Diagnostic and Therapeutic Model for Real Time Management of Diabetes A. E. Babalola O. M. Omisore B. A. Ojokoh Department of Computer Science, Institute of Biomedical & Health Department of Computer Science, Federal University of Technology, Engineering Shenzhen Institutes of Federal University of Technology, Akure, Nigeria. Advanced Technology, Chinese Akure, Nigeria. asegunloluwa@gmail.com Acad. of Science, Shenzhen, China. bolanleojokoh@gmail.com ootsorewilly@gmail.com ABSTRACT Sharma & Majumdar, 68% of women between 21 and 52 years live Diabetes is a major health problem inherent to people at all age with obesity [4]. Obesity occurs when excess fat accumulates in body thus reduces life expectancy. In conventional medicine, groups in developing countries. Conventionally, diagnosis of this obesity is not regarded as chronic disease but it leads to serious condition was based on blood sugar level however, its effect can health conditions like diabetes mellitus, cardiovascular diseases be traced from other symptoms such as Body Mass Index, and (CVD), which have high mortality rates [5]. Blood Pressure. This paper presents a decision support model that can be used by diabetic patients and medical practitioners for The adoption of Information Technology in modern societies has diagnosis and therapy of diabetes. Fuzzy Logic was adopted for experienced a shift of paradigm in health condition management. diagnosing pre-diabetic and diabetic patients’ data from Obafemi Most people consume foods without considering their health state Awolowo University Teaching Hospital Complex, Ile-Ife, Osun- either because they were not properly guided or due to State, Nigeria. Therapy is provided as personalized diet unavailability of medical experts. Medical procedures have been recommendation using person correlation coefficient and users’ supported by technology advancement to minimize the rapid preferences. System evaluation shows adorable performance on growth of chronic diseases like diabetes [6]. Moreover, recommender system (RS) assumes information from beneficiary both operations. or his closer neighborhood to give suggestions for making optimal CCS Concepts decision while faced with different choices [7]. • Information systems ➝Information retrieval ➝Retrieval Recent developments in different fields such as Health and tasks and goals ➝ Recommender systems Commerce RS have adopted the development of expert systems to support their business logics [8]. Such advancement could play a Keywords major role in disease control by providing accurate and reliable component; Diabetes Mellitus; Medical Diagnosis; diagnosis results, acknowledgement of risk status. The central Recommender System; Fuzzy Logic; Diet Personalization problems aim is building artificial life, reasoning and programming, knowledge representation, and understanding 1. INTRODUCTION cognitives of natural language with adoption of Artificial As mechanization, urbanization, globalization, financial and Intelligence (AI) in machines. For instance, Ali & Mehdi [9] social developments led to richer and better life in human daily applied fuzzy concept to reduce risks associated with conventional affairs, modifications and alterations in diets have brought about practices in health diagnosis. AI techniques were mostly applied in greater chances of developing certain diseases (FAO, 2013). Food diagnosis of cardiovascular, parasitic, and viral related diseases choice selection has immense effect on health as hale diets help [10] however, obesity and diabetics has recently received some sustain balanced body weight, enhanced growth, and boost immune recognition [6][11][12]. In Babalola et.al. [6], real time diagnosis system thereby promoting good mental function for daily activities. system was proposed to detect severity of diabetes in patients. Medical research has shown that healthy foods strengthens the In e-Health, RSs can be utilized for therapy by predicting or immune system thereby presenting people a greater chance of recommending types of food people can take to preclude them countering free radicals and warding off diseases[1]. Simply, from certain ailments. As a result, an intelligent meal planning healthy diets are dietary taken to develop and repair body cells and system was proposed in [13] and similarly, Napat et al. [14] tissues for body effective function. Contrarily, poor dietary