International Journal of Computer Applications (0975 – 8887)
Volume 132 – No.12, December2015
Study of Diet Recommendation System based on Fuzzy
Logic and Ontology
Shital V. Chavan S.S. Sambare
Department of Computer Engineering Department of Computer Engineering
Pimpri Chinchwad College of Engineering Pimpri Chinchwad College of Engineering
Pune-44 Pune-44
ABSTRACT intelligence and knowledge of the dietician is approximate
People are becoming very conscious about their health. and not exact in nature. The medical parameters are not
Universe is changing at every moment and results in seasonal precise in nature. They may be precise to a certain degree. The
variations. These seasonal variations affect on human body. degree of precision is important and therefore the fuzzy logic
Every individual having their own unique prakriti from their concept is used to represent the fuzziness of the parameters.
birth. Prakriti is one type of energy to maintain physical and Ontology is used to represent knowledge in tree like manner
mental state of human body. Imbalance in these energies with concept and related instances. The use of fuzzy logic
results in illnesses and different diseases. For maintaining with ontology offers a solution to handle uncertainties and
health and prevent diseases, diet is the best medicine in vagueness in information. In diet recommendations most of
Ayurveda. People follow different diet plans recommended the time dieticians tell patients to take food in quantities like
from different dietician. This recommendation is manual use spices in small amount, fruits in large amount, oil in
recommendation, in computer aided world this is taken as moderate amount. So this small, moderate ,normal, low ,high
slow process. For recommending diet instead of manually, and medium are not in number format so represent those
different approaches from computer science are used. In this words in computer language is not easy task and also for use.
paper different existing diet recommendation approaches This paper provides a review on different techniques used for
studies based on fuzzy approaches and ontology and recommending diet for people using fuzzy logic and ontology.
knowledge representation methods, and proposed a method to First terminologies related to Ayurveda are discussed in
recommend diet based on prakriti of person and current section 2, in section 3 fuzzy related concepts are discussed, in
season. section 4 Fuzzy ontology related concepts are discussed, in
section 5 Related work is discussed, in section 6 proposed
Keywords work is discussed and in section 7 at last conclusion
Prakriti, Season, Fuzzy logic, Ontology, Diet
recommendation. 2. AYURVEDA RELARED CONCEPTS
1. INTRODUCTION 2.1 Prakriti
“Diet” is one of the three supporting factors of life, plays a In Ayurveda, each and every individual is unique. Their size
key role in maintaining health and prevention of diseases. In and shape are different, even their physiological and
Modern Science the balanced diet is one which contains a psychological characters and behavior are also different. This
variety of foods in such quantities and proportions that the is because they have different proportion of dosha (vata, Pitta
need for energy, proteins, vitamins, minerals, fats, and Kapha) at the time of birth which decides their
carbohydrates and other nutrients is adequately met for constitution. Once this constitution is set, it is permanent for
maintaining health. According to Ayurveda the diet which that individual [3]. Type of different Prakriti is shown in
nourishes both the mental and physical built is called figure 1.
"balanced diet"[1].
2.2 Season
Health of human being is totally depend on diet. Human body According to Ayurveda year is divided into “six seasons”.
is made from “tridoshas” and is called as “prakriti”. In Each season consist of two months and in every season
Ayurveda "Prakriti (Dosha)" is called as root of body. particular prakriti raises their proportion in human body and
"Prakriti" plays important role in healthy or sick body. When which results in illnesses. For example in winter season vata
all these "Dosha" are in their normal proportions then person prakriti people facing problems related to health like pain in
remains healthy but the moment their proportion is disturbed Joints. Figure 2 shows relationship between prakriti and
person falls sick. In the human body these "Dosha" are present season.
with certain proportion. Disorder in human health because of
disturbance of the combination of these "Dosha"[2]. Seasonal 2.3 DIET
variations affect on human body. As physical activities is “Diet” is the best medicine in Ayurveda to keep maintains
changes seasonality. Proper diet helps to maintain health. health and diet is given in planned way, there is no need for
There are various different techniques used to recommending separate medical treatment [1]. For different prakriti and
diet for people having different diseases also for normal users. depend on season every individual have to take unique diet.
Fuzzy technique giving good results in real word problem to Same diet is not used for number of people.
recommend personalized diet with different uncertainty and
ontology is used to represent this knowledge.
