Chapter 2REVIEW OF LITERATURE2.

1       BACKGROUND                                                                                      The growth of ICT (Information CommunicationTechnology) and related application playing major role in business and has opennew interest for researchers. E-Health application has great influence inmedical. E-Health applications assist clinicians in various aspect frommanagement, reporting and financial outcomes. Decision Support Systems havebeen emerged and being increasingly used by many applications, like WebEHR, PMSand E-Health according to their feasibility in automatically extracting usefulinformation and predicting and recommending appropriate results to consumers.Therefore, this research applies a decision support system prototype to enhancethe prediction mechanism for diabetes patients.Ten-Ho Lin and Von-WunSoo 1 presentedtheir research by using three datasets, stat-log heart disease, breast cancerand diabetes and used Minimal Description Length (MDL) principle.

They proposeda greedy search algorithm of the minimum description length to prune the fuzzyARTMAP categories one by one.  Researchconcluded with the experiments that fuzzy ARTMAP pruned have better performancethan MDL principle.JosephL. Breault 2, worked on Pima Indian Diabetic Dataset (PIDD) publicallyavailable UCI Machine Learning Lab. They find the accuracy of data miningalgorithms using diabetic dataset with 8 variables. 392 cases all non-diabeticgave the accuracy of 65.1% in prediction.

ROSETT a tool used for prediction inrough sets. The test sets were classified according to defaults of the naïveBayes classifier, and the 10 accuracies ranged from 69.6% to 85.5% with a meanof 73.8% and a 95% CI. Accuracy of predicting diabetic status on the PIDD found82.6% on the initial random sample, which exceeds the previously used machinelearning algorithms that ranged from 66-81%.

Best services for writing your paper according to Trustpilot

Premium Partner
From $18.00 per page
4,8 / 5
4,80
Writers Experience
4,80
Delivery
4,90
Support
4,70
Price
Recommended Service
From $13.90 per page
4,6 / 5
4,70
Writers Experience
4,70
Delivery
4,60
Support
4,60
Price
From $20.00 per page
4,5 / 5
4,80
Writers Experience
4,50
Delivery
4,40
Support
4,10
Price
* All Partners were chosen among 50+ writing services by our Customer Satisfaction Team

Mean accuracy 73.2% found whileexecuting classifier on 10 random records.Thus, Chang-Shing Lee, Mei-Hui Wang 3 present five layerfuzzy diabetes ontology (FDO) for diabetes decisions with semantic decisionsupport agent (SDSA)that based on knowledge construction mechanism. Proposedsystem construct semantic description decision based on fuzzy rules fordiabetes patient. This novel five layer fuzzy ontologies co-related to generatefinal description as decision result and cross validated with medical staff.

Buche et al. 4 used rules and ontologies inrelational models and designed query schemes based on fuzzy for heterogeneousstructured, imprecise and incomplete data. Lee et al. 5 summarize news byusing fuzzy ontology.

Goncalves et al. 6 proposed inverted hierarchicalneuro-fuzzy BSP system for classification of pattern and extraction of rules indata. Grelle et al. 7 presented architecture that use agent paradigm powerfuland simple to control the hybrid complex environment. Lee et al. 8 developed meeting scheduling agent based on genetic fuzzy.

However, a better known that vague and imprecise knowledge cannotproduce by the classical ontology in several real world applications. Magni and Bellazzi 9proposed a model to note variations relation from blood sugar with respect totime. Quan et al. 10 developed an automatic fuzzy ontology generationapproach for semantic web. Faezeh, Hossien,Ebrahim 11 presented a technique for clustering (FACT) based on patterns. Automaticsemantic help desk support based on fuzzy ontology proposed by Quan et al. 12.

Leonarda and Antonio 13 used MLP (Multilayer Perceptron)neural network to detect diabetic by symptoms in retinal images. Networktrained for the optimal global thresh hold that can minimize pixelclassification errors. System performances are evaluated by means of anadequate index to provide percentage measure in the detection of eye suspectregions based on neuro fuzzy subsystem. Campos-Delgado et al. 14 Designed anddeveloped controller that based on fuzzy expert knowledge to circulate bloodglucose level in a body.

