Big Data Analytics and Intelligence: A Perspective for Health Care

Half Title Page
Title Page
Copyright Page
Contents
About the Editors
About the Authors
Preface
Chapter 1-Big Data Analytics and Intelligence: A Perspective for Health Care
	1. Introduction
	2. Big Data Overview
	3. Big Data Applications in Health Care
		3.1.1. Levels of Staffing. Staffing levels are set by administrators of the particular organization and these factors are influenced by various forces such as budgetary considerations and features of local nurse labor markets. The administrative departmen
			3.1.2. Outcomes. Capturing and analyzing the patient information helps in generating a summarized report so that it can be used in later stages for better understanding. Even though it has resulted in great success still this method is very challenging be
			3.1.3. Conclusion. A difference can be seen in healthcare sections where staffing is less when compared to institutions where staffing is more. Most of the researches that were conducted suggest that if nurses appointed are less than required it creates u
		3.2. Electronic Health Records. Needs and Advantages
			3.2.1. Introduction. The most important task of EHR is to help in understanding the medical background of patients with the help of electronic mechanism rather than using traditional techniques of maintaining papers or folders. This helps in reducing time
			3.2.2. Importance for Improving Efficiency and Productivity. One of the main aims of maintaining EHR is that it helps to retrieve information’s regarding the patients whenever required. Lab results can be gathered from decades ago with less amount of time
			3.2.3. Application. The application of EHRs ranges from government sectors to financial sections of various industries. Few of the applications and the expected outcomes from the particular industries are as follows.
			3.2.4. Aggregated Data. Since decisions are to be taken based on the past experiences, most of the organizations collect high quality data in raw format. These data are mainly procured from the data collected from inpatient and outpatient data and details
			3.2.5. Integrated Data. The main disadvantage of maintaining paper-based health records in that it can be used to combine other paper health records and store as the same. Since this mechanism lacks the ability to integrate with other paper forms of infor
			3.2.6. Conclusion. In order to modernize the infrastructure in healthcare sector it is required to adopt and implement EHRs-based systems. A survey conducted to understand the importance of EHR shows that it helps to identify patients with serious health
		3.3. Enhancing Patient Engagement
			3.3.1. Introduction. The healthcare industry like any other sector of the society works mainly to gain profit and survive in the business field. Since patients are the most important factor in the healthcare institutions it is important to ensure they are
			3.3.2. Patient-reported Outcomes. Most of the people who invest in a healthcare sector are mainly interested in improving and expanding the existing business. There are few techniques that are followed to increase the profit. One of the techniques is by c
			3.3.3. Values of Patient-reported Outcomes. By identifying the needs of the stakeholders and other staffs in a healthcare organization has helped in understanding the importance of report-based mechanism. Most of the surveys conducted on report-based deci
			3.3.4. Conclusion. The development of technology and mechanisms used for treatments have helped in getting much better and appropriate accurate way of solving diseases. It has also helped in identifying methods by which health conditions can be improved w
		3.4. Big Data to Understand Cure for Cancer
			3.4.1. Introduction. In the healthcare-related industry, the concept of big data was introduced recently. Big data analytics has helped in taking many positive measures to improve the procedures involved with patient care. It allows a more efficient depen
		3.5. Predictive Analytics in Health Care
			3.5.1. Introduction. The main aim of predictive analytics in health care is to help various organizations and healthcare sectors to discover the data and convert it into information that can be used to improve business decisions (Al Mamoon et al., 2013).
			3.5.2. Application of Predictive Analytics. Critical care intervention provides surveillance type of mechanism for solving and making alerts so that it will help in reducing risk faced by patients from infections and unwanted drug allergies.