Dietician having experience and skill for interpreting and
analyzing health related issues by seeing person. The
20
International Journal of Computer Applications (0975 – 8887)
Volume 132 – No.12, December2015
change from one person to another; also, the domain experts
opinions are also vary. Hence, the effectiveness of the type-1-
based system will goes down when there are high uncertainty
levels that are related with the diet domain [5].
Type-2 FLSs used to handle the uncertainties in the group-
decision-making process as they can model the uncertainties
between expert opinions using type-2 fuzzy sets. A T2FS is
characterized by a fuzzy Membership function, the
membership value for each element of this set is a fuzzy set in
0, 1 and in between 0 and 1 , in T1FS, where the membership
function value is a crisp number in exactly 0 or 1. The MFs of
T2FSs are three-dimensional and include a footprint of
uncertainty (FOU). To model the group uncertainties T2FSs
provide additional degrees of freedom that can make it
possible, which involve the different opinions and preferences
Figure 1. Types of Prakriti of experts [5]. The T2FSs can model the requirements of a
person specification that is reflective of all the experts’
opinions and which can then be used to provide a good
recommendation for the diet [5].
4. FUZZY ONTOLOGY
Ontology is a method used for representing a piece of
knowledge and its relationship to other pieces of knowledge
[6]. Because ontology is not sufficient to deal with uncertainty
or vagueness, they are extended with fuzzy data to compose
fuzzy ontology, a fuzzy ontology provide a degree of truth to
an ontology, which describes the degree to which an item is
an instance of a class of items, this is useful for constructing a
set of knowledge for a domain [6].
Fuzzy ontology is representing hierarchical relationship
between concepts in a particular domain and which can be
viewed as a graph. As many fuzzy Concepts are rather vague
than precise and the fuzzy ontology is capable of dealing with
fuzzy knowledge [7].
Fuzzy ontology is defined as the tuple <I, C, R ,F, A>[8]. I is
Figure 2. Relationship between Season and Prakriti the set of individuals, also called instances of the concepts. C
is the set of concepts. Each concept C ∈ C specifies a fuzzy
3. FUZZY RELATED CONCEPTS set on the domain of instances, that is, C: I → [0, 1]. The set
of entities of the fuzzy ontology will be indicated by E where
3.1 Fuzzy Logic E = C ∪ I. R is the set of relations. Each R ∈ R is a n-ary
Fuzzy logic is represented using fuzzy sets, sets that express fuzzy relation on the domain of entities so that R: En → [0,
uncertainty, fuzzy logic is based on the concept of 1]. This part of the fuzzy definition appears to apply to
membership degrees and fuzzy logic is designed to relations between concepts, and relations between concepts
mathematically represent uncertainty for dealing with the and instances. A special role is held by the taxonomic
inbuilt vagueness in some domains. Normally logic is relationship T: E2 → [0, 1] among the entities. The C: I → [0,
comprised of only two values i.e. true and false and has its 1] function seems to define the is-a between an instance and a
constraints in dealing with problems related to the real world concept C. F is the set of the fuzzy relations on the set of
domain. Fuzzy logic uses logical values between 0 and 1[4]. entities. A is the set of axioms defined using a suitable logical
language [8].
3.2 Fuzzy Set
In the fuzzy theory, fuzzy set F of universe U is defined by a 5. RELATED WORK
membership function. It is denoted by µF(x) such that μF(x): Chang-Shing Lee, Mei-Hui Wang, and Shun-Teng Lan et al
F→ [0, 1], 1 if x is totally in F, μF(x) = 0 if x is not in F. For [9]. Proposed an Adaptive Personalized Diet Linguistic
any element x of universe X, membership function μF(x) Recommendation Mechanism based on Type-2 Fuzzy Logic
equals the degree to which x is an element of set F. Degree System (T2FLS) and Genetic Fuzzy Markup language
having a value between 0 and 1 which represents the degree (GFML). First, by using domain expert an adaptive dietary
of membership and also called as membership value of assessment and recommendation ontology is constructed and
element x in set F [4]. then a T2FS-based GFML is constructed, the fuzzy
knowledge base and the fuzzy rule base of the proposed
3.3 Type-1 and Type-2 Fuzzy Logic mechanism is describing by using genetic algorithms. The
The type-1 FLSs can operate well under specific operation input type of information is about the eaten meal and
conditions. The linguistic and numerical uncertainties can proposed mechanism is used to improve their unhealthy eating
make problems in determining the exact and precise habit and then become healthier and healthier.