Yager and Petry 15 presented and approach forsummarization of data using ontologies and proposed a framework that resolvescontradictory evidence for decision making. A query enrichment approachproposes by Knappe et al. 16. Thus, Cranfield and Pan 17 built relationshipbetween ontology engineering and model driven architecture. Polat and Gunes 18used principal component analysis and designed an expert system to diagnose thediabetes disease in a patient. For instance, Calegari and Farina 19 conductedscale free analysis with fuzzy ontologies. An intelligent healthcare agent thatuses ontology for respiratory waveform recognition designed by Lee and Wang20.

­2.2       TECHNIQUEDatamining use techniques and process to examine large scale data collectedperiodically. The term also refers to a collection of tools used to perform theprocess.

Collection of tools available to assist the process of Data mining.Use of Data Mining application for chronic disease like diabetes is an emergingresearch field. Data mining algorithm issued for testing the accuracy inpredicting diabetic status. Diabetic status predicted through Data Miningalgorithms and accuracy can be obtained for validating the results.

There iskeen interest in Classification for the researchers of DM and ML. Classificationpredict a class that appears in data but not provided in new instance.Classification train the model on data provided and generated rules and classrelation between the attributes for predication. Algorithm finds relationshipbetween attribute and co-relate the new instances for predication. At the end anew instance with same attributes given to algorithm for the prediction ofclass that is un-seen in new record, thus algorithm have to predict the classof coming record as final result.Radha and Rajagopalan 21 proposed an application based onfuzzy logic for diagnosis of diabetes. Fuzzy sets and linguistic variable areused to predict the diabetes in a body.

Un-certainty dealt by the mathematicalcalculations and with fuzzy logic. Paper also presented with computer basedfuzzy logic with minimum and maximum realtionship values that sepcify fuzzyframe work. To stong more this relationship fourtypatients data collected forexperiments.

Zhou, Purvis and Kasabov 22 presented a technique to selectmembership function that based on statistical analysis for fuzzy neuralnetwork. Experiments done on medical data to validate the proposed method. Diabeteshas two main types that is type-1 and type-2 categorize by American DiabetesAssociation 23. Type-1 is found mostly in children and young adults, wheretype-2 is the most common form of diabetes found in diabetes patients. Intype-2 diabetes patient’s body does not produce enough insulin or producedinsulin does not consumed by the body cells.

Polat et al. 24 worked great to develop a learning systemthat diagnose diabetes. Corchado et al.

25 presented an intelligent schemefor dynamically reporting of nursing task including their activities andmonitoring of patient care. Weng and Chang 26 developed research documentrecommendation and then create user profiles by using ontology. Lee et al. 27presented an intelligent decision support system for CMMI applications thatconstruct automatic ontology for uncategorized documents. Bechhofer et al. 28 workon inflexible web for dynamic linkage using vocabularies and ontologies. Anagent can communicate with other agents in an environment whether virtualentity or physical. Research on intelligent agents increased due to advancementof internet technologies and semantic web.

Kahramanli and Allahverdi30 come up with diabetes classification system using hybrid neural networkapproach that is capable of concepts modeling and describe relationshipsbetween concepts. Ontologies have impact and study in natural languageprocessing, e-commerce, multi agent systems and medicine.Thus Pedrycz and Rai 31 proposed an agent that workcollaboratively for data analysis. Hudelot et al. 32 presented fuzzy relationontology that interprets images. Jeatrakul and Wong 33 conducted aresearch and produced comparison of five binary Neural Network classifiers.

They used three benchmark datasets obtained from UCI machine learningrepository and observed the performance of Back Propagation Neural Network(BPNN), Radial Basis Function Neural Network (RBFNN), General Regression NeuralNetwork (GRNN), Probabilistic Neural Network (PNN), and Complementary NeuralNetwork (CMTNN).    2.3 DOMAINThe main objective of a recommendersystem is to generate recommendation that is yet un-seen based on personspecific preferences. Recommender system generates recommendation with highestrating that is produced with different machine learning techniques. There aredifferent classifications of recommender system by knowledge based filtering,content based filtering, collaborative filtering and hybrid approaches.

There is great interest in researchersfound through literature for automatic disease diagnosis. Sapna and Tamilarasi34 presented a technique based on neuropathy. Due to diabetic mellitus nervedisorder is observed in patients.

