Development of a dietary menu planning and counselling software with artificial intelligence PhD thesis
Erzsébet Mák SEMMELWEIS UNIVERSITY Doctoral School of Pathology
Supervisor:
Dr. István Szabolcs, University Professor, Head of Department, Doctor of the Hungarian Academy of Sciences
Official Reviewers: Chairperson: External Members: Dr. András Fogarasi, Head Physician, Med. Habil. PhD Dr. Ágnes Szilvás, Head Physician, PhD Internal Members: Dr. Péter Sótonyi, University Associate Professor, PhD Zoltán Balogh, College Associate Professor, Head of Department, PhD. Board of Examiners: Chairperson: Dr. Gyula Tamás, Professor Emeritus, CSc. External Member: Dr. Ákos Pap, University Professor, PhD. Internal Member: Dr. Mihály Végh, College Professor, PhD.
Budapest 2010
Introduction As a preliminary of my scientific work mention must be inevitably made of an earlier work by the Department of Information (IT) Systems at the University of Veszprém, a software based on artificial intelligence developed for the dietary management of cardiologic patients. This software, however, did not quite live up to professional expectations: although the generated menus planned numeric parameters in a fairly reliable way, 5-day menus often failed to be in compliance with common sense either in terms of individual meals or in terms of days. A further problem was to meet variety requirements, in fact, recommendations were made which seemed to fully contradict common sense. I got involved and started to deal with this topic upon the request of the Veszprém developers. It was obvious that the developers possessed a high level of professional know-how, because the professor, associate professors and PhD students of the university, who worked on the software, had already proved their expertise in developing a number of successful expert systems for various disciplines. Thus the conclusion was drawn that in this case the problem might lie in the lack or inaccuracy of such a description of dietary knowledge that can be effectively grasped by the tools of information technology.
Objectives 1. As a dietician I cherished the idea to extend the applicability of this counselling system beyond the scope of cardiologic patients and make it also useable for patients of other so-called lifestyle diseases (diseases of civilisation). For this reason, one of my objectives was to survey the demand for the future software among Hungarian patients suffering from the most characteristic, non-infectious, chronic diseases that can be also influenced by diet, such as cardiologic diseases, obesity, diabetes, oncologic diseases, allergy and coeliac syndrome. I attempted to highlight the demands of the target group of the program to be developed with regards to form of use, depth of information needed, willingness to pay, guarantee of authenticity and time consumption. 2. To develop an artificial intelligence-based dietary menu planning and counselling software which takes individual user needs into consideration, follows up-to-date
dietetic guidelines and, in particular, produces output menus which will not be in contradiction with common sense. Within this context, my direct objectives included: a. To develop a dietary ontology b. To work out harmonising rules in a form interpretable for information technology c. To describe sensory properties in a form interpretable for information technology
Hypotheses on the demands for software use 1. I assumed that the majority of patients interviewed have a preference for use via internet, with the most detailed information possible, free of charge, with medium fill in time and minimum waiting time for the results. 2. The sex of the interviewee does not significantly influence the demand for software use. 3. Age significantly influences the demand for software use. 4. The profession (type of work) of the interviewee does not significantly influence the demand for software use. 5. There are significant differences among disease groups in the demand for software use. 6. The selected form of use (Web, CD, via professional staff) has an influence on the demand for the use of the software system.
Methods Method of demand survey At certain specified times of the period 2006 – 2009 I carried out a survey with crosssectional questionnaires among non-infectious chronic patients being treated at polyclinics. The population surveyed was to include patients suffering from non-infectious chronic diseases. However, due to the lack of accurate data and lists I was unable to perform the sampling procedure in a representative way, thus the results obtained can only be regarded as approximate information. Criteria of inclusion were that the person should appear at the
polyclinic and be older than 18. The patient was assigned to a certain disease group on the basis of the visited polyclinic department where he/she filled in the questionnaire, because this characterised his/her type of need for medical attention at that specific time. Patients newly diagnosed were excluded from the survey, because their first shock reaction might have significantly distorted their response.
Methods used in developing a dietary menu planning and nutrition counselling software 1. Research and analysis of domestic recommendations for non-infectious chronic diseases. 2. Development of a dietary knowledge base by means of the Protege ontology editor on the basis of new systems of aspects. 3. Analysis of the menus generated by the MenuGene program (from the dietetic systems developed so far) in order to accurately reveal the dietetic causes of the problems and find practical solutions to them. 4. Evaluation of widely-used and validated examination methods of nutritional habits in order to design a medical-dietetic history sheet, which can be adapted to the expert system as an input interface.