antecedents and consequent membership functions during the
FLS design. As time goes every user behavior and preferences Chang-Shing Lee, Chin-Yuan Hsu, Su-E Kuo, Akio Naito at
al [10]. This paper presented diet recommendation system
21
International Journal of Computer Applications (0975 – 8887)
Volume 132 – No.12, December2015
based on genetic fuzzy markup language to describe the Mei-Hui Wang, and Hani Hagras et al [5]. This paper
knowledge base and rule base of the diet domain, including proposes a new ontology model, which is based on interval
six food categories of some common food. To construct the type-2 fuzzy sets, called type-2 fuzzy ontology, for knowledge
fuzzy food ontology the domain experts first define the representation in the field of personal diabetic-diet
nutrient facts of the common food. GFML-based learning recommend. First, plan the diet goals are taken from domain
mechanism combining the genetic learning mechanism with experts for the involved diabetes and create the nutrition facts
the fuzzy markup language and is carried out to infer the of common Taiwanese food. Second, daily input eaten items
possibility of dietary healthy level for one-day meals. of diabetic patient are taken. Third, the ontology-creating
mechanism constructs a T2FO, including a type-2 FPPO, a
Wahidah Husain, Lee Jing Wei, Sooi Li Cheng and Nasriah type-2 FFO, and a set of type-2 FPFOs. Finally, the T2FS-
Zakaria et al [11]. Proposed a Personalized Diet based system retrieves the built T2FO to recommend a
Recommendation System for Cancer Patients to help patients personal diabetic meal plan.
for manage their daily food intake. The proposed system
combines Case-based Reasoning, Rule-based Reasoning and Chang-Shing Lee, Mei-Hui Wang, Zhi-Wei Chen, Chi-Fang
Genetic Algorithm to recommend suitable diet as per patients Lo, Su-E Kuo, Hui-Ching Kuo and Hui-Hua Cheng et al [12].
information. Using Case-based Reasoning set of diet plans In this paper property of fuzzy ontology is proposed for
taken from the cases which are already present in the system diabetes and cardiovascular diseases. Using type-2 fuzzy sets
are suggested, whereas Rule-based Reasoning is used to filter and fuzzy inference approach and giving dietary healthy level
out cases which are irrelevant from the system and select the from users eating habits. Following table 1 shows study of
approximately matching case to be suggested to the patient. five papers related to diet recommendation systems:
The Genetic Algorithm technique ensures that the diet menus
suggested are customized according to each patient's personal
health conditions.
Table 1. Survey Table
Sr. Paper Title Techniques Input Output
No
1 Adaptive Personalized [Link]-2 FS Eating Healthy
Diet Linguistic Habit diet
Recommendation [Link]
Mechanism Based on fuzzy
Type-2 Fuzzy Sets and markup
Genetic Fuzzy Markup language
Language.( 2015)
2 Genetic Fuzzy Markup [Link] Eaten Health
Language for Diet markup food level
Application.(2011)
language
[Link]
algorithm
3 Application of Data [Link] Cancer Diet
Mining Techniques in a algorithm patient Plan
Personalized Diet Inform-
Recommendation System [Link]-based ation
for Cancer Patients. reasoning
(2011)
[Link]-
based
reasoning
4 A Type-2 Fuzzy [Link]-2 Diet Diet
Ontology and its fuzzy set goal Plan
Application to Personal and
Diabetic-Diet [Link]-2 user
Recommendation. (2010) fuzzy profile
22
International Journal of Computer Applications (0975 – 8887)
Volume 132 – No.12, December2015
ontology
5 Property and Application Type-2 Eaten Health
of Fuzzy Ontology for fuzzy food level
Dietary Assessment. Ontology and
(2010) BMI
4. On collected database fuzzy ontology is apply
6. PROPOSED WORK 5. Using input values Type-2 fuzzy profile is created.
In proposed system as shown in figure 3, type-2 fuzzy logic 6. For personal diet recommendation fuzzy food
and fuzzy ontology is integrated to recommend diet for user. ontology and Type-2 fuzzy profile ontology is
Type-2 fuzzy logic is used to deal with uncertain things which created.
are come from group of different dieticians. Fuzzy ontology is 7. Diet menu is recommended to user.