			3.5.3. Text Mining Medical Records. With the increase in the use of EHRs has led to a situation which has forced to adopt data-mining techniques to understand the data retrieved from various health reports. The content of the Electronic Health Record (HER
			3.5.4. Conclusion. The decision support systems mainly search for large percentage of unstructured text and use them for decision-making after it is analyzed and made meaningful. These data after analyzing are stored in databases so that it can be used in
		3.6. Need for Security and a Mechanism to Reduce Fraud in Big Data
			3.6.1. Introduction. There is a huge need for big data in health care as well, due to rising costs in countries like the India. Studies show that there will be an increase in the demand for implementing big data analytics tool for improving the healthcare
			3.6.2. Three Levels of Security. The security tools need to be implemented at three various levels that are not present in the network. These levels are as follows:
			3.6.3. Challenges in Securing Big Data. Few of the challenges that are faced in securing big data are as follows:
			3.6.4. Conclusion. From various studies and research conducted it enables to understand that the end-users are mostly responsible for data after it is being converted into useful information. Most of the users of big data implements security measures but
		3.7. Telemedicine
			3.7.1. Introduction. Telemedicine is the process that includes giving and taking data in return regarding various drugs from one site to another through electronic medium. The main aim of telemedicine is to help in improving the health condition of the pa
			3.7.2. Application of Telemedicine. The origin of telemedicine mainly started by providing assistance through communication-based technology. Most of the research based on the history of telemedicine indicates that it was first used for delivering prescri
			3.7.3. Conclusion. As days passes the risk involved with various diseases are increasing and alarming. Technological devices are being invented to avoid contact directly with the patients and these devices are being updated day by day. This technological
		3.8. Applications of Big Data
	References
Chapter 2-Big Data Analytics in Health Sector: Need, Opportunities, Challenges, and Future Prospects
	Introduction
	Big Data
	BD Definitions in the Health Sector
	BD Needs in the Health Sector
	The Health Care Analytics Environment
	EHRs
	EMRs
	Sensor Data
	Internet of Things
	BDA Techniques, Tools, and Technologies in Health Sector
	Opportunities in Health through BDA Use
	Challenges and Strategies
	Few Strategies to Overcome the Challenges of BDA in the Health Sector
	Conclusion and Prospects
	References
Chapter 3-Use of Classification Algorithms in Health Care
	Introduction
	Data Mining in Health care
	Classification Algorithms used in the Healthcare Industry
		Determining the Value of K. In the KNNs algorithm, finding the optimal value of K is of utmost importance as it influences the result and accuracy obtained on the data to a great extent. The k-value in KNN is referred to as the instances in the training s
		Naïve Bayes Algorithm
			Understanding the Working of a Naïve Bayes Classifier. The basic assumption made while building a Naïve Bayes Classifier is that each attribute is independent of the other in the prediction of the output variable. Moreover, none of the attributes make an
		Support Vector Machines
			Understanding the Concept of Hyperplanes. In simple terms, the data points in the training set of a particular dataset define a vector object termed as the hyperplane. Conventionally, the maximum margin hyperplane in SVMs is used in classification problem
		Decision Trees
			Entropy. For a particular finite set S, Shannon’s Entropy is denoted as H(s). The uncertainty of the data is measured by entropy. Entropy can be defined by equation 7.
			Information Gain. The splitting of data is followed by a decrement in value of Entropy which decides the Information Gain. The information serving with the greatest value of information gain makes the best decision tree which means that the branches of th
			Gini Index. A population is said to be pure if two items which are selected randomly belong to the same probability and class. The basic working of a Gini Index is based on categorical variables, namely, True or False. The homogeneity in the data is said
		Chi-square
		Pruning
		Random Forest
			Training a Random Forest Classifier. The training of the random forest classifier takes place through the technique of bagging. A bagging technique is also termed as bootstrap aggregation which involves random sampling of the training set of data in such
		Logistic Regression
		Sigmoid Function
			Understanding the Hypothesis Representation of Logistic Regression. The hypothesis representation of linear regression is shown in equation 12.
		Concept of Cost Function
		Neural Networks
		Activation Functions
		Linear Activation Function
	Sigmoid Activation Function
		Hyperbolic Tangent Activation Function
		Rectified Linear Unit Activation Function
		Leaky Rectified Linear Unit Activation Function
		Softmax Activation Function
		Gradient Descent
		CNN and RNN
		Ensemble Learning
		Boosting
		Bagging
		Combination Methods
			Averaging. This is the most popular combination method. The various types of averaging are explained below:
			Voting. Voting is considered to be among the most common and fundamental methods for nominal outputs. Various techniques are:
			Stacking. Stacking is a general method of training a learner on the combination of the output of individual learners, called first-level learners or base learners (Wolpert, 1992). The combiner is referred to as meta learner or second-level learner.
