the appropriate vaccine, which in many cases takes many years. Until then, it is necessary to resort to many solutions, including solutions that rely on information technologies and artificial intelligence (AI). In this context, this paper proposes a new solution based on the ontology and rules of intelligent reasoning. Initially, the virus environment is analyzed, followed by the extraction and editing of the main elements of the ontology using the “Protégé” software. In the last step, the proposed solution is tested, by choosing the city of Adrar in southwestern Algeria, which was particularly affected by COVID-19. Three scenarios were shown for different cases. The efficiency of the proposed solution was confirmed through the instructions it provides in the event of symptoms appearing in a person. In addition, this solution helps the competent authorities know the location and extent of the epidemic by informing the local communities.

1. Introduction

Viruses are one of the most well-known types of microorganism and the most common type of infection that affects humans, as they can be transmitted to humans in various ways: through the air, food, or direct contact with an infected person, animal, or surface [1]. Viruses are famous for their ability to mutate and adapt to be able to live and coexist. Throughout history, humanity has undergone several pandemics caused by viral infections, such as swine flu, bird flu, and finally COVID-19 [2].

During the emergence of viruses, countries strove to find solutions to limit their spread among individuals, whether manually or through applications or programs, and the use of digital health applications emerged as a mechanism to confront the repercussions of many viruses. One of the most important examples is the experience of China and the World Health Organization in the field of digital health applications: AI, big data, 3D printing, and digital medicine as mechanisms to mitigate the spread of the virus and limit its spread [3]. The use of digital health applications at the global level enhances the efficiency of healthcare services in confronting viruses, and artificial intelligence plays a major role in addressing many problems [4]. In the AI family, deep neural networks and recurrent networks are used to predict activities, such as physical and chemical properties, and consistently improve information processing and decision making by developing powerful algorithms. AI is useful not only in treating infected patients but also in monitoring their health [5].

Big data and AI are naturally equipped to play an important role in the containment of viruses. AI contributes to analyzing data and building risk-based models to help analyze infection-transmission scenarios, monitor disease, provide for early detection, diagnose and follow up infected cases, and achieve comprehensive monitoring of all individuals to ensure their compliance. In addition, an analytical descriptive approach was used, with researchers concluding that AI and big data contribute to the prediction of future strategies by providing information about features of viruses such as their physical and chemical properties [4]. Besides this, the use of ontology and the modeling of functional interactions between genes, combined with a growing interest in systems biology, represent an excellent opportunity to create global interaction models for animal viral infection [6].

When applied alone, anti-virus strategies are not sufficient. They are not sufficient to completely eradicate viruses, as software solutions are necessary and complementary to medical treatments. First, one such example is the “close contact detector” launched by China, which helped reduce infection rates by informing individuals that an infected person was close to them [7]. Second, the Singapore-developed “TraceTogether” is a tracking program that can access information related to those infected with the virus. The app tracks the location and proximity of infected people to other people, notices individuals who come in contact with a confirmed case, and identifies potential locations for the virus to spread [8]. Third, the “Sickweather” app relies on social sharing by allowing people to post information about their area and share information about infection with others [9]. On the contrary, if a person is looking for an answer for the appropriate treatment to alleviate the symptoms of infection, the “MyChart” application provides the necessary health instructions, in addition to providing information through which you can assess a person’s health [10].

In this work we propose a new solution based on ontology and the rules of intelligent reasoning. Based on this smart solution, the sudden emergence and rapid spread of many dangerous viruses that have appeared in recent years can be prevented. This incited researchers to work on developing solutions to limit the spread of a specific virus and not most viruses that share certain symptoms. For this reason, we opted for developing this solution to cope with the emergence of strange viruses that require a long time to be identified accurately. This accurate identification is needed to develop appropriate vaccines. Following this, the appearance of strange symptoms in a person automatically leads to informing official institutions, which in turn take urgent measures, such as applying a policy of distancing to limit the spread of the virus or providing preventive treatments. The advantages of the developed solution can be summarized in the following:

−

Using the ontology enables the necessary updates to be made in the database without causing a bug in the solution;

−

With the emergence of new viruses, new concepts can be added at the right place and time without re-developing the entire solution;

−

This solution is characterized by accuracy, speed, and extensibility;

−

This solution is open to all future developments.