used to generate food ontology with help of fuzzy concept,
fuzzy variable and fuzzy sets. This fuzzy ontology helps to 7. CONCLUSION
represent uncertain things in efficient manner and knowledge In this paper different diet recommendation system are studied
is extracted based based on fuzzy logic approaches and ontology. The use of
fuzzy logic with ontology will offers a solution to handle
on membership value of variables. Database is created by uncertainties and vagueness in information. From study we
collecting data related to diet for different prakriti and season proposed a method for diet recommendation using type-2
wise diet from different websites, Ayurvedic books and fuzzy logic and fuzzy ontology techniques. Type-1 fuzzy
dieticians. logic having disadvantage of when dealing with more number
Working of proposed system is as follow: of uncertain things increases and not giving proper
[Link]-2 fuzzy logic is used to handle uncertain things
1. User enter their health information.(age, sex, coming from different opinions and overcome disadvantage of
weight, height) . Type-1 fuzzy logic. Ontology is integrated with fuzzy
2. User enters prakriti type. concepts to represent uncertain things.
3. User enter current season.
Ayurvedic Ayurvedic Ayurvedic
Websites Books Doctors
Database
Module
Food
database
User Fuzzy Food
Information Ontology
Type-2 Fuzzy Personalized
User Prakriti personal Food Diet Plan
Type Ontology
Type-2 Fuzzy
Current season Profile ontology
Input Module Processing Module Output Module
Figure 3. Architecture of Proposed Work
23
International Journal of Computer Applications (0975 – 8887)
Volume 132 – No.12, December2015
8. REFERENCES [7] Nilavu Devadoss, Sivakumar Ramakrishnan. 2015.
[1] Dr. Neeru Nathani. 2013. An appraisal of the concept of Knowledge Representation using Fuzzy Ontologies – A
Diet and Dietetics Ayurveda. Asian Journal of Modern Review. International Journal of Computer Science and
and Ayurvedic Medical Science (ISSN 2279-0772) Information Technologies, Vol. 6 (5).
Vol.2, no.1. [8] V.V. Cross. 2014. “Fuzzy Ontologies: The State of the
[2] S. Kher, An jana Jain, M.K. Ojha, G.N. Sharma. 1996. Art. IEEE.
Intelligent Diagnosis of Human Disorders based on [9] Chang-Shing Lee, Mei-Hui Wang, and Shun-Teng Lan.
Ayurveda. 2015. Adaptive Personalized Diet Linguistic
[3] Chaudhary Sunil Kumar and Tripathi N. S. 2013. Study Recommendation Mechanism Based on Type-2 Fuzzy
Of Seasonal Variations In Different Prakriti Persons Sets and Genetic Fuzzy Markup Language. IEEE
With Special Reference To Serum Urea. International Transactions on Fuzzy Systems, Vol. 23, No. 5.
Journal Of General Medicine And Pharmacy (Ijgmp) [10] Chang-Shing Lee, Mei-Hui Wang,Zhi-Wei Chen. 2011.
Issn 2319-3999 Vol. 2, Issue 3, 23-28. Genetic Fuzzy Markup Language for Diet Application.
[4] Tanumeet Kaur, Amardeep Kaur. 2012. Extension of A 2011 IEEE International Conference on Fuzzy Systems
Crisp Ontology to Fuzzy Ontology. International Journal June 27-30.
Of Computational Engineering Research [11] Wahidah Husain, Lee Jing Wei, Sooi Li Cheng and
([Link]) Vol. 2. Nasriah Zakaria. 2011. Application of Data Mining
[5] Mei-Hui Wang, and Hani Hagras. 2010. A Type-2 Fuzzy Techniques in a Personalized Diet Recommendation
Ontology and Its Application to Personal Diabetic-Diet System for Cancer Patients. 2011 IEEE Colloquium on
Recommendation. IEEE Transactions On Fuzzy Humanities, Science and Engineering Research
Systems, Vol. 18, No. 2. (CHUSER 2011), Dec 5-6.
[6] Joshua M. Krbez, Adnan Shaout. 2013. Fuzzy Nutrition [12] Chang-Shing Lee, Mei-Hui Wang, Zhi-Wei Chen, Chi-
System. International Journal of Innovative Research in Fang Lo, Su-E Kuo, Hui-Ching Kuo and Hui-Hua
Computer and Communication Engineering (An ISO Cheng. 2009. A Novel Type-2 Fuzzy Ontology and Its
3297: 2007 Certified Organization) Vol. 1. Application to Diet Assessment. IEEE/WIC/ACM
International Joint Conferences on, vol.3, 15-18.
IJCATM : [Link] 24