Implementation Of Vikor Algorithm In Web-Based Recommendation System For Obstetrician Selection

id


I. INTRODUCTION
The process of childbirth is a normal event that is important in a woman's life.The term "normal" describes that labour and birth have occurred for many years [1].The population growth in the world is increasing due to the increasing number of births that occur [2].In general, the birth process is carried out in a maternity hospital with the help of an obstetrician [3].Selection of the right obstetrician according to individual needs is an essential factor.However, many parents-to-be couples face difficulties choosing the right obstetrician to help, from health consultations for pregnant women to assisting in the hospital's delivery process.This uncertainty in choosing an obstetrician significantly impacts the experience and outcome of labour [4].Therefore, an approach is needed to help prospective parents make wise decisions.One proposed solution is developing a web-based recommendation system to help users consult [5].A recommendation system is a method developed to predict an item and display the best item value for the user.This system can filter relevant information and provide obstetrician recommendations that best suit the preferences and criteria of each pair of prospective parents.This system aims to assist users in navigating the various choices of obstetricians efficiently, considering the role of obstetricians is crucial in the delivery process.Using a recommendation system can reduce the burden of information users receive and guide them in making better decisions [6].This system will create a model based on user input regarding the user's references [7].
The model developed will be used by the system so that it will filter information and adjust it to the criteria entered by the user so that it displays the recommendations that the user wants.Previous research created a decision support system for selecting obstetricians using the Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) method [8].In this study, calculations were carried out using five criteria or attributes to determine which obstetrician the pregnant woman would choose.The criteria used are delivery rates, consultation rates, hospital position, obstetrician popularity and experience.The research was conducted on five users with different weight values for each user.The results obtained from the recommendation system using TOPSIS [9] were a 40% difference between the system results and the actual data results, so the recommendation system had significant and inaccurate error accuracy.Research conducted by Lahby et al. [10]  including TOPSIS (Technique for Order of Preference by Similarity to Ideal Solution) and VIKOR (VIseKriterijumska Optimizacija Kompromisno Resenje) [11].The research aims to choose which algorithm is best for mobile network selection in heterogeneous wireless networks.The study conducted four simulations with four traffic types: background, conversational, interactive and streaming.The results show that the ranking of the VIKOR method in the four simulations is superior to the TOPSIS method, so it can be concluded that the VIKOR method provides better results than the TOPSIS method.Based on previous research, this research will create a recommendation system for obstetricians using the VIKOR method with the hypothesis that better results can be achieved.The VIKOR approach involves several stages, including determining criteria weights, matrix normalization, Utility Measure and Regret Measure calculations, and VIKOR index calculations [12].This recommendation system will be tested through scenarios and user surveys using the USE questionnaire [13].The scenario testing results will show the effectiveness of implementing the VIKOR method in this system.In contrast, the user survey results will provide insight into user satisfaction with the system being developed.This research has important implications for improving the quality of decision-making [6] in selecting obstetricians.By applying the VIKOR method in an application, this research can make a real contribution to developing a more sophisticated and efficient recommendation system.Thus, this research emphasizes the importance of applying reliable decision-making methods in complex scenarios.

II. METHODS System and Database Design
The system design was created by creating a flowchart [14] using diagrams.netto describe the workflow of a web-based obstetrician recommendation application using the VIKOR method.The database used was also designed using phpMyAdmin [15] to create a MySQL database [15].Database design involves a table structure that supports the storage of obstetrician criteria and information to be assessed.