The major research contribution of our work may be summarized as follows:

−

We propose a smart solution to provide preventive guidance about an individual’s health. The inputs to this solution are data about the person and his surroundings in terms of health status and virus spread.

−

The proposed solution is based on information and communication technologies; in addition, from a media point of view the field of health and diseases is an open and distributed system that includes a large amount of information, which requires good structuring.

−

An ontological approach is advanced to develop the proposed solution. The solution’s knowledge base includes information about the general symptoms of viruses and how to prevent them. Ontology technology has many advantages, such as flexibility in the use and updating of information and its reliance on collaborative and intelligent thinking, through a set of pre-developed rules that work in concert with each other to provide immediate guidance and advice to people to reduce the spread of infection.

To accomplish this work, a series of steps was followed:

−

The first step represents a field study of everything related to the viruses. The results of this step are the extract of information, which includes concepts, concepts’ attributes, relationships between concepts, examples related to concepts, and rules of intelligent reasoning.

−

In the second step, all this information obtained with the “Protégé” program is edited and checked using syntax checking to display edited information.

−

Finally, the final step is to implement the proposed solution using several scenarios.

The remainder of the article is organized as follows: After this introduction, Section 2 is devoted to presenting the study environment. In Section 3, the proposed methods and obtained results are presented. This includes the design and implementation of the proposed solution. Various scenarios are addressed. A detailed description of the rule is given; analysis and discussion follow in Section 4, and concluding remarks are then given in Section 5.

2. Presentation of the Study Environment

According to the latest data from the World Health Organization, a large number of virus-specific symptoms have been encountered [11]. Often these viruses share many symptoms, such as fever, a change in facial features, and diarrhea. Where one of them appears, medical intervention must be carried out. The first symptoms may appear within 5 to 6 days after infection, but this period can extend to 14 days. The most common symptoms can be mentioned:

−

Fever: a person is considered to have a fever if his temperature is 38 degrees Celsius or more, and it may also be accompanied by flu symptoms such as headache, cough, sore throat, shortness of breath, and chest pain [12].

−

Dry cough: coughing without sputum output. In rare cases, the cough may be accompanied by blood [13].

−

Sore throat: a sore throat is often an early symptom, and due to the inhalation of respiratory viruses, they enter the nose and throat first and may multiply there early, which leads to throat pain and irritation [14].

−

Pains and aches: these include muscles and joint pain.

−

A persistent sense of “nasal douche”: early nasal symptoms that often appear with loss of sense of smell (or decreased sense of smell) and loss of taste.

The patient needs to have a medical file to monitor his health condition accurately (

Table 1. Examples of patient-file information.

Information	Description	Type
Identifier	Patient’s number is the same as the medical file number.	Alphanumeric
Name	Patient name	Alphabetical
Sex	Gender (male or female)	Alphabetical
Nationality	Patient’s nationality	Alphabetical
Date of birth	Patient’s date of birth	Alphanumeric
Marital status	Marital status of the patient (single/married)	Alphabetical
Profession	Patient’s profession	Alphabetical
Address	Patient’s address	Alphanumeric
Mobile number	Patient’s mobile number	Digital
Social Security number	Patient’s social security number	Digital
Date of each consultation	Date of each patient consultation	Digital
Name of another person	Name of another person to contact if necessary	Alphabetical
Hospital phone number	Phone number of the hospital where the patient is being treated	Digital
Hospital email	Email of the hospital where the patient is treated	Alphanumeric
Name of the attending physician	Name of the physician treating the patient	Alphabetical
Fever	A symptom that affects a person after being infected with the virus	Alphabetical
Headache	A symptom that affects a person after being infected with the virus	Alphabetical
Cough	A symptom that affects a person after being infected with the virus	Alphabetical
Conjunctivitis	A symptom that affects a person after being infected with the virus	Alphabetical
Dry cough	A symptom that affects a person after being infected with the virus	Alphabetical

). This information is very important for the realization of the knowledge base, which will be approached in the next part of this work.