lifestyle presented a knowledge-based approach for personalized food is a key contributor to development of chronic diseases such as recommender. Furthermore, clustering analysis is adopted in obesity, diabetes, and cardiovascular diseases [2]. Phanich et. al [15] to recommend food items for diabetic patients. In the world today, millions of people suffer from poor health Hsu et al. [16] also developed an online system that searches food conditions as a result of inappropriate diets [3]. For instance composition databases, calculates dietary intake, and provides the obesity, which has potentials of causing more severe problems, has guidance for decision making in nutrition counseling. being a popular health condition around the world. In a finding by Despite the adoption of RS for diet related diagnosis and preventive-therapy, suitable food selection are still difficult for some people especially in the presence of many factors. Hence this work extends the model in [6] with recommender component that predicts type and quantity of foods that can be taken for effective CoRI’16, Sept 7–9, 2016, Ibadan, Nigeria. management of diabetes. The remaining parts of this paper is organized that Section 2 presents review of related works; Section 3 presents the diagnostic RS for managing pre-diabetic and diabetic patients. Experimentation, Results and evaluation of experiments carried out are presented in Section 4. Lastly, conclusion and futures works are presented in Section 5. 64 Figure 1. Architecture of a Fuzzy Based Diet Recommender System 2. SYSTEM DESIGN AND ANALYSIS 2.2 Food Recommendation Architecture of the model for real time management of diabetes is All patients whose data are passed for diagnosis are lamed pre- presented in Fig. 1. The baseline model was described in [6] though diabetic. However, diagnostic component of the model is used without enhancement for recommendation thus, this model shows determine the true health status of patients. People could be lamed capability of diagnosis of diabetes and food recommendation base diabetic if their body weight signifies being overweighed or obese. on patient health data. To determine if a patient is obese, the Body Mass Index value is obtained using Eq 2. 2.1 Fuzzy-Based Diagnosis The system architecture has four major components working inter- connectedly to perform diagnosis of pre-diabetic and diabetic patients and recommendation of food items base on patient diabetic or obesity level. The diagnostic part combines efforts from where w(kg) is the patient’s weight in kilogram, and h 2(cm2) is Graphical User Interface, Knowledge base, and Fuzzy Logic height in centimeter. components to determine patients’ health status. Recommender Then patient is classified as being underweight, normal, component has roles to play when a user demands overweight, or obese base on categorization in Eq 3. recommendation anonymously or as a basis for food therapy. A detailed functions of the first three components that is: User Interface, Knowledge Base and Fuzzy Logic; were detailed in [6]. { In a nutshell, the functions are: a. Patients’ data were collected and stored for processing through graphical user interface in an efficient manner; Therapeutic action is triggered if a person classified obese is also b. Raw and processed data are stored in the database component confirmed diabetic by the diagnostic processes, or request is of knowledge base together with if-then rules upon which voluntarily made for food recommendation. fuzzy logic component operates; and c. Fuzzy inference component operates on user’s data for 1) Recommendation for Diabetic Personality purpose of diabetic diagnosis. Fuzzification and In cases where lamed patients are diagnosed diabetic, the defuzzification were applied to handle imprecise and recommendation component uses Broca Index to compute ideal uncertain information innate in patients’ data. body weight (bw) of the patient. Broca Index is an ideal body mass measurement developed for standard weight computation [17]. The The defuzzification process translates output from fuzzy index value is obtained as in Eq 4. inference engine to crisp values through computationally simple and accurate technique: Centroid of Gravity (CoG). Given as an aggregated membership function with as center of the Broca Index value is used to determine the activity level (AL) of a membership function, the output value is determined as with Eq1. patient. Activity level of a patient has intrinsic characteristic with ∑ total calorie of energy such patient requires daily, and it is determined by combining BMI value of patient with his/her work ∑ category. Classification of works done by people based on expected energy to achieve optimal result is reported in [18]. This research combines classification advice by authors with BMI values in categorizing patients to AL groups, as presented in Table 1. 