Evaluation Criteria
This research considers five main criteria in selecting an obstetrician, namely: 1. Hospital Accreditation [16]: Accreditation of the hospital where the doctor practices 2. Insurance [17]: BPJS Health services provided by obstetricians 3. Obstetricians Experience: Obstetrician practice experience in the field of obstetrics 4. Consultation Rates [18]: Consultation rates charged by obstetricians 5. Childbirth Rates [19]: Childbirth rates charged by obstetricians Data was obtained via the official website of the Hospital Accreditation Commission (http://www.kars.or.id/).In contrast, data on BPJS Health services, obstetrician practice experience, delivery rates and consultation rates were obtained manually from the Alodokter website ( https://www.alodokter.com/) and the SehatQ website (https://www.sehatq.com/).There are 15 lines of criteria data taken from these sources.
Application Workflow Fig 1 .Web Application Flowchart Figure 1 is a flowchart or workflow of the running process of a web-based obstetrician recommendation system application using the VIKOR method.Users who use this application will be asked to enter five criteria, namely: 1. Input hospital accreditation 2. Input BPJS (Insurance) services http://ijstm.inarah.co.id 1360 3. Input the obstetrician's experience 4. Input consultation rates 5. Input childbirth rates After inputting the data, calculations are carried out using the VIKOR method according to the previous formula.After the calculation is complete, the obstetrician's ranking will be displayed according to the criteria entered by the user.

VIKOR Method
According to El-Santawy [20], the VIKOR method (VIseKriterijumska Optimizacija Kompromisno Resenje in Serbian, which means Multicriteria Optimization and Compromise Solution) is a method introduced as a technique that can be applied in MCDM.This method focuses on ranking and selecting existing alternatives based on conflicting criteria.According to Zimmerman [21], MCDM consists of two models, namely Multi-Objective Decision Making (MODM) and Multi-Attribute Decision Making (MADM).The difference between the two models is that MADM is used to solve problems in continuous space, while MADM is used to solve problems in discrete areas.The VIKOR method is included in MADM because it solves problems related to assessing or selecting several alternatives using predetermined criteria [22].After the collection of Q values is obtained, the ranking will be carried out by sorting the best data, which has the lowest Q value, to the worst data, which has the largest Q value.

Testing and Analysis
Testing is done by testing the recommendation system through scenarios and user surveys.The results of scenario testing will measure the effectiveness of implementing the VIKOR method in this system.Meanwhile, the user survey results will provide insight into the level of user satisfaction with the system developed using the USE Questionnaire.The USE Questionnaire is a method used to measure the usability of a recommendation system.USE Questionnaire is a package comprising four research variables: usefulness, ease of use, ease of learning and satisfaction [24].Each variable has questions that will be used to evaluate a recommendation system.Testing the recommendation system with the USE Questionnaire is expected to provide information and empirical evidence about the usability of using the recommendation system, which can illustrate whether the system is following user needs or not so that it can provide convenience and satisfaction to users.Respondents were asked to provide answers based on a Likert scale with five alternative answers.The scores obtained from these questions will then be calculated using a specific formula and converted into predetermined eligibility categories.The following criteria for using the Likert scale can be seen in Table 1.
Table 1 After getting the value for each question, usability measurement is carried out by calculating the answers from respondents using the following formula.
=     × 100 The results of the eligibility percentage for each variable are then converted using the feasibility category table in Table 2 Analysis of the test results will provide information about the extent to which this recommendation system has succeeded in meeting user needs in selecting the appropriate obstetrician.

III. RESULT AND DISCUSSION System Implementation
At this stage, the system is implemented using the Grocery CRUD administrator website, designed based on the PHP CodeIgniter framework [25].To access this system, the user must log in first; then, the user will be directed to the main page of the administrator's website, as shown in Figure 3.The main page immediately displays a list of obstetricians who are the objects of research.4 shows the main page of the obstetrician recommendation system using the VIKOR method.Website users will receive information about the recommendation system website and the form that must be filled in to get the obstetrician's recommendation results.No recommendation results will appear if the user does not fill out the form completely.

Fig 5. Results of Obstetrician Recommendations
Figure 5 is an image that displays the results of the recommendations after the user fills out the criteria form completely.The recommendation results are displayed in a table and contain the obstetrician's photo, the obstetrician's ID, the obstetrician's name, the hospital where the obstetrician practices, hospital address, hospital contact, consultation fees, maternity fees and the obstetrician's practice experience.