3. Methods and Results

Ontology is a formal and explicit representation of information associated with a specific domain [15,16,17]. In medicine, ontology is interested in the genesis of medical entities: diseases, clinical signs, clinical syndromes, symptoms, lesions, biological anomalies, and radiological anomalies [18]. The proposed solution includes three levels, which are knowledge, processing and intelligent reasoning, and the displaying of results (Figure 1).

−

Level of knowledge: includes concepts, attributes, associations between concepts [19,20,21].

−

Level of processing and reasoning: includes reasoning rules, inference engines, classification algorithms [22,23,24].

−

Level of display of the results: includes the demonstration of the results in the form of text messages, vocal alerts, or video.

3.1. Concepts

Conceptualization refers to an organized set of concepts, which can be defined as an entity composed of three distinct elements [25]:

−

The term(s) express(es) the concept in language;

−

The meaning of the concept is also called “notion” or “intension” of the concept;

−

The object(s) denoted by the concept, also called “realization” or “extension” of the concept.

These elements are usually presented as constituting the vertices of a “semantic triangle”. For this solution, a distinction can be made between some concepts that represent knowledge about the virus (

Table 2. An extract of the solution concepts.

Concept	Description
Person	This can be a normal person or a doctor who cares about providing first aid and treatment to people who show symptoms of the virus.
Local Authorities	This relates to institutions and associations that work to address citizens′ problems.
Fever	A person′s temperature is high, which is a sign of infection with the virus.
Cleaning Hands	One of the necessary precautions that must be taken is washing hands with an antiseptic or anti-bacterial and anti-viral substance.
Headache	Headache is one of the signs that appears in a person infected with the virus

).

3.2. Attributes (Or Properties) of Concepts

Attributes are sometimes called roles or properties, since, after defining some classes, it is necessary to describe the internal structure of the concepts [22]. Each of the proposed solution concepts has a set of attributes (

Table 3. Examples of concept attributes.

Attribute	Type	Description	Concept
Name Patient	Alphabetical	Patient name	Person
Hospital Address	Alphanumeric	Hospital address	Hospital
Doctor Phone	Numeric	Doctor’s phone number	Doctor
Name of Drug	Alphanumeric	Concerns drugs for first interventions	Treatment

).

3.3. Solution Relationships

A relationship is the creation of a combination of two or more concepts of an objective relating to data structuring and the completion of action [26]. While some conceptual connections can be expressed using the properties held by concepts, others must be represented using independent relationships [27]. The proposed solution includes some relationships that work together in a conceptual data model (

Table 4. Examples of solution relationships.

Relationships	Descriptions	Concepts
Fever Has Treatment	Fever needs treatment	Fever, Treatment
Doctor Treat Patient	Doctor is responsible for treating the patient	Doctor, Patient
Hospital Include Doctor	Hospital includes a set of medicines	Hospital, Doctor

).

3.4. Instances (Individuals)

Ontology, and all the individual instances of the classes, constitute a knowledge base [28]. Instances are the extensional definition of ontology; these objects convey knowledge about the domain of the problem.

3.5. Intelligent Reasoning

The reasoning is concerned with the exploitation of acquired knowledge to produce new knowledge, as it uses inference mechanisms that allow solving many problems [29]. Additionally, different reasoning mechanisms are used, depending on the objectives of the system to be prepared [16,17]: logical reasoning, inference by classification, filtering, inheritance, and rule-based reasoning (

Table 5. Examples of rules for the solution.

