Prediction the Number of Dengue Hemorrhagic Fever Patients Using Fuzzy Tsukamoto Method at Public Health Service of Purbalingga

DHF (Dengue Hemorrhagic Fever) is still a major health problem in Indonesia. One of the factors that led to an increase in dengue cases is uncertain climate that causes dengue fever is difficult to be predicted. Prediction is an important thing that is used to determine future events by identifying patterns of events in the past. When knowing the events that happen, it will make everyone to make better preparation for everything. This research is aimed at determining the accuracy of Tsukamoto Fuzzy method in the number of dengue patients in Puskesmas Purbalingga. Tsukamoto Fuzzy method can be used for prediction because it has the ability to examine and identify the pattern of historical data. Tsukamoto fuzzy that used to predict the number of dengue fever patients at Puskesmas Purbalingga has several stages. The first stage is the collection of climate data includes precipitation, humidity, water temperature and the data of dengue fever patients in Puskesmas Purbalingga. The next stage is processing the data that has been obtained. The last stage is to make predictions. Based on the results of the implementation by Tsukamoto Fuzzy method in predicting the number of dengue fever patients in Purbalingga for twelve months in 2016, it was obtained a percentage error (MAPE) of 8.13% or had an accuracy rate of 91.87 %. With the small value of MAPE and high accuracy, it shows that the system can predict well.


INTRODUCTION
Dengue Hemorrhagic fever is a major health problem in Indonesia.The number of cases and the spread of this deadly disease is increasing in line with the increasing mobility and population density.The population of Indonesia is more than 250 million people and the fourth rank in the country with the highest population in the world.WHO records Indonesia as the country with the highest dengue Hemorrhagic fever case in Southeast Asia [1].Global warming that causes the weather becoming erratic makes the number of dengue Hemorrhagic fever is now difficult to know.
According to Health Department of Purbalingga Regency among the regencies in Indonesia, Purbalingga Regency is an area suffering from dengue Hemorrhagic fever.It can be seen from the figures incidence of Dengue Hemorrhagic Fever (DHF) equal to 64 per 100,000 population or 572 cases with 3 deaths (CFR = 0,5%) during 2013.This happened because until now there is no vaccine for dengue Hemorrhagic fever that can be used to humans.DHF control efforts in Indonesia are still focus on surveillance programs, Mosquito Nest Eradication (MNE), and Periodic Monitoring (PM).
The increase of number of patients and victims who died showed the less prepared precautions taken.Health Department is still waiting for reports of dengue cases before fogging [2].This rationale makes it necessary to make models that describe the dynamics of the number of cases of dengue fever.Models with sufficient accuracy can be used as a tool for decision-making and distribution of resources to cope with epidemics [3].
Prediction is an important thing used to know the future events by recognizing the patterns of events in the past.While knowing the events that will happen, it makes everyone better prepare everything, both for human life and their property.
Basic Fuzzy logic is a fuzzy set theory.Fuzzy logic contained within the membership function.According to [4] the membership function is a curve that shows the mapping of points of input data into membership values that have intervals between 0 and 1.One application of fuzzy logic that has developed today is immense Fuzzy Inference System (Fuzzy Inference System/FIS), a computational framework that is based on fuzzy set theory, fuzzy rules IF THEN shaped and Fuzzy reasoning.There are three methods in Fuzzy inference systems that often used, namely Tsukamoto, Mamdani and Takagi Sugeno.This study uses Tsukamoto Fuzzy method because this method has several advantages among others, this method can do faster computation, more intuitive, accepted by many, and better suited to the input received from the man not by machines [5].
Fuzzy logic can be used on systems that have non-precise inputs or more linguistic inputs.Research related to fuzzy logic, including the ranking of construction projects [6], analysis of learning relationships with student satisfaction [7], detection of reasoning ability [8], employee performance appraisal [9], best decision making [10], [11], prediction Number of visits [12], [13] and predicted number of students [14].
Based on the background above, it is discussed regarding: (1) how to implement the Tsukamoto Fuzzy method in predicting the number of dengue patients at health public service of Purbalingga (2) what level of accuracy resulting in the number of dengue patients at health public service of Purbalingga

METHODS
There are several stages performed systematically in this study, starting from the preparation stage, then data processing, and predict the number of dengue hemorrhagic fever patients.