5. Analysis of the harmonising rules of menu planning discussed in the professional literature and finding their shortcomings.
6. Analysis of the sensory properties (of dishes) discussed in the professional literature and finding their shortcomings.
Results
Results of demand survey The survey confirmed that the sex of the interviewee does not significantly influence the demand for software use. The results of statistical tests did not prove any significance influence of age on the demand for the use of the software. However, age seemed to significantly influence the selected form of use and acceptable waiting time. The profession of the interviewee does not significantly influence the demand for the use of the software. Among disease groups there are significant differences in software demand concerning the form of use, the required depth of information, the guarantee of authenticity, and the
acceptable waiting time for obtaining results. The selected forms of use (Web, CD, via professional staff) do not significantly influence the demand for the use of the counselling system.
Development of the dietary menu planning and counselling software system The raw material data base created on the basis of the USDA National Nutrient Data Base is the largest domestic data base with specification for 190 ingredients and 7454 basic materials and ample pictorial information that can be used both in Hungarian and in English. The recipe data base was based on the knowledge contained in college textbooks. In ontology development we reproduced a significant part of the content of the textbooks by the detailed descriptions of the concepts. When examining the concepts, we did not rely on their position within the hierarchy, which is already easily handled by human logic due to custom, but we concentrated on those necessary and satisfactory conditions, on the basis of which it can be decided if a certain object belongs to the scope of a concept or not. As a first step it was necessary for us to find the most general building block (defined as component), which can serve as a common basis for all grouping aspects. The concept of component is in a partitive relationship with the concept of food (dish): it is contained by it, it is a part of it. With this consideration, it has become simpler to systematise the properties of food. Thus it is possible to decide if a basic material or a technological operation can be replaced by another, and, ultimately, the number of components and their place within the hierarchy can also determine the level of difficulty of the recipe. Also a new approach is needed to process and systematise technological operations. The chronological order followed so far is not suitable for their inclusion in the ontological system. It seemed to be more feasible to order technological operations on the basis of their effects on those properties of the food (dish) which can be matched with various requirements. First these requirements were specified and then the technological operations known were matched to them. The problem encountered here was that the number of opportunities on the basis of which one could determine an operation seemed to be almost infinite. We have drawn the conclusion that it is worth-while specifying various operations only in the form of concepts and if the ontology contains the kinds of raw material that are
capable of meeting the given requirements and knows the quantity units, the genetic algorithm should be able to generate concrete recipes from these data. Harmonising rules are entities with which we can make the menu to comply with taste and common sense logical requirements. These rules are to be applied to dishes within a meal, for subsequent main meals, for subsequent days and for periods of 1 week, 10 days and 30 days. First some general guidelines for sensory properties, technological operations, the frequency of occurrence of components (“knight’s move” rule), taste and traditions were formulated, which can be interpreted by the system on the basis of hierarchy and relations. With the accurate and detailed description of sensory properties we have determined the reasonability of putting a certain food (dish) on the menu plan. In this context, we have worked out an entirely new system of organoleptic indices, which, through scores and the application of a fitness function, selects suitable items to be put on the menu. As input to the expert system we have elaborated a special nutrition medical history sheet to be filled in by the users to an extent relevant to their level of information demand concerning the use of the software.
Conclusions An expert counselling system has been developed which takes the needs of individual users into account with unprecedented accuracy, follows up-to-date dietary guidelines during dietary planning and also produces common sense output menus. On the basis of the test menus generated, significantly better menus have been produced by this system than by counselling programs utilising traditional systematisation principles. The reason for this lies in the developed dietary ontology containing technology and recipe concepts and the description of harmonising rules and sensory properties, which can also be interpreted by genetic algorithms, as well as the creation of the most powerful basic material data base in the Hungarian market with the largest number of items and the most detailed energy and nutritional value data. By means of the algorithms this expert system can generate an almost unlimited number of recipes merely on the basis of recipe concepts. The conceptual formulation of recipes and their material requirements has been made possible by introducing the concept of component, with which we have been able to solve the problem of the
systematisation of food properties, the substitution of technological operations with one another and the decision of the level of difficulty of recipes. The results of the preliminary demand survey have cast light on the fact that the most significant difference in the demand for the use of the software is in terms of disease groups and age. In order to achieve a high level of user satisfaction, it is reasonable to work out multi-level input and output interfaces in the counselling system. By utilising these principles of systematisation and the results obtained, the opportunity is given to develop expert systems for other disciplines, such as sports science.