65 Table 1. Activity level (AL) categorization. recommendation actuates on voluntary request by anonymous user BMI Sedentary Active Very active using Pearson Correlation Coefficient for similarities measures. Obese 30 35 40 Pearson Correlation Coefficient (PCC) measures distance between Overweight 25 30 35 items that are linearly related. Unlike Euclidean distance measure, Normal 20 25 30 PCC observes correlations of variables in range of -1 to +1, hence Underweight 15 20 25 accuracy of score is maintained when data is not normalized. In any voluntary recommendation request, users are obliged to specify a Therefore, a preferred three-square meal ration is applied base on set of food items preferable to them as a meal. PCC is employed to case study disease. In the case of diabetes, 3:4:3 ration is applied for retrieve groups of food-item combination(s) that are found closest Breakfast, Lunch, and Dinner respectively. The basic goal is to eat to user’s specification. For instance if a user specifies an item- healthy foods in reasonable proportions alongside time of the day set and where X,Y or Z are since diabetic patients are keenly monitored to avoid blood sugar food items in a user’s choice, and is a major nutrient in food-item spikes. Since more energy is required in afternoon period, it is X, then groups of food items are recommended from the database important such patients consume more calories in the afternoon following correlation procedures in Eq. 7. rather than morning or night. Summarily, rations needed for ∑ ̅ ̅ consumption by diabetic patients is given as Eq. 5 . . . . . . . . . . . . (7) √∑ ̅ √∑ ̅ where ̅ and ̅ are mean values of nutrients and with confidence where kcal is the kilocalorie for Period of a day, is ration for , and value such that . Two set of food are correlated if AL is activity level category of patient. they have a high confidence value. The periodic kilocalorie intake is further shared to three macro food 3. EXPERIMENTS AND RESULTS nutrients: carbohydrate, protein and fat. Sharing percentages of This section reports details of experiment carried out to validate the macro food nutrients strongly depend on diabetic level of patient. proposed model. Details of the dataset used and results are detailed According to American Diabetes Association [19], low in this part. Results from some related works were taken as basis for carbohydrate meals are good to keep blood glucose levels in performance measure. diabetic patients within normal range, and offer tasty meals that satisfy hunger. Such lowness inversely depends on diabetes level of 3.1 Dataset the patient. Relationships between diabetes level and percentages of The dataset used were sourced from multiple agents including macronutrients in diabetic meals are given in Table 2. Percentage nutritionists, diet related publications and websites. Since diabetic ratios were computed and displayed as users’ guide. The output is a patients can only feed on certain foods, we design a template made seven-day food plan recommendation based on food roster pre- from foods mostly consumed by diabetic people only. The template stored in the database. Dynamism in recommendation depends on is a flexible seven-day calendric roster shaped with the help of history, allergic foods, favorite foods, and diabetic health status of nutritionists. patients. Guided by nutritionists, we established some relationships between Table 2. Diabetic Levels and Macro Food Nutrient Percentage. diabetes level and amount of kilocalories patients can consume Diet Carbohydrate Protein Fat from macro nutrients in foods to provide diet personalization for Normal 60% 20% 20% diabetic patients. Finally, a list of 70 food items consumable by diabetic patients were crawled from diet related websites, and Mild Diabetic 56% 23% 21% analyzed. The macro- and micro- features were elicited and stored Severe Diabetic 50% 26% 24% in the