Recommendation System Scenario Test
The scenario trial of the obstetrician recommendation system using the VIKOR method was carried out using obstetrician data, as shown in Table 3.After obtaining obstetrician data that will be used in VIKOR calculations, the criteria weight data obtained from the user will be compiled along with the maximum value, which is the best value for each criterion and the minimum value, which is the worst value for each measure, which is displayed in After that, matrix normalization was carried out according to the obstetrician's data in Table 3 and the maximum and minimum values contained in Table 4 using formula (1).After normalization is carried out, the normalized data will be multiplied by the weight value for each criterion with formula (2).After obtaining the results of multiplying the normalization matrix with the criteria weights, the next step is to find the S-value (utility measures) using formula (3) and the R-value (regret measures) using formula (4).After obtaining the S and R values from each obstetrician's data, the next step is to find the maximum and minimum values of the S and R values.The S and R values from each obstetrician's data, as well as the maximum and minimum values of S and R, will be used to find the Q value or VIKOR index with formula (5) and the value v = 0.5 After obtaining the Q value for each data, we will immediately rank it by sorting the Q values from the smallest Q value to the largest Q value shown in Table 5.The VIKOR ranking results in Table 5 will be converted into a table in the obstetrician recommendation system web application shown in Figure 6.

Fig 6. Recommendation Results in Web Applications
The figure shows that the data from manual calculations and the data from measures in the system show the same results.
System User Survey Test A system user survey test was conducted to determine the responses of several respondents who used this application.The questionnaire used for user survey testing is the USE questionnaire.After the questionnaire results are obtained, a calculation is carried out to determine the value of the distributed USE questionnaire.Value calculations are carried out for each variable in the USE questionnaire: the variable's usefulness, ease of use, ease of learning and satisfaction.According to formula (6).
http://ijstm.inarah.co.id 1364  6 is the result of calculating the USE questionnaire for each variable and the average for the entire questionnaire.After calculating each variable, a percentage value was obtained for each variable, namely the usefulness variable with a percentage of 79.76%, the ease of use variable with a rate of 84.81%, the ease of learning variable with a rate of 85.48%, and the satisfaction variable with a rate of 83.31%.Measuring the overall USE questionnaire value produces a feasibility percentage value of 83.34%, which shows that the results of testing the obstetrician recommendation system using the VIKOR method have a value that is "very feasible" according to the feasibility category in Table (2).

IV. CONCLUSION
In this research, a web-based obstetrician recommendation system was successfully implemented using the VIKOR method using the PHP programming language.This system can provide obstetrician recommendations according to the criteria determined by the user with the weight of each measure, which can be adjusted according to preferences.In this research, scenario testing and user surveys were carried out.The results of scenario trials on this system show the effectiveness of implementing the VIKOR method.The obstetrician data that has been entered, the matrix normalization process, and the calculation of the VIKOR index value consistently produce recommendations that match user preferences.This provides an overview of the system's ability to process and analyze data accurately to provide meaningful recommendations.Inuser survey testing is carried out using the USE questionnaire.
This testing gave us insight into user responses and satisfaction with this system.The average result of user survey testing is 83.34%, which means it shows an excellent level of feasibility with four variables: usefulness, ease of use, ease of learning and satisfaction, all of which meet the requirements and obtain positive responses from users.Overall, this research has proven that a web-based obstetrician recommendation system using the VIKOR method has the potential to provide an effective solution in selecting an obstetrician.Nonetheless, this research has the potential for further development by improving more comprehensive data integration and current updates regarding obstetricians and hospitals.Developing additional features, such as user reviews and the possibility of selecting by location, could also improve the system's usability. V.

Fig 2 .
Fig 2. Flowchart of the VIKOR Method Figure 2 is a flowchart of the VIKOR method in general, with the procedure[23] explained below: 1. Normalize the decision matrix used in the VIKOR method calculations with the following formula:   = (  * −   )   * −   −

Fig 3 .
Fig 3. Main page of the Administrator Website

Fig 4 .
Fig 4. Main page of the Recommendation System Figure4shows the main page of the obstetrician recommendation system using the VIKOR method.Website users will receive information about the recommendation system website and the form that must be filled in to get the obstetrician's recommendation results.No recommendation results will appear if the user does not fill out the form completely.

.
Criteria for Using the Likert Scale

Table 4 .
Criteria Weights and Maximum and Minimum Values

Table 6 .
USE questionnaire calculation results