New scientifically relevant results 1. Introduction of the concept of component: We have defined it as the smallest unit to serve as the common basis of aspects for grouping food for all recipes. 2. Technological concept: For the execution of a certain technological operation we do not specify the accurate list of basic materials but only their groups or, occasionally, their subgroups. Thus the algorithm will recommend a version of preparing the food best fitting to the requirements and will generate unique recipes on its own. 3. Ontology of technological operations: A new system of aspects have been used to set a hierarchy for technological operations. Instead of the existing chronological order, technological operations have been taken into account with regards to their suitability for special requirements. 4. Ontology of food groups: This new grouping system regards the presence of certain food properties as the basis for setting up the hierarchy as opposed to function, technology, basic material, traditions, etc. used earlier. 5. Harmonising rules: A system of rules for dishes within a meal, for subsequent main meals, for subsequent days and for periods of 1 week, 10 days and 30 days has been defined in a form that can be processed by IT tools. 6. Organoleptic index: The effects of raw materials, foods, technological operations are assigned a computed average value (score) based on specified aspects of sensory properties. This will then be rated as good or bad by the program on the basis of a fitness function and inserted in or discarded from the menu, accordingly.
7. A survey on the users’ demands for the planned software has been first concluded in Hungary among non-infectious chronic patients. As different diseases affect the population to various degrees, from a scientific point of view it is remarkable that patients suffering from a certain disease have significantly different demands from patients affected by an another disease group. Consequently, patients should be approached in different ways in the different areas of medical attention in order for us to achieve a high level of patient satisfaction.
Summary An artificial intelligence-based dietary counselling system has been created with the amalgamation of a number of disciplines. Our objective was to develop a counselling system which takes the needs of the users into account with unprecedented accuracy, follows up-todate dietary guidelines and also produces common sense output menus. We have tried to overcome the shortcomings of dietary expert systems developed so far by a new interpretation of the possible aspects of menu planning, an extension of the raw material data base, partial automation of the recipe data base, a conceptual interpretation of technological operations and a formulation and systematisation of harmonising rules. In order for us to achieve these endeavours, new concepts and systematisation methods had to be developed, such as the concept of component, the application of the so called “knight’s move” rule, organoleptic index, and the technological concept. In order to implement all these and to make them usable, new IT methods have been developed as needed. In addition, we wanted to create a marketable software which can achieve a high level of user satisfaction. For this purpose, we have been first to survey the demands of non-infectious chronic patients, and on the basis on their responses, we have worked out our expert system, which could counsel patients, lay users, doctors, nurses, etc. and also support the work of dieticians. We have analysed our hypotheses according to sex, age, disease group and form of use in 6 points (form of use, requirement for depth of information, degree of willingness to pay for use, guarantee of authenticity, time devoted to fill in medical history, waiting time accepted for obtaining a response). We have concluded that sex, age and the selected form of use (Web, CD, via professional staff) do not significantly influence the use of the counselling
system. The patient’s profession (the type of work the patients does) and especially his/her disease, however, have a significant impact on use.