database schematically as shown in Table 3. Very Severe Diabetic 45% 30% 25% These features offer cognitive help in personalized and anonymized The model present patients with substitutes for allergic foods using food items recommendation. Food combination strictly follows a correlation measure, this is detailed in voluntarily request model base on nature combination where nature of food items is recommendation. Finally, for conveniences, food items are assumedly derived from their appearance. Nature can be any of the converted into grams on display to enable patients prepare their options in Table 4. meals correctly and independent of nutritionists. Conversions of the macro nutrients adopted from [6] is presented in Eq 6. 3.2 Experimental Result Data of thirty pre-diabetic patients from Obafemi Awolowo University Teaching Hospitals (OAUTHC), Ile-Ife, Osun State, Nigeria was fed into the system for purpose of diagnosis and food therapy. Aside these macro nutrients, traces from other food nutrients such Table 3. Classification of Food Items for Possible Combination as vitamins, water are also considered for personalization. 2) Recommendation on Voluntarily Request. In first part of the recommendation module, the model exhibits a curative mechanism for pre-diabetic patients who were registered with the system and diagnosed. On another side, the model has a preventive component that predicts food items for users who were never lamed pre-diabetic nor diagnosed by the system. This 66 Table 4. Feature Set of Foods for Personalized Recommendation Macro Nutrient In Food Multi Vitamins Others Description Carbohydrates Protein Fats A C E B6 B12 D Salt Fibre Gram/Serving Nature Figure 2. Patients’ Information Form (Case Study of Patient 013) This is to validate preciseness of personalization in recommendation made by the proposed system. All program codes are implemented with HTML, PHP, JavaScript and SQL. The HTML tags and Java scripts are employed to structure the outlook Figure 3. Result of Fuzzy-Based Diagnosis (Case Study of Patient 013) and behavior of the web pages respectively. At each session, anthropometric data and vitals of patients are captured with web 3.3 Diet Personalization and Recommendation interface displayed as Fig. 2. Following the diagnosis processes In this phase two procedures carried out seamlessly are evaluated by the Fuzzy-Based component, users’ Diet personalization and recommendation of food items for users. Personalization and Recommendation is operated on at actual During personalization, ideal body weight of a patient is used to modules. determine the total Energy Required Daily (ERD) in kilocalories. This includes certain Proportion of Breakfast, Lunch and Dinner Upon successful fuzzification process, the diagnosis result for each (PoB, PoL, PoD) of food items with proportions depending on patient is displayed as in Fig. 3. The interface has two parts: the first selected eating formula, and pre-configured percentage of macro shows result of fuzzy operation on left side, and a summary panel food nutrients for different levels of diabetes as explained in Table on the right side. The later displays diagnosis status and level of 2 of Section 3.To compute ideal body weight of a patient, Broca severity of diabetes in patients. Also on the later side are controls Index estimated from patient’s height is utilized. Then, the total for personalizing foods to be recommended for patients, alongside amount of food to be consumed per day is computed following with patient diet history. Data of the 30 subjects and respective procedures in Section 3. diagnosis are given in Table 5. Diagnosis by fuzzy component shows Patient 013 is severely diabetic, hence appropriate eating Patient 013, in his session with the system, supplied values in Fig. 2 formula to be selected is 3:4:3. and has severe diabetic as diagnosis result. The patient claims a height value of 1.50 hence a Broca Index of 50 was computed and To recommend personalized food items, “Continue to Prediction” thereby, a total of 1750 kcal is to be consumed daily. Still on button in Fig. 3 is clicked. Once the procedure is triggered, the personalization, selected eating formula was applied to determine recommender acquires diagnosis result and portions for breakfast, lunch and dinner as given in Table 6. Table 5. DiAgnosis Result of 30 Pre-Diabetic Patients Table 6. Breakfast, Lunch and Dinner Food Proportion. Breakfast Lunch Dinner Portion Portion Portion PfB = ⁄ * kCal ⁄ ⁄ => 525 => 700 => 525 selected eating formula as essential information needed to fine-tune Furthermore, the three-square meals were subdivided by major the pre-designed seven days template for food personalization and food nutrients: carbohydrate, protein, and fat; hence we applied recommendation. daily proportion for each of breakfast, lunch, and dinner to a suitable diet configuration base on patient’s diabetic severity to compute appropriate proportion of macro food nutrient as in Table 7. 