		Model Evaluation
			Confusion Matrix. Confusion matrix, also known as error matrix (Stephman, 1997) is the most commonly used method of model evaluation. It is a matrix that is used to describe a model’s performance. For example, given a binary classification problem where t
		Receiver Operating Characteristics Curve
		E-health
			Technologies in E-health. The basic working of e-health is generally supported by the government and various partners at different nationality levels. These partners at the country and global level contribute to endorsing the application of Information an
			Advantages of E-health. The consumers and providers of e-health are all subjected to the benefits as much as their barriers. The benefits of e-health are discussed hereunder:
	References
Chapter 4-Big Data Analytics in Excelling Health Care: Achievement and  Challenges in India
	Introduction
	Importance of Data in Modern Society: Data as Fuel of Modern Economy
	Concept and Emergence of Big Data
	Big Data Analytics: Fundamental Concepts
	Objectives
	Research Methodology
	Analysis I
	Analysis II
		Historical Perspective of Big Data Analytics
		Typology of Big Data Analytics
		Analysis III
		Applications of Big Data Analytics in Health Sector
		A. R&D in Heath Sector (Descriptive Analysis: It Would Be Manifested Largely by Descriptive Analytics)
		B. Treatment Protocol Perceptive Analytics
		C. Advance Awareness Campaigns Based on Predictive Analysis
	Achievements
	Challenges
	Conclusion
	References
Chapter 5-Predictive Big Data Analytics in Healthcare
	1. Introduction
	2. Sources of Big Data in Healthcare
		2.1. Clinical Data
			2.1.1. EHRs/EMR. EHRs or Electronic Medical Records (EMR) are an outcome of the digitalization of healthcare services provided by both government and non-government service providers. A lot of vital information regarding the patient, his disease, diagnosi
			2.1.2. Lab Testing. A large number of hospital information systems and patient diagnostics are based on laboratory test reports. Laboratory testing involves testing a specimen and consequently involves the generation of a lot of data. In addition to patie
			2.1.3. Omics. Omics is a field in biology involved in the study of the constitution of the cells of an organism encompassing various molecules inside it. For example, Genomics is concerned about understanding genes and the functions that they perform, Pro
			2.1.4. Administrative Data. Generally, administrative data in healthcare is generated at each point in a healthcare system. Administrative data contain patient registration details including the demographics of the patient, pharmaceutical data based on th
		2.2. Claims Data
			2.2.1 Medical and Prescription Claims. Medical claims data, usually with third party agencies like insurance firms can be used as a Big Data source for prediction of several health issues and their possible indicators. Typically, a healthcare claims form
		2.3. Clinical Research Data
			2.3.1. Clinical trials. Clinical trials (Clark, 2019) are often conducted in healthcare for testing new drugs, vaccines, or therapies. With diseases becoming resistant to existing drugs, it is becoming overwhelming for healthcare service providers to incr
		2.4. Patient – Generated Data
			2.4.1. Social Media Data. Digital data sources – mobile, social, real-time – can provide significant information about the disease and health dynamics of the population. “Digital Disease surveillance” harnesses digital data to monitor and predict disease
			2.4.2. Wearable Sensors. Wearable devices are gaining adoption in the healthcare field. They aid in continuous and remote monitoring of patients, thereby reducing the cost of administering healthcare, preventing emergencies and improve the quality of care
	3. Areas of Application of Predictive Big Data Analytics in Healthcare
	4. Advantages of Predictive Big Data Analytics in Healthcare
		4.1. IT Infrastructure Benefits
		4.2. Operational Benefits
		4.3. Organizational Benefits
		4.4. Managerial Benefits
		4.5. Strategic Benefits
	5. Challenges of Predictive Big Data Analytics in Healthcare
		5.1. Infrastructural Concerns
			5.1.1. Inadequate Integration of Healthcare Systems. As is evident from different sources of Big Data prevalent in the healthcare sector, most of the data come from a variety of different data owners. A big concern in healthcare, especially in a developin
			5.1.2. Poor Management of Healthcare Information. There are two perspectives to this scenario. The first perspective points to the lack of availability of adequate Information Technology (IT) infrastructure, to manage the information about healthcare. The
			5.1.3. A High Volume of Digital Healthcare Data Inflow with Increasing Velocity. Big Data accumulated from a wide range of healthcare big data sources is increasing at an unprecedented rate. This can be attributed to the slow but evolving process of digi
		5.2. Challenges Arising Due to the Technique
			5.2.1. Need to Address “Heterogeneity” in Data. Big Data-based predictive algorithms for healthcare applications need to take into account the heterogeneity of data (Calster, Wynants, Timmerman, Steyerberg, & Collins, 2019). The heterogeneity in data impl