Related publications
Articles: 1. Mák E.: A szívbeteg gyermekek étkeztetése? Élelmezésvezetők Lapja 1998. II. évf. 5. sz. 19. 2. Mák Erzsébet: Cukorbetegek cukra Élelmezésvezetők Lapja 1999. III. évf. 4. sz. 19. 3. Mák E: Időskorban kicsit másképpen?! Élelmezésvezetők Lapja 1999. III. évf. 5. sz. 19-20. 4. Mák E.: Ételkészítés daganatos betegeknek Élelmezésvezetők Lapja 2000. IV.. évf. 11. sz. 28. 5. Mák, E. /Semmelweis Egyetem Eü. Fk. Dietetika Tanszék/ (2001): Sporttáplálkozás, Élelmezésvezetők Lapja, V/4. 27-30. 6. Szalai M., Mák E., Réti A., Szigeti J., Farkas L., Varga L.: Védőgázos csomagolású marinádozott csirkehús vizsgálata, különös tekintettel az érzékszervi tulajdonságokra, Acta Agronomica Óváriensis 2003. 45.1. 69-76. 7. Henter, I.; Mák, E.: Egy felmérés tanulságai, Új Diéta, 2003/II. 13-14. 8. Henter, I.; Mák, E.: Egy felmérés tanulságai, Új Diéta, 2003/III. 14-15. 9. Karamánné Pakai A., Mák E., Németh K., Dér A., M. Szögedi I.: A táplálkozás és a szűrővizsgálatok jelentősége a nőgyógyászati daganatok megelőzésében, Új Diéta, 2008/2 4-5. 10. Karamánné Pakai A., Németh K., Fekete J., Mészáros L., Dér A., Ramona D., S. Ponyókay, Mák E., Bujtor A., Domján P., Balázs P.: A méhnyakrák-szűrés eredményességét befolyásoló tényezők – egy felmérés tükrében; NŐVÉR 2008/21. 2. 3-9. 11. Karamánné Pakai A., Mák E., Németh K., Dér A., M. Szögedi I.: A táplálkozás és a szűrővizsgálatok jelentősége a nőgyógyászati daganatok megelőzésében, Új Diéta, 2008/2 4-5. 12. Mák E., Gaál B., Vassányi I., Karamánné Pakai A., Szabolcs I.: Egészségügyi szoftverek mesterséges intelligenciával – étrendtervező szoftve; Magyar Orvos XVI. 2008.11. 36-38. 13. Karamánné Pakai A., Németh K., Dér A., Mák E., Mészáros L., Lampek K., Oláh A., Balázs P., Investigating the reasons why hungarian women avoid organized screening for cervical cancer. Buletting of Medical Sciences, 2008. 81. (4) 271-275.
14. Mák E., Veresné Bálint M., Pálfi E., Németh I-né, Lichthammer A., Gaál B., Karamánné P. A., Szabolcs I.: Improving the quality of life of the population through the Internet by surveying demand for the use of diet planning software, New Medicine 2010 – Befogadó nyilatkozat mellékelve 15. Mák E, Gaál B., Vassanyi I., Pintér B., Németh I-né, Kozmann Gy.: Interneten keresztüli és manuálisan felvett táplálkozási napló komplexitásának összehasonlító elemzése, Lecture Notes in Computer Science 2010 - Befogadó nyilatkozat folyamatban 16. Mák E., Tóth T., Karamánné P., Fehér F-né.: Az ételcsoportosítás új rendszere, Új Diéta 2010 – befogadó nyilatkozat mellékelve
Books, parts of books: 1. Bencsik, K.; Mák, E.: Hogy fel ne menjen a pumpa – A magasvérnyomásról, Focus Vitalis Kiadó, Budapest, 2003. 2. Mák, E.; Neducza, J.: Fitnes-diéta, Focus Vitalis Kiadó, Budapest, 2003. 3. Bencsik, K.; Fehér, F-né.; Horváth, G; Mák, E.; Varga, Zs.: Mit? Mivel? Hogyan? Ételkészítési technológia és kolloidika (szerk: Bencsik K.), főiskolai jegyzet Semmelweis Egyetem Eü. Fk. Budapest 2003. 4. Bencsik, K.; Fehér, F-né.; Mák, E.: Közétkeztetési szakácskönyv, főiskolai jegyzet, Semmelweis Egyetem Eü. Fk. Budapest 2004. 5. Veresné Bálint M (szerk): Gyakorlati dietetika – 1/3 fejezet, 33-42o.: Mák, E.: A diétás étlaptervezés módszere, szempontjai, Semmelweis Egyetem Eü. Fk. Budapest 2004. Presentations: 1. Gaál B., Vassanyi I., Mák E., Kozmann Gy.: Életmód- és táplálkozás tanácsadó szakértői rendszer bemutatása, 324. Tudományos Kollokvium az MTA Élelm. Tud. Komplex Bizottsága, a Központi Élelm.-tud. Kutatóint, és a Magyar Élelmezésipari Tud Egy. rendezvénye, 2006. április 27. 2. Mák E., Gaál B., és mtsi: Életmód- és táplálkozás tanácsadó szakértői rendszer, Magyar Táplálkozástudományi Társaság Vándorgyűlése, Keszthely, 2006. október 57. 3. Mák E., Gaál B.: Intelligens étlaptervezési szoftver, Magyar Tudomány Napja, Budapest, 2006. nov. 8.