67 Finally, gram equivalents of items in food template were Dinner determined as personalized diet recommendation. Diets recommended by the model are presented to users in a single interface with ranking done based on users’ favorites. This results in having foods that a user likes at top of the list while allergic ones come last in the auto-adjusted 7-day plan. specific alternatives across different food items of the same Nature Pearson correlation explained in Section 3 was used to pair food (See Table 4), and compute correlation score for each combination. items for meals recommended in any seven-day plan. The main For instance, coefficient algorithm combines each of Amala, Eba, role of Pearson coefficient is to generate balanced diet meal by Fufu, Pounded Yam, Wheat, and Semolina; with alternative items observing correlation among food items in the database. In in other food natures to check correlation. combination like Amala + Okra + Mackerel Fish + Orange, each As a result fixed content of food roster template are updated with food item are connected with meal combinations that has best correlation. This is a direct Table 7. Proportion of macro food nutrient (KCAL) for user “013” function of Confidence Value (CV) varying between 0-1. However, Carbohydrate Protein Fat only combinations with were considered for recommendation as in Table 8. Breakfast Another important part is personalization of food items which are recommended by filtering patient’s allergen. Actually, recommendation interface has information displayed in two parts. The left side contains basic data of patient together with information on quantification of food-per-day (FPD), diet Lunch configurations with gram equivalent of foods, and patient’s allergic foods, while the right side is actual recommendation made for patient. For each item in recommended plan, the diet system sorts out available substitutes for allergies using Euclidean distance given as Table 8. Pearson Coefficient Correlation of Food Items S/N Food Item Combination CHO Protein Fat Salt Fibre A C E B6 B12 D CV Amala 20 3 0 0 2.1 57 51 73 55 724 0.5 Carrot 5.61 1 0.3 0 3.1 120 7000 206 220 0 0 1 0.885 Mackerel Fish 5.9 21.1 2.8 5.9 0 0 0 0 0 0 0 Ewedu 0.3 1.8 1.4 0.5 0 37 0 115 129 883 0.5 Boiled Pumpkin Vegetable 1.4 0.2 0 0 0.3 1398 1300 200 0 0 Pounded Yam 20 3 0 0 0.8 73 13 196 13 837 0.3 2 0.883 Beef Meat 0 7 3 1.7 0 14 35 191 21 402 0.6 Carrot 5.61 1 0.3 0 3.1 120 7000 206 220 0 0 Eba 20 3 0 0.3 0.2 82 61 76 0 1286 0.4 Beef Meat 0 7 3 1.7 0 14 35 191 21 402 0.6 3 0.881 Carrot 5.61 1 0.3 0 3.1 120 7000 206 220 0 0 Boiled Pumpkin Vegetable 1.4 0.2 0 0.3 0.3 1398 1300 200 0 0 0 Eba 20 3 0 0.3 0.2 82 61 76 0 1286 0.4 Carrot 5.61 1 0.3 0 3.1 120 7000 206 220 0 0 4 0.876 Mackerel Fish 5.9 21.1 2.8 5.9 0 0 0 0 0 0 0 Boiled Pumpkin Vegetable 1.4 0.2 0 0.4 0.3 1398 1300 200 0 0 0 Pounded Yam 20 3 0 0 0.8 73 13 196 13 837 0.3 Carrot 5.61 1 0.5 0 3.1 120 7000 206 220 0 0 5 0.854 Duck 0.9 19 28 0.2 0 210 0 700 180 3 0 Boiled Pumpkin Vegetable 1.4 0.2 0 0.4 0.3 1398 1300 200 0 0 0 : Table 9. List of Food Substitute(Pap) ̂ ̂ (√∑ ) √ where ̂ and ̂ are alternative food items and are jth respective values of their nutrients. In Fig. 4, Patient 013 indicates allergy to Pap, amongst other recommended food items, a sample of food items filtered out as Finally, the recommendation on voluntarily request only substitute with the same nature is shown in Table 9. Substitutes are requires the Pearson correlation between the food items to arranged in ascending order of their distance measure to Pap as locate a set of foods that are similar to user’ selection. Previous Agidi, Quaker Oats, and Corn Flakes. works emphasized this in different RS [26][27]. 