Patient suffering with diabetes have effectedby diabetic neuropathy easily. According to the research there are fiftypercent chances patients have disease that effect nerves systems. Body internalorgan like heart, stomach, etc., are known as automatic nerves.

Researchconducted by authors used neuropathy and risk factors to make fuzzy relationsand make an equation for this. After that they used Multilayer Perceptron MLPwith Fuzzy Inference System and liked it with fuzzy relation equation to buildbinary relations.Santi Waulan etal. 35 presented a new classification technique called Multiple Knot Spline(MKS-SSVM) and conducted experiments using PIDD dataset. At first theypresented MKS-SSVM theoretical then compare SSVM application to diagnosediabetes. Effectiveness was observed through these MKS-SSVM techniques throughexperiments conducted for evaluation.

Chang and Lilly 36proposed an approach named evolutionary to extract compact fuzzy classificationsystem.Pardha Repalli37, conducted research and found relation between age groups and activitiesof people who are suffering from diabetes disease. They worked to find outfactors that cause individual to be diabetic.

Statistics given by the Centersfor Disease Control states that 26.9% of the population affected by diabetesare people whose age is greater than 65, 11.8% of all men aged 20 years orolder are affected by diabetes and 10.8% of all women aged 20 years or olderare affected by diabetes. The dataset used for modeling and analysis have 50784total instances with 37 attributes. They computed a new variable age_new asnominal variable, dividing in to three group’s young age, middle age and oldage and the target variable diabetes_diag_binary is a binary variable.

Researchshow that 34% of the population whose age is below 20 years were not affectedby diabetes. 33.9% of the population whose age is above 20 and below 45 yearsnot affected by diabetes. Population whose age greater than 45 were 26.8% andfound not effected by diabetes.

Asmaa S. Hussein, Wail M. Omar, Xue 38 discussedimportance of medical recommender in their research. They presented recommendersystem as a major support in growing medical field.

With the increasing casesof chronic disease healthcare support system providing real time recommendation for medical patient to overcome theloses. DSS provide assistance in controlling the disease and risk analysisprediction for 24/7 remote monitoring. Its challenge to provide real timeaccurate recommendation in medical due to complexity       in data e.g. unbalance, large, noisy and missing data.Rajesh,et.

al.39 proposed a technique by applying C4.5 algorithm and found 91%accuracy by applying on dataset. Priya,et.

al. 40 Comeup with research on Neural Networks using Rapid Miner tool. Their modelsuggested the high classification performance comparatively with other models.

Model produced better accuracy while comparing with other techniques. BlancaS. Leon, et.al.

41 Worked and proposed an model to control type 1 diabetesmellitus. They used Recurrent Neural Networks (RNN) for controlling the glucoseinsulin in patients. RNN experiments control the complexity of glucose verywell in the experiments conducted by researchers.        RajAnand, et.al. 42 Presented an approach and diagnosis with PCA (principlecomponent analysis) and HONN (Higher Order Neural Network), found that PCA is fasterwith coverage and have lower mean square error.

They presented this novelapproach with Pima Indian Diabetes dataset.In another research paper Samantha Sod see 43 describedthe importance of Decision supports system that recommender systems areapplications of information systems providing recommendations on the base of historypreferences and being used vast in the field of commerce, marketing and entertainment.RaviSanakal, et.al.

44 Conducted a research using FCM andSVM (SMO) and comparethe resultant accuracy. Objective of the research was to find out besttechnique in the diagnosis of Diabetes. They found FCM a better predictor thenSVM having 94.3% of significant accuracy comparatively with second one that is59.5%. Inanother research conducted by Veena Vijayan, et.

al 45 worked for themaximization of K-Nearest Neighbor, K-means, Amalgam Knn and Adaptive Neuro FuzyInference System. Study showed that EM (Expectation Maximization) have leastclassification accuracy. Amalgam Knna and ANFIS posses the better classificationaccuracy. They have found that Amalgam KNN have mutual feature of KNN and Kmeans.

Radha,et al. 46 conducted a research of using C4.5, SVM, K-NN, PLR and BLR fordiabetic patient prediction. Research concluded that BLR has highest accuracywith 75% and lowest computing time with 0.

27 error rate.Experiments conducted on dataset chosenfrom UCI Machine Learning Repository 47 that concludes better performance ofproposed technique than K-means clustering.