4. B Gaál(1), I. Vassányi(1), Gy Kozmann(1), E. Mák(2), I. Szabolcs(2): Expert System for Lifestyle and Nutrition Counseling - Életmód és táplálkozás-tanácsadó szakértői rendszer (1 Pannon Egyetem, Információs Rendszerek Tanszék, 2 University of Pannonia, Dept of Information Systems Hungary, 2 Semmelweis Egyetem, Dietetikai és Táplálkozástudományi Tanszék Semmelweis University, Dept. of Dietetics and Nutrition Sciences Hungary) BEST for Health - Bio-Economy Sustainable Technology for Health BEST 2006 „Sustainable Development and Innovation Technologies ,- Information- Telecommunication Technologies-for Bio-Economy, Bio-medicine and Health” „Fenntartható Fejlődés és Innovációs Technológiák különös tekintettel az Információs és Telekommunikációs technológiákra, amelyek a biogazdaságot, a bio-medicinát és egyészségügyet támogatják” Budapest, 2006. aug. 1719. 5. Mák E., Gaál B.: Intelligens étlaptervező szoftverek megoldáskereső módszerei, Magyar Tudomány Napja, Budapest, 2007. október. 7.
Poster presentations: 1. M. Bálint, E. Mák, I. Szabolcs, M. Pankotai, K. Németh: Studying the habits of food consumption and the quality of lifestyle in the case of elderly people living in different homes The 6th European Forum for Dietitians (EFAD) Geneva, 2-5 June. 2005. p. 24 2. E. Mák, M. Bálint, I. Szabolcs, K. Horváth, E. Pálfi: Studying the nutrition intake and mental status in the case of elderly people living in different homes The 6th European Forum for Dietitians (EFAD) Geneva, 2-5 June. 2005. p. 20. 3. Mák E., Gaál B., Vassányi I., Szabolcs I.: Igényfelmérés célzott betegek körében intelligens étrendtervező szoftverre; Magyar Táplálkozástudományi Társaság XXXIII.
Other publications Articles: 1. Mák E: Só nélkül finomat? Élelmezésvezetők Lapja 1998. II. évf. 3. sz. 18. 2. Mák E.: Mit ehet, aki „nem ehet”? Élelmezésvezetők Lapja 1998. II. évf. 4. sz. 19. 3. Mák E.: A vizet isszuk, vagy esszük? Élelmezésvezetők Lapja 1998. II. évf. 6. sz. 19.
4. Mák E.:Mit tanítanak az iskolákban? Élelmezésvezetők Lapja 1998. II. évf. 8. sz. 1920 5. Mák E. : Vajon mindenki egyformán várja a karácsonyt? Élelmezésvezetők Lapja 1998. II. évf. 10. sz. 26. 6. Mák E: Burgonyavariációk Élelmezésvezetők Lapja 1999. III. évf. 1. sz. 19. 7. Mák E.: Frissen várjuk a tavaszt Élelmezésvezetők Lapja 1999. III. évf. 3. sz. 8-9. 8. Mák E.: Mit találunk külföldön? Élelmezésvezetők Lapja 1999. III. évf. 6. sz. 18-20. 9. Mák Erzsébet: Azt eszem, amiben hiszek – Hinduizmus Élelmezésvezetők Lapja 1999. III. évf. 11. sz. 30-31. 10. Mák E.: Azt eszem, amiben hiszek – Kóser konyha I. Élelmezésvezetők Lapja 2000. IV. évf. 1. sz. 14-15. 11. Mák E.: Azt eszem, amiben hiszek – Kóser konyha II. Élelmezésvezetők Lapja 2000. IV. évf. 2. sz. 16-17. 12. Mák E.: A vegetáriánus konyha? Élelmezésvezetők Lapja 2000. IV. évf. 3. sz. 26. 13. Mák E.: Karácsonyi ünnepkör, Új Diéta 2001/3-4. 40. 14. Mák, E.: Az élelmezés minőségének fejlesztési lehetőségei a gyermekintézményekben, Élelmezésvezetők Lapja, 2001. V/12. 8-9. 15. Koszonits, R.; Mák, E.: Amit egy dietetikusnak a védőgázos csomagolásról tudni érdemes, Új Diéta, 2003/1. 13-14. 16. Mák E., Fehér F-né: Mit adhatunk egyéves kor alatt a bölcsiben?, Élelmezés 2006/9. 16-18. 17. Mák E., Fehér F-né: Étlaptervezés a bölcsödében, Élelmezés 2006/10. 18-20. 18. Dér A., Karamánné Pakai A., Németh K., Marx Gy., Papp É., Jakabfyné Sági G., Mák E., Vattay P.: Krónikus stressz vizsgálata profitorientált munkahelyen dolgozók körében, Magyar Orvos 2009. 1-2. 36-40.