			5.2.2. Predictive Models Aim at Discrimination, Rather Than Calibration. Most of the predictive Big Data algorithms for healthcare applications aim at the identification of discrimination in data, for example, they predict who may or may not have a diseas
			5.2.3. Need for External Validation. The most common testing technique for validation of predictive algorithms is to test run the same on test data. These data are obtained by splitting the input data set into test-training data. There is a need to extern
		5.3. Data-related Concerns
			5.3.1. Data Quality and Integration Issues. There is a dearth of standardization and common protocols for healthcare data. As such, there is a dearth of metadata available on a particular algorithm. This limitation restricts assessment of data quality iss
			5.3.2. Data Maintenance and Storage Issues. The healthcare data owing to security and privacy concerns cannot be made publicly available. This increases the cost of accessing the data. Additionally, this restricts the development of data standardization a
		5.4. Security/Privacy Concerns
			5.4.1. Informed Consent. Patient privacy and security are of the utmost concern in healthcare. This therefore poses a huge challenge to healthcare analytics applications as the data has to be obtained and shared with third party agencies and patient may n
			5.4.2. Data Protocols. In healthcare, there is no appropriate data governance structure. There are no established protocols for structuring, storing, and sharing the data in the healthcare structure. Furthermore, due to the lack of standardization, differ
		5.5. Organization-related Concerns
			5.5.1. Organization Culture. About 60 per cent of the healthcare organizations fail to develop an overall vision and a strategy for the implementation of analytics. Resultantly, the top management, as well as the entire organization, has no idea about the
			5.5.2. Staffing and Training. The field of analytics in all industries is facing acute shortage of skilled professionals. It is important to note that there is a dearth of skilled professionals and managers who can draw knowledge and interpret the outcome
	6. Conclusion and Future Scope
	References
Chapter 6-Smart Nursery with Health Monitoring System Through Integration of IoT and Machine Learning
	1. Introduction
	2. Literature Survey
		2.1. LDR Sensor
		2.2. Moisture Sensor
		2.3. pH Sensor
		2.4. Air Quality
			2.4.1. PPD42NJ Particle Sensor Unit. In view of the light dispersing strategy, it identifies airborne particles ceaselessly. Heartbeat yield that relates to fixation per unit volume of particles can be acquired, with utilizing a unique location strategy d
			2.4.2. PS2 Pollen Sensor. This sensor distinguishes the molecule with light dispersing technique, utilizing one light producer and two light receptors, and segregates Japanese cedar or cypress dust from different particles with two variables “dissipated l
			2.4.3. MQ2 Sensor. The Grove – Gas Sensor (MQ2) module is (Prabaharan et al., 2017) valuable for gas discovery. It is appropriate for identifying H2, LPG, CH4, CO, alcohol, smoke, or propane (Anson et al., n.d.). Because of its high affectability and qui
			2.4.4. MQ9 Sensor. Similar to MQ2 sensor, MQ9 module is valuable for gas recognition. It is appropriate for distinguishing LPG, CO, and CH4. Because of its good affectability and faster response time, estimation can be taken at the unique time. The affect
			2.4.5. MiCS-2714 Gas Sensor (NO2). The essential rule behind the SGX metal oxide sensors is that the opposition of the recognizing layer in sensor changes within the sight of the objective gases. Or then again in basic terms, decreasing gases evacuate a p
			2.4.6. MiSC-2614 Gas Sensor (Ozone). The MiCS-2614 is a reduced ­Modern Ozone Solution (MOS) sensor. The MiCS-2614 is a powerful Micro-­Electro Mechanical Systems (MEMS) sensor for ozone discovery, similarly for gas spill location and open air quality obs
			2.4.7. PM 2.5 Sensor. PM 2.5 laser dust sensor is an advanced widespread molecule fixation sensor, it can be utilized to acquire the quantity of suspended particulate issue in a unit volume of air within 0.3 to 10 microns, in particular, the concentration
			2.4.8. Carbon Filter. Carbon channel alludes to a technique for filtration that utilizes activated carbon to remove the contaminants of water or air. Carbon channels utilize a procedure where the poisons cling to the carbon particles as the water or air g
			2.4.9. HEPA Filter. High-efficiency Particulate Air (HEPA) channels in vacuum cleaners will, in general, utilize two very extraordinary components to clean the airstream. To begin with, there are one or more external channels that work like strainers to s
			2.4.10. Normal AC Filter. An air filter is generally made of a spun fiberglass material or from creased paper or fabric encased in a cardboard casing. Its basic function is to clean the air that flows through your warming and cooling framework.