Preparation phase
The preparation phase begins with a literature study that supports the success of the research.In addition, this study also conducted a search of data needed for this research, namely by taking data to the health public service of Purbalingga to obtain data on the number of patients and BMKG to obtain data of weather variables for three years from 2014 until 2016.

Data processing
After getting the required data, then the data is processed.This is necessary so that the secondary data obtained can be incorporated into the prediction method Tsukamoto Fuzzy.In addition, a standard is required to process the data, in this case related to the determination of data types and also the determination of predictive time limits.Data on the number of patients and weather variables used are data per month for twenty months.The predicted month limit used in this study is twenty months before the prediction month (2014-2016), ie, the first twenty months (2014-2015) function as training data, the next twelve months (2016) as test data.

Prediction the Number of DHF Patients
The procedure of prediction the number of DHF patients using Tsukamoto Fuzzy method is as follows.a. Determine the scope of training parameters b.Determine the election limit for months to be trained c.The training process uses Fuzzy Tsukamoto against the preceding months d.Incorporate test data on training system to get corrections e. Prediction the number of DHF patients According to Kusumadewi [15] in Tsukamoto's method, any consequence of the ifthen-form rule must be represented by a fuzzy set with a monotonous membership function.As a result, the inference output of each rule is given explicitly (crisp) based on α-predicate (fire strength).The end result is obtained by using weighted averages.To obtain the crisp output value / Z value firmly, searched by changing the input (a fuzzy set obtained from the composition of fuzzy rules) into a number in the fuzzy set domain.This method is called the method of defuzification (affirmation).The defuzification method used in the Tsukamoto method is the Center Mean Defuzzifier (CAD) centered defuzification method presented in the equation below Equation 1.
According to the above description then the rules that can be formed can be presented as in Table 2.The final stage in the above case is the defuzzification process using the fuzzy tsukamoto method in Matlab software which will produce data or prediction in the form of the number of dengue patients during the next year.

RESULTS AND DISCUSSION
In this study, predictions were made using monthly input data for twenty months before the prediction month.In the training process of predicting the number of dengue fever patients, the input data is the number of dengue patients and climatic variables for twenty months from 2014-2015.The exposure of the results will be presented in the form of tables and graphs.Table 3 shows the comparison of predicted results with targets for 2014-2015 while the comparison comparison of predictive training results for the 2014-2015 targets is shown in Figure 7.The second stage is processing the data that has been obtained by determining the type of data and predictions month limit, and (3) The third stage is a training method Tsukamoto Fuzzy with a predetermined input.The results shows that the percentage error (MAPE) is currently 8.13% predicted or otherwise has an accuracy rate of 91.87%.The MAPE value is small enough and the accuracy is high enough then it shows the system can predict well.

Figure 8 .
Figure 8.Comparison Chart Prediction Test Result with Target Year 2016 4. CONCLUSION

Table 2 .
Results of Rule BaseBased on the results in accordance with the rule-based knowledge base, as shown in Table2, it can be inferred rules that formed on fuzzy inference as follows.

Table 1 .
The data used for the prediction is data of Public Health Service of Purbalingga dengue patients from May 2014 through December 2015.Establish a universal set of conversationsThe universe of speech of each variable is the range of possible values of the existing data.The fuzzy variable can be shown in Table1.Fuzzy Variables

Table 3 .
Comparison of Predicted results with year 2014-2015 Graph Comparison of Predicted by Target Training Results Year 2014-2015When conducting the test, the weight gained from the training was called and processed with different input from the training process but similar.Testing is done with the data for twelve months and produce MAPE is 8.13% which can be seen in Table4, while the graph of test results using Tsukamoto Fuzzy method in predicting the number of dengue patients during the twelve months of 2016 is shown in Figure8.

Table 4 .
Comparison of Predicted to Target Results with Target Year 2016