DSS BASED ON DOCTOR PERSONAL RECOMMENDATIONSON A SPECIFIC DISEAS TO ENHANCE THE REVENUE MUHAMMAD SHAHID(08-arid-538) Baraniinstitute of information technologyPirMehr Ali ShahArid Agriculture University RawalpindiPakistan2018 DSS BASED ON DOCTOR PERSONAL RECOMMENDATIONS ONA SPECIFIC DISEAS TO ENHANCE THE REVENUE by MUHAMMAD SHAHID(08-arid-538) Athesis submitted in partial fulfillment oftherequirements for the degree of Masterof ScienceinComputerScience Barani institute of informationtechnologyPir Mehr Ali ShahArid Agriculture University RawalpindiPakistan2018 CERTIFICATE I hereby undertake that thisresearch is an original one and no part of this thesis falls under plagiarism.If found otherwise, at any stage, I will be responsible for the consequences.StudentName: Muhammad Shahid Signature:_________________RegistrationNo: 08-arid-538 Date: ____________________ Certifiedthat the contents and form of thesis entitled “DSS Based on DoctorPersonal Recommendations on a Specific Disease to Enhance the Revenue” submitted by Muhammad Shahid has been foundsatisfactory for the requirements of the MS (CS) degree.
Supervisor:_____________________ (Mr.Muhammad Asif) Member:______________________ (Dr.Umair Abdullah) Member:______________________ (Dr. Mohammad Jamil Sawar) Date of Viva Voce:________________ External Examiner: _________________Director: _______________________________________Director AdvancedStudies: ________________________ DEDICATION Idedicate this thesis to my parents and my teachers whose encouragement isalways a source of motivation and determination for me. Their support andprayers enabled me to complete my work successfully and they are credited forall my achievements. I am thankful to my friends for their support which alwayscontributed to my work. CONTENTS PageList ofAbriviations ixList of Tables xList of Figures xiAcknowledgements xiiAbstract xiii1 INTRODUCTION 1 1.
1 BACKGROUND 1 1.2 PROBLEM STATEMENT 1 1.3 SCOPE 1 1.4 AIMS AND OBJECTIVE 2 1.5 THESIS STRUCTURE 22 REVIEW OF LITERATURE 5 2.1 BACKGROUND 5 2.2 TECHNIQUE 7 2.3 DOMAIN 103 A PROPOSED DSS FOR DIABETES PREDICTION 14 3.
1 THE OVERVIEW OF PROPOSED DSS 14 3.2 CONCEPTUAL DIAGRAM OF PROPOSED DSS 15 3.3 DETAILED DIAGRAM OF PROPOSED DSS 16 3.4 PROPOSED DSS DESIGN 17 4 RESULTS AND DISCUSSION 19 4.1 EXPERIMENT 194.2 PROPOSED DSS TRAINING /LEARNING 204.
3 PROPOSED DSS PREDICTION 21 4.4 PROPOSED DSS OUTPUT 214.5 PREDICTION WITH PROPOSED DSS 22 CONCLUSION AND FUTURE WORK 23 SUMMARY 23 LITERATURE CITED 24 LIST OF ABBREVIATIONSDSS DecisionSupport SystemBIIT BaraniInstitute of Information TechnologyEHR ElectronicHeath RecordsPMS PhysicianManagement SystemPHR PatientHealth RecordsAI ArtificialIntelligenceSVM SupportVector MachineSMO SequentialMinimal OptimizationKNN K-nearestNeighborsANN ArtificialNeural NetworkMLP MultilayerPerceptronPIDD PimaIndians Diabetes DatasetML MachineLearning LIST OF TABLES Table No.
Page4.1 Summary of Experiments 19 LISTOF FIGURES Figure No. Page 3.1 System Architectural Diagram 143.2 Conceptual Diagram of Proposed DSS 153.3 DetailedDiagram of Proposed DSS 16 3.4 Proposed DSS Inteface 174.1 Rules Snapshow 204.
2 Proposed DSS Training/LearningInterface 214.3 Proposed DSS Prediction Interface 214.4 Proposed DSS Output 22 ACKNOWLEDGEMENTS Immensegratitude to Allah Almighty, the Most Beneficent and The Most Merciful whoblessed me with the desire to seek knowledge and to explore some of the manyaspects of His creation.
I would like to express my sinceregratitude to my supervisor Mr. MuhammadAsif for the continuous support for my research. I am thankful to him forhis patience, motivation, enthusiasm, and immense knowledge. His guidancehelped me in carrying out research work and writing this thesis.
I attributethe completion of my degree to his encouragement and effort and without himthis thesis would have not been possible. One simply could not wish for abetter or gracious supervisor.I would like to thank members of mysupervisory committee; Dr. MohammadJamil Sawar and Dr. Umair Abdullah for their support, invaluable advice andguidance. I am deeply grateful to my father Mr.
Muhammad Afzal and my motherfor immense support and high level of motivation throughout my study career. Iwould like to thank all BIIT staff for their cooperation.I would love to express my sincerethanks to my colleagues and friends who have always been a real source ofmotivation for me. Last but not the least, I would like to thank my familyespecially my loving parents, my sisters, brothers, my wife and my children forsupporting me and encouraging me throughout my life.
ABSTRACT Withthe rapid growth of Health Care Information Systems, there has been muchadvancement at different levels e.g. clinically, healthcare providers orientedand patient specific.
DSS (Decision Support Systems) are based on historicaldata in terms of objectives reality, so success rates are very high in thesetypes of systems. Historical data can beused to generate useful patterns for decision making. Healthcare support systems like recommender systems play a vital role in differentaspects of in-clinic and out-clinic patient care.DSS assistits users in different aspect; an accurate DSS may facilitate its user in predictionmore accurately. In healthcare IT systems DSS help clinical staff moreintelligently on the base of knowledge extracted from historical data toprovide better care of patient. DSS comprises of multiple tools to produce adecision in the healthcare system e.g. EHR (Electronic Health Records).
Thesetools generates different random reminders and alerts that provide both patientand clinicians timely intimation about suffering diseases. DSS has numerousbenefits that includes, assist a care provider for better care of patient,future prediction of patient suffering, producing precaution and quality carethat leads to patient satisfaction. DSS works jointly with EHR, PMS and PHR toincorporate timely and quickly through systematic alerts and guidelines. DSS maywork as a standalone system in any health facility to corporate clinician’sstaff.
Inthis research, a decision support system is proposed to predict a patient orientationtowards a specific disease that is diabetes. In the prediction two standardtechniques are used. First one is C4.5; a decision tree algorithm known as mostappropriate predictor in supervised learning and second one is support vectormachine (SVM). By implementing two famous standard predictors of data mining,we developed a prototype of DSS to facilitate the clinicians, doctors andphysicians to estimate the diabetes orientation in a patient through biologicalreadings.Thisproposed DSS will be integrated with standard HealthCare Systems e.g.
in EHR(Electronic Health Records), PMS (Physician Management System), PHR (PatientHealth Records). Users of this system include Providers, Physicians or anyother prescriber who is using a standard healthcare system in a practice orclinic. By training through a standard diabetes data set, generation of new predictionof patient diabetes is the core research objective of this research. This candirectly or indirectly increase the rating of HealthCare Systems and providersatisfaction over HealthCare system. Overall this system is built on historicaldata to assist its users and for prediction of diabetes rather found positiveor negative.