		2.5. Conductivity Sensor
		2.6. Temperature Sensor
		2.7. Pressure Sensor (Barometric Pressure)
		2.8. Pressure Sensor (Soil Pressure)
	3. Methodology
		3.1. Data Acquiring and Preprocessing
		3.2. Data Modeling
	4. Proposed Solution
	5. Results
	6. Conclusion
	7. Future Scope
	References
Chapter 7-Computer-aided Big Healthcare Data (BHD) Analytics
	1. Introduction
		1.2. Methods and Technology Progress in Big Data
	2. Motivation
	3. Anatomy of Big Data
	4. Benefits of BHD Analytics
	5. BHD Applications in Real Clinics
		5.1. Big Health Data Applications
			5.1.1. Prediction/Diagnosis. The actual magic of big healthcare lies in early, non-invasive, timely, and precise diagnosis of the disease. The number of smartphone health apps has touched the seventh sky, and Apple Store contains 325,000 health mobile app
			5.1.2. Human Genomics/Genome-wide Association Studies and Expression Quantitative Trait Loci. The inspection of a genome-wide set of genetic variations in diverse persons to understand the correlation between the variants with the corresponding trait lies
			5.1.3. Epidemics/Internet-based Epidemic Surveillance. It is now a fact that the number of related searches is high as soon as the sum of persons possessing the widespread symptoms is tall. The study offered by Google on influenza is the perfect example i
			5.1.4. Recommendation System in Health Care/Recommendation. In the recommendation, the collaborative filtering method is well-thought-out a category of recommender system that estimates operators’ lookouts round an object denoting the same to a wide clust
			5.1.5. Sensor-based Health Condition and Monitoring. With the advancement in the hardware of smart sensors, the sensors are produced at an exponential rate. The healthcare monitors the patients remotely using the data generated by smart sensors. The IoT v
	6. Conclusions
	References
Chapter 8-Intrusion Detection and Security System
	1. Introduction
	2. Literature Review
	3. System Architecture
		3.2. Infrared Sensor
		3.3. MQ-2 Sensor
		3.4. LDR Module
		3.5. DHT11
		3.6. Relay
	4. Hardware Assembly and Implementation
		4.1. ATmega328P
		4.2. Infrared Sensor
		4.3. MQ-2 Module
		4.4. LDR Module
		4.5. DHT11 Module
	5. Working
	6. Future Scope
	6. Conclusion
	References
Chapter 9-Decision Making with BI in Healthcare Domain
	1. Introduction
	2. Vision
	3. Main Contribution
	4. Business Intelligence
		4.1 Evolution of BI
		4.2 BI in Healthcare Industry and its Benefits
	5. Literature Survey
	6. Problem Identified
	7. Proposed Solution
		7.2 Data Warehouse Design
			7.2.1 Fact Table. It is at the center of star schema and consists of facts, metrics, or any quantitative information. It also contains the primary key of every dimension table present in star schema. It contains Drid and Pid which is primary key of variou
			7.2.2 Dimension Table. A dimension table stores either the attributes of a dimension or its hierarchies. Dimension tables are generally small in size and contain text values or descriptive values. They contain a single unique id(primary key) which links t
		7.3 Dataset Used
		7.4 Extraction, Transformation, and Loading Process
	7.5 Implementation
		7.5.1 At Doctor Level. To show how BI solutions are beneficial at doctor level an illustration is shown. Considering the query-with the size of tumor present in patients, the nature of the cancer can be studied such that, appropriate treatment and medicat