Rule	Description
R1	If a person has a high body temperature, this rule advises him to measure his body temperature with a medical thermometer and to apply a distance from people of more than 2 m, as well as wear a mask. In addition, the importance of this rule lies in reminding the patient to use a thermometer to ascertain this value accurately, as well as the importance of having this device near anyone.
R2	Using a thermometer to measure body temperature. If the temperature is greater than or equal to 38 degrees Celsius, this rule suggests that the person undergoes an urgent examination by a doctor with the application of the rule of a more than 2 meters’ spacing between people, as well as mask wearing. This rule was mainly based on the fulfillment of only one condition, which is that the body temperature is greater than or equal to 38 degrees, because the latter is one of the main indicators of the virus, and a rise in body temperature above 38 affects body functions in humans.
R3	If a person loses their sense of smell, this rule suggests rapid medical intervention to test for the virus, with the application of the rule of more than 2 meters’ spacing between people, as well as mask wearing.
R4	If someone feels a headache, this rule advises seeking medical intervention and testing for the virus, with the application of the rule of more than 2 meters’ distancing between people, as well as mask wearing.
R5	If a person feels a sore throat, this person should seek medical intervention and get tested for the virus, with the application of the rule of more than 2 meters’ distancing between people, as well as mask wearing.
R6	If a positive case of the virus is detected in a person, there are many precautions to be taken by other people, such as physical distancing between people, wearing a mask, and using disinfectants.
R7	If a person is in a public place, this rule suggests applying physical distancing and leaving a distance between people that is greater than 2 m. This rule also suggests wearing a mask.
R8	If a person talks to others, or coughs or sneezes, he should put a tissue on his mouth and wash his hands with an alcoholic solution.
R9	If a person is in a closed place (for example, a room), he must wear a mask, as well as open the windows, to ensure a change in air in the place and thus reduce the transmission of the virus from one person to another.

and Appendix A).

3.6. Implementation

This section is reserved to present the main elements of the ontology implementation, such as data editing, and the tools used.

3.6.1. “Protégé” Software

“Protégé 3.5” (Stanford Center for Biomedical Informatics Research, Stanford, CA, USA) is a stand-alone open source platform, which provides a graphical environment enabling the editing, visualization, and control (verification of constraints) of ontologies. The “Protégé 3.5” knowledge representation model is derived from the frames model [19]. The latter contains classes (to model the concepts), slots (to model the attributes of the concepts), and facets (to define the values of the properties and the constraints on these values), as well as instances of the classes.

3.6.2. Choice of a Specification Language

The choice was directed towards OWL, which is the standard language of representation and specification of ontology, compared to its predecessors RDFS and DAML-OIL, which are insufficient to codify the ontology of this memory in terms of semantic functionalities [30]. In addition, coding ontology in OWL form has the advantage of making it reusable.

3.6.3. Ontology Editing and OWL

Creating ontology is a complex task, as this process takes several lines of code and requires significant effort and time, especially if this ontology is coded directly in the ontology language without using any tool. For this purpose, the program “Protégé” has been designed to avoid all complications and create an ontology and hierarchy structure (Figure 2).

3.6.4. Verification of the Semantic and Grammatical Aspect

One of the important steps that we must go through is the process of validating the semantic and grammatical aspects, because adding new information and functions in the ontology may result in semantic contradictions. To tackle this task, there are many suggestions, including the use of lexical and structural features between ontologies to check for semantic deviations [31]. Another avenue consists in discovering semantic anomalies [32]. For ontology in this work, automatic validation is performed with Deep Learning (DL), which is based on the use of Pellet 1.5.2 logic. This choice is guided by the availability of the logical plug-in included with Protégé-OWL 3.4.4. Both methods involve repeated operations until more anomalies can be identified. For this solution, no error appeared and the ontology has been syntactically validated (Figure 3).