68 3.4 System Evaluation Quantification of the summary is a better way to describe user’s Evaluation is necessary for validation of application systems in preference on recommendation features, hence we had to compute general, but effectiveness is a measure of focus in personalizing a central value for rating purpose. In this evaluation, we assumed diet recommendation systems. Since the system handles two the parameters have equal weights, however weights of Likert important aspects of life, evaluation is separated as diagnosis error scale values differs as in Fig. 4. Therefore, the system has an and user’s preference on recommendation features. For diagnosis, average rating of 74.8%, that is, 187 points out of 250 maximum we checked the sensitivity of underlying mathematical models to points. observe how it responds to inputs. This was done by comparing automated diagnosis results with conventional human approach, 4. CONCLUSION hence utilized Eq. (9) to compute the ratio of properly diagnosed Diabetes Mellitus is a serious health condition that causes mal- patients (True Positive value) to total number of patients diagnosed absorption of foods to be used as energy in human body. It is with the system. costly to manage and as a result, appears as a major factor for high mortality rates in developing countries. Importantly, preventive measures against primary cause, which sometimes is malnutrition or obesity, aid healthy diet lifestyle, improve blood pressure, control blood sugar level and decrease the risk of health Table 10. Diagnosis Results by Proposed System and Expert complications. This paper presents a scalable computer aided model for management of diabetes. Fuzzy logic is proposed for diagnosing diabetic levels in patients, and a personalized recommendation approached towards maintaining balanced macronutrients needed by patients just after diagnosis. Diagnosis stage of the proposed system is basically to determine if a lamed patient is diabetic, in other words to determine the level of diabetes. However, the major proposal in this study is personalization of recommended foods. Recommendation initially emulates a seven-day template which serves as a basic food roster, Therefore, results obtained from manual approach were compared and subsequently modified to suit different patients. Modification with observations by the proposed system as presented in Table 10. is based on diagnosis results of patients hence reaching a goal of On comparing the diagnostic potential of the two approaches, the personalization. The work done is a typical personalization in proposed system demonstrates a sensitivity value of 73.3%, hence which foods recommended varies directly with total ERD by the model’s response to changes in input values is similar to patients. Hence, the recommended seven-day food plan for each human experts. We also evaluated effects of personalization in patient is filtered based on the diet history and allergies of such recommended diets. A routine call was included for dieticians to patients. Different sets of food items are selected and passed for communicate their feedback about performance of the proposed ranking at each session. system with regards to quality of recommended diets and patient personalization levels. This is to ensure reliability of parameters The system has promising diagnosis accuracy and reliable average used in personalization and recommendation of food items. As point for recommendation. Broca index adopted in this research is shown in Fig. 4, five factors were used to observe users’ more efficient at recommending ERD for users, however, might preference, each could attain one of five Likert-scale values: 5- not be the best for very severe diabetic patients, because there is Excellent, 4-Very Good, 3-Average, 2-Fair, and 1-Poor. great need to reduce their carbohydrate intake. Basically, accuracy of the system could be better if calorie recommendation method is adopted. Hence, future work scan adopt switching to Harris Benedict method for recommending daily calorie intake. Furthermore, it is good to note that Harris Benedict method also has limitation of over estimating calorie intake hence, only suitable for patients with very severe diabetes. Figure 4. Evaluation Form for Personalization in Diet Recommendation 5. ACKNOWLEDGMENT Great thanks to Dietician Ogundana T. and other staff at This evaluation part was based on view and responses of 10 Department of Dietetics, OAUTHC, Ile-Ife, Nigeria, for their randomly selected experts on diagnosis and therapy of diabetes at assistance during data collection and interpretation, rule-base Obafemi Awolowo University Teaching Hospitals (OAUTHC), formulation, and model evaluation. Ile-Ife, Osun State, Nigeria. Assessments done by experts using data of thirty pre-diabetic patients is summarized in Table 11. 