			7.5.2 At Organization Level. In this section, an attempt is made to show how important and beneficial it is to have BI at organization level. It can help organizations to study the kind of patients they treat, for example, whether they are emergency cases
	8. Conclusion
	References
Chapter 10-Assistance for Facial Palsy using Quantitative Technology
	1. Introduction
		1.2. Background
	2. Literature Survey
	3. Problem Identified
	4. Comparative Study of Already Existing Solution
	5. Proposed Solution
	6. Pros and Cons of Solution
	7. Conclusion
	8. Future Scope
	References
Chapter 11-Constructive Effect of Ranking Optimal Features Using Random Forest, Support Vector Machine and Naïve Bayes for Breast Cancer Diagnosis
	1. Introduction
	2. Related Work on Breast Cancer Prediction
	3. Machine Learning Classifiers
		3.1. Random Forest
		3.2. Support Vector Machine
		3.3. Naïve Bayes
	4. Statistical Analysis
	5. Proposed Methodology
	6. Experimental Results and Discussion
	7. Conclusion
	References
Chapter 12-Intelligent Establishment of Correlation of TTH and Diabetes Mellitus on Subject’s Physical Characteristics: MMBD* and ML Perspective in Healthcare
	Introduction
		Stress
		Migraine
			Causes. The main cause of migraine is drinking. Drinks, especially wine and too much caffeine, are also the main cause of migraine. Stress at work or home can also cause this problem. Getting too much sleep is or getting too less sleep both is harmful and
		Tension-type Headaches (TTH)
			Causes of TTH. A headache is a type of tension that can be caused by different types of foods, activities, and so on. The major causes of headaches types are smoking, eye strain, fatigue, alcohol, cold or flu, dry eye, caffeine and sinus infections.
			Symptoms of TTH. Symptoms of TTH include pressure around the forehead, dull headaches, tenderness, and scalp. TTH pain is usually moderate, but sometimes it can be severe, depending on the duration of the headache you are facing (Chauhan et al., 2018d; Ra
		Obesity
		Coronary Artery Disease (CAD)
			Symptoms. A type of chest pain is associated with heart disease. Angina can cause the following emotions in the chest:
			Treatment. There is no cure for CAD. However, there are several ways to manage things. Treatment includes lifestyle changes such as healthy eating, quitting smoking, and regular exercise. But some people need medicine and treatment.
		The Cases When to See a Doctor
		AI in Healthcare
		Machine Learning
		ML in Healthcare
		Big Data (BD)
		Applications of BD in Healthcare
		Internet of Things (IoT)
		IoT in Healthcare
		Role of Technology in Addressing the Problem of Integration of Healthcare System
		Literature Survey/Previous Findings
	Our Experiment Results, Interpretation and Discussion
		About the Study and Analysis
	Results and Discussion
		Diabetes and Insulin Consumption
		A Novelty in Our Work
	Future Scope, Possible Applications, and Limitations
	Recommendations and Consideration
	Conclusions
	References
Chapter 13-A Machine Learning Approach Toward Meal Classification and Assessment of Nutrients Value Based on Weather Conditions
	Introduction
	Role and Value of Nutrients in Food
	Life-threatening Diseases Caused by Unhealthy Food
	Effect of Weather on Food
	Summer Eats
	Winter Eats
	Spring Eats
	Autumn Eats
	Food Security
	Problem Identified
	Malnutrition
	Carbohydrates
	Deficiency of Carbohydrates
	Deficiency of Fats
	Deficiency of Protein
	Deficiency of Vitamin
	Deficiency of Minerals
	Machine Learning
	Advantages and Disadvantages of SVMs
	Deep Learning
	Proposed Solution
	Future Scope
	Conclusions