Books, parts of books: 1. Fehér
F-né.,
Mák
E.:
Táplálkozástani
alapismeretek
a
csecsemő
és
gyermekgondozásban, főiskolai jegyzet, Semmelweis Egyetem Eü. Fk. Budapest 2006. 2. Zoltán Ö. T. (főszerk.), Barna M. (szerk).: Általános dietetikai ismeretek – Egészségügyi felsőoktatási záróvizsga tesztkérdés-gyűjtemény, 14. fejezet 177-227. Bíró Gy., Szépvölgyi G-né., Mák E.: Táplálkozástudomány, táplálkozáspolitika, táplálkozás-epidemiológia Medicina Könyvkiadó, Budapest, 1999.
Presentations: 1. M. Barna, and E. Mak: Breast – feeding in Hungary, Milanopediatria 2000. Nutrition Genetics Environment, from infancy to adulthood Milano 2000. május 25-28. 2. Fehér F-né, Mák E.: Szakirodalmi szemle a közétkeztetés szempontjából, Élelmezésvezetők Fóruma Kaposvár, 2000. április 3. Mák E., Németh I-né: A gyermekélelmezés minőségügyi kérdései - konferencia élelmezésvezetők és pedagógusok részére, Élelmezésvezetők Fóruma, Budapest, 2002. május 4. Mák E., Veresné B.: Az oktatás és a sporttáplálkozás kapcsolat, MDOSZ Szakmai konferenciája, Budapest, 2004. 5. Szabolcs M., Mák E., Frank J.: Hallgatók sportolási szokásai MDOSZ Szakmai Konferenciája, Budapest, 2004. 6. Mák E.: A ketogén diéta elmélete és gyakorlat, NUTRICIA rendezvénye neurológusoknak, Visegrád, 2006. március 31- április 1 7. Mák E.: A helyes táplálási elvek 6 hónapos kor után, MESZK Orsz. Védőnői Tagozat I. konferenciája, 2006. Budapest, június 15. 8. Mák E., Gaál B.: Életmód- és táplálkozás tanácsadó szakértői rendszer, Magyar Táplálkozástudományi Társaság Vándorgyűlése, Keszthely, 2006. október 5-7. 9. Simonova E., Mák E., Kiss E.: A sütőstúdiók szerepe a fenilketonúriás betegek edukációjában,
Magyar
Táplálkozástudományi
Társaság
Vándorgyűlése,
Székesfehérvár, 2009. október 8-10.
Poster presentations: 1. Varga Zs., Román M., Mák E.:Kefir-like dairy products for patients with galactose intolerance, The 5th European Forum for Dietitians (EFAD) Budapest, 4-7. May, 2003. p. 48. 2. Mák E., Havasi A., Richter É., Kiss E., Hacsek G.: A ketogén diéta legfontosabb ismérvei, Gyermekneurológiai Konferencia, Nyíregyháza, 2008.április 24-26. 3. Tóth T., Fehér F-né., Dusa F., Ruda V., Mák E.: Ketogén diétában alkalmazható ételek érzékszervi bírálata Magyar Táplálkozástudományi Társaság XXXIV. Vándorgyűlése Székesfehérvár, 2009. október 8-10.
4. Mák E., Rózsavölgyi K., Répási E., Tóth T., Gilingerné Pankotai M.: Burgonyaételek alkalmazhatóságának vizsgálata ketogén diétában Magyar Táplálkozástudományi Társaság XXXIV. Vándorgyűlése Székesfehérvár, 2009. október 8-10. 5. Mák E., Vikidár E., Ruda V., Tóth T., Gilingerné Pankotai M.: Zöldségételek alkalmazhatóságának vizsgálata ketogén diétában Magyar Táplálkozástudományi Társaság XXXIV. Vándorgyűlése Székesfehérvár, 2009. október 8-10.