3.6.5. Presentation of a Scenario

To test the proposed solution, the city of Adrar in southwest Algeria was chosen. This city has been particularly affected by COVID-19 since March 2020. It is very important to have a database at our disposal for each person in Adrar city, such as his health condition, the diseases and treatments he has previously undertaken, and other information detailed in Table 1.

Scenario 1

In the first scenario, two people are tested. The first has a high body temperature whereas the second does not. The first step consists in updating information about the two people. Following this, this solution offers many tips using the rules of intelligent reasoning, such as using a thermometer to properly measure temperature and the application of a distance of at least 2 m between people. The final step consists in recommending the use of the mask (Figure 4).

Scenario 2

Similarly, in the second scenario, two people are examined. After measuring the temperature of the two people, it was found to be above 38 degrees Celsius for the first, whereas the second person has a temperature below 37 degrees. Using the rules of intelligent reasoning, the solution offers a set of tips. For the first person, it is a question of carrying out the medical treatment, applying an estimated distance of 2 m between persons, and wearing a mask. For the second person, the advice given concerns the application of the distance of 2 m between people and wearing a mask (Figure 5).

Scenario 3

In the third scenario, six people were selected, and each was distinguished by signs and actions that could be a sign of infection with COVID-19 (Figure 6). After entering each person’s information into the solution’s database, advice was gained that could limit the spread of COVID-19 (

Table 6. Signs and actions of the third scenario.

Person	Signs	Actions
Person 1	He talks to others Coughing Sneezes	Washing hands with alcohol-based liquid Covering the mouth with hands while speaking Covering the mouth with a tissue when sneezing or coughing
Person 2	Coughing Sneezes	Washing hands with alcohol-based liquid Covering the mouth with a tissue when sneezing or coughing
Person 3	Coughing	Washing hands with alcohol-based liquid Covering the mouth with a tissue when sneezing
Person 4	He talks to others Coughing	Washing hands with alcohol-based liquid Covering the mouth with a tissue when coughing Covering the mouth with hands while speaking
Person 5	He talks to others	Washing hands with alcohol-based liquid Covering the mouth with hands while speaking
Person 6	/	Washing hands with alcohol-based liquid

).

4. Analyze and Discussion

In the healthcare sector, AI approaches have been widely used due to the urgent challenges posed by the spread of viruses, particularly in the area of awareness and precaution [33]. Among the artificial intelligence approaches that have emerged with great effectiveness is an ontology that we have relied on to develop the solution.

Table 7. Description of the scenarios.

Symptoms	Solution Suggestions	Means Used	Notes
When a person feels an increase in his body temperature	This solution suggests that the person check their temperature with a thermometer	Thermometer	Undertake this as a precaution
When a virus appears nearby (even without showing any symptoms)	This solution suggests using a mask and washing your hands regularly with alcoholic liquid or soap and water	Mask, alcoholic liquid, and water	Prevention must be performed in daily life
When someone talks to you	This solution suggests covering the mouth and nose with the hands or by bending the elbow	/	This should be performed, especially in public
When the body temperature is greater than 38 degrees	This solution advises a medical intervention without forgetting to apply the rules of prevention	Thermometer, mask, alcoholic liquid, and water	Medical intervention must be performed very urgently
When sneezing or coughing	This solution recommends using a tissue and keeping away from the others, at a distance of not less than two meters	Tissue paper	It is also recommended to keep a distance of at least 2 m when someone else coughs or sneezes
When a person is in closed places (home, workplace, etc.)	It is recommended to open the windows to allow the air to change	Windows	Adequate ventilation is very important to avoid the spread of the virus
With the onset of any of the known symptoms of COVID-19, such as fever, dry cough, fatigue, loss of taste or smell, body aches, headache, sore throat, nasal con-gestion, red eyes, and diarrhea and skin rash	The proposed solution advises people with these symptoms to seek medical attention as well as get tested for COVID-19	Thermometer, mask, alcoholic liquid, and water	Medical intervention must be performed very urgently

summarizes the scenarios previously described:

From the results obtained, the following can be concluded:

−

This solution is a very effective preventive method;

−

This solution can help the competent authorities to know the location and extent of the epidemic;

−

This solution features the ability to quickly share updated information by encouraging behaviors such as hand washing and social distancing;

−

The proposed solution advises people who suffer from these symptoms to seek medical attention and follow the preventive instructions, and thus the individual has an important role in achieving acceptable results, through his cooperation and carefully following the instructions and recommendations, as well as reporting when a viral disease is suspected;

−

Unlike other solutions, this solution can identify a diseased condition in the individual, by containing the symptoms that affect the person in the knowledge base of the ontology;

−

If a new (unknown) virus appears, there are two possibilities. The first is that it shares some of its symptoms with the viruses defined in the database. In this case, it can be dealt with. In the opposite case, it may pose a danger, but once it is identified, its identifiers can be entered into the database.

Structurally, we know that the environment for this study (humans and viruses) is a dispersed (the system elements are geographically distributed) [34], dynamic (the system elements are changing), and open (the system interacts with other systems) system [35]. These last characteristics are among the main reasons for choosing the ontology approach. Using this approach, many advantages can be gained compared to the previous solutions, including:

−

When a new virus appears, the knowledge base can be easily updated with information about this virus without a defect in the solution;

−

Unused information can be deleted from the knowledge base without going through the steps of developing the solution;

−

In this solution, the ontology approach uses medical information from multiple sources that may use different vocabularies. To operate it, an ontology defines common terms in natural language and characterizes them in a computer-understandable logical language (semantic aspect). Thus ontology can support the automatic creation of generic data models that can aggregate data from different sources. This feature is not available in other artificial-intelligence approaches;

−

Unlike other artificial-intelligence approaches, works based on this ontology are open to future developments, and their knowledge base can be easily exported and used by other works and researchers.

Finally, it should be recalled that almost all research questions (hypotheses) related to the development of a smart solution to address and prevent the risk of virus spread have been answered. In addition, comparing it to other works [36,37,38,39], we conclude that this solution has strong additions, whether in response (through results) or adaptation (ontology-based solution) to all developments. Accordingly, many recommendations regarding investment in this solution can be made by researchers specialized in the field of medicine. The establishment of a scientific project at the level of research centers is a qualitative initiative to develop more efficient solutions and thus protect humanity from the threat of viruses. It is also important to generalize these solutions at the global level and provide them to poor countries for free.

5. Conclusions

Due to the nature of the virus, where it is difficult for health systems to detect them. We thus face a challenge for health authorities around the world in the early detection of cases. With the spread increasing in more than one country, quarantine or early detection at points of entry alone is not enough, so screening systems need to be strengthened to detect disease transmission at the community level. In this context, an intelligent solution based on ontology and the capitalization of knowledge has been developed.

The results obtained showed the great importance of the quality and accuracy of the advice provided, thanks to which it was possible to limit the spread of the virus. From the viewpoint of exploitation, the proposed solution has many advantages, among which are:

−

Flexibility during use;

−

The ability to update the data without repeating the design stages;

−

Adaptation to most cases (the emergence of new strains of the virus);

−

This solution can be considered in addition to being a marketing tool; it can also be considered as a platform for students and researchers in the field of health and epidemiology.

About the works to be carried out in the future, the following may be cited:

−

It is important to enrich the knowledge base with further, recently discovered information on the evolution of the virus;

−

It is very important to develop the work in coordination with health experts and doctors to make it an important tool in the daily work of health organizations;

−

It is important to add information regarding vaccines and drugs to eradicate the virus;

−

It is very important to include new technologies, such as big data techniques, in the development of solutions;

−

Due to the nature of the environment of this study, it is characterized by geographical distribution. It is, therefore, possible to propose a hybrid solution between ontology and multi-agent systems.