6. REFERENCES [1] Dharkar S. and Rajavat A. (2011),“Web Data Mining for Table 11. Summary of Responses from Diabetes Experts. designing of Healthy Eating System”,International Journal of Internet Computing, Vol 1(1):19-24. [2] Eyre, H., Kahn, R., Robertson, R., Clark, N., Doyle, C., Gansler, T. and Thun, M. (2004),“Preventing Cancer, Cardiovascular Disease, and Diabetes: ACommon Agenda for the American Cancer Society”, Cancer Journal for Clinicians, Vol54(4): 190-207. [3] Kim, J., Lee, J., Park, J., Lee, Y., and Rim, K. (2009). Design of Diet Recommendation System for Healthcare Service 69 Based on User Information.Fourth International Conference Clinical System for Diabetes Diagnosis, Global Journal of onComputer Sciences and Convergence Information Science, Engineering and Technology, Special Issuein Technology, pp. 516-518. Medicine and Pharmacology, Issue 3, pp. 23-31 [4] Sharma M. and Majumdar P. K.(2009), “Occupational [12] Lu Liu, Jie Tang, Yu Cheng, Ankit Agrawal, Wei-keng Liao, lifestyle diseases: An emerging issue", Indian Journal of AlokChoudhary (2013), Mining Diabetes Complication and Occupational Environment Medicine”, Vol 13(3): 109-112. Treatment Patterns for Clinical Decision Support, [5] Poirier P, Giles T, Bray G, Hong Y, Stern J, Pi-Sunyer F, International Conference on Information and Eckel R. (2006), "Obesity and cardiovascular disease: Knowledge Management, Oct. 27–Nov. 1, 2013, San pathophysiology, evaluation, and effect of weight loss", Francisco, CA, USA. Arteriosclerosis Thrombosis and Vascular Biology, Vol. [13] Aberg, J. (2006),“Dealing with Malnutrition: A Meal 26(5):968-76. Planning System for Elderly”. In AAAI Spring Symposium: [6] Babalola A. E., Omisore O. M., Ojokoh B. A., and Adewale Argumentation for Consumers of Healthcare, pp. 1-7. O. S. (Accepted), “Real Time Diagnostic System for [14] Napat S., Marut B., Ye M. T., Thepchai S., Ponrudee N. Detecting Severity of Diabetes”, International Conference on (2010), “A Knowledge-based Framework for Development of Fuzzy Systems & Data Mining, 11-14, December, 2016, Personalized Food Recommender System”, Fifth International Macau, China. Conference on Knowledge, Information and Creativity [7] Jung, H., and Chung, K., (2015),“Knowledge-based Dietary Support Systems, November 25-27, Chiang Mai, Thailand Nutrition Recommendation for Obese Management”, [15] Phanich, M., Pholkul, P., and Phimoltares, S. (2010). Food Information Technology and Management, pp 1-14. recommendation system using clustering analysis for diabetic [8] Ojokoh, B., Omisore, M., Samuel, O., and Ogunniyi, T. patients. In Information Science and Applications, pp. 1-8. (2012),“A Fuzzy Logic Based Personalized Recommender [16] Hsu, C., Huang, L., Chen, T., Chen, L., and Chao, J. (2011). System”, International Journal of Computer Science and A Web-Based Decision Support System for Dietary Analysis Information Technology and Security, Vol. 2:1008-1015. and Recommendations. Telemedicine & e-Health, 17(2), 68- [9] Ali A. & Mehdi N. (2010), A Fuzzy Expert System for Heart 75 Disease Diagnosis, Proceedings of the International Multi [17] American Diabetes Association (2008), “Standards of Conference of Engineers and Computer Scientists, Hong Medical Care in Diabetes”, Diabetes Care, 31(1):12-54 Kong, March 17-19, Vol I, pp 1-6 [18] Lim-Cheng N., Fabia G., Quebral G., and Yu M. (2014), [10] Mendis, S., Puska, P., and Norrving, B., (2013), “Global Atlas "Shed: An Online Diet Counselling System", DLSU Research on Cardiovascular Disease Prevention and Control in Congress 2014, De La Salle University, Manila, Philippines. Geneva”, World Health Organization in collaboration with the [19] Magee D.J., Zachazewski J. E., and Quillen W. S. (2007), World Heart Federation and the World Stroke Organization. “Scientific Foundations and Principles of Practice in pp. 3–18 Musculoskeletal Rehabilitation”, pp. 349-354, Saunders [11] Afrand P., Yazdani N., Moetamedzadeh H., Naderi F., and Elsevier, St. Louis Missouri. Panahi M. (2012), Design and Implementation of an Expert 70