Chronic obstructive pulmonary disease restricts airflow in the respiratory system. After the age of 40 or 50, the symptoms of this disease become more apparent. According to studies, there were 251 million cases of COPD worldwide in 2016, which is expected to become the third leading cause of death in 2030. The symptoms of this disease include shortness of breath, chronic coughing, and phlegm over several days. While COPD cannot be cured, it can be identified by its symptoms, which will help to improve quality of life and reduce mortality risks. The fundamental principle in this type of disease is early diagnosis and timely access to medical care (including hospitalization). With the advancement of smart and remote technologies in many fields, especially health and treatment, humans have been able to improve and solve problems more efficiently. In recent years, context awareness has become one of the new and practical concepts in the area of health diagnosis and treatment, due to its combination with fuzzy expert systems. Based on this, a system can be context-aware, which includes information to describe the activities of users of a program, so the information is the context. Because the diagnosis process is patient-centered, this information provides them with the possibility of quick notification of abnormalities, which in turn prevents wasted time and money. As a result, this information provides a suitable background for accurate and early prediction of abnormalities for efficient treatment. It seems necessary to design a system that compensates for the previous shortcomings and uses modern concepts of context and context awareness since no study has been conducted regarding the diagnosis of COPD disease by tissue-awareness method and direct communication with the user has not been used in their treatment methods.
The purpose of this model is to create high-level context information based on low-level data. To achieve this goal, low-level context data are converted into different fuzzy sets that form inputs for the fuzzy inference engine, and then they are converted into high-level meaningful information, and based on this knowledge base, conclusions are drawn about the risk level of the disease. There are several factors that can be used to diagnose the level of COPD, including: 1) Air pollution level 2) Disease symptoms 3) Age 4) Cold and hot environmental conditions (temperature). In order to investigate the effective tissues in determining the severity of the disease, 60 patients with chronic obstructive pulmonary disease during one year (2018) in 25 medical centers of Ilam city were asked to provide their medical history and personal characteristics, as well as they completed their questionnaire under the supervision of a doctor. The slope of belonging to a set slowly changes into "not belonging to it" in fuzzy sets, unlike classical sets. There are four basic steps in a fuzzy inference system: fuzzification, knowledge base, decision making unit, and defuzzification. Using the existing rule base, the decision-making unit obtains the diagnostic levels of COPD disease (normal, moderate, severe), which can then be used with the de-fuzzification of the outputs. All the data received from the patients were entered into the fuzzy system and the output was compared with the recorded diagnosis of the specialist doctor to evaluate the system. In order to evaluate and perform the fuzzy system in the diagnosis of COPD, the Kappa test, accuracy, sensitivity, and F1 scores were used.
Results and Discussion:
The determination of effective individual and environmental contects for diagnosing chronic obstructive pulmonary disease was first prepared using library research and risk assessments of determining factors, and then based on a lung disease specialist's opinion. As a result, the fuzzy expert system was designed based on the parameters of the symptoms (such as coughing, chest phlegm, feelings of heaviness in the chest, breathing, sleeping, energy level, range of physical activity, relief from leaving the house - despite the physical condition), air pollution level, age, cold and hot environmental conditions (temperature), etc. After determining the membership functions, it is necessary to create the rule base. According to four system inputs, 183 "if-then" rules were performed. By using the Kappa statistical test, we compared the fuzzy expert system's diagnosis with that of the relevant doctor's final diagnosis. The Kappa statistical test had a value of 0.77, which indicated a very strong relationship between the variables. In the confusion matrix, which is used to evaluate the model's performance, the accuracy value was equal to 0.8167, the sensitivity value was equal to 0.8182, and the F1 score was equal to 0.8791.
A fuzzy context-based expert system is a new approach that provides accurate and quick information to patients by inferring individual and environmental contexts. The purpose of this study is to gain a deeper understanding of the user's current condition and to determine the level of their illness with the addition of the user's participation in receiving personal and environmental contexts, as well as his participation in decision-making. The approach presented in this article can be repeated and implemented in other situations. This method can be applied in different regions and cities or different division levels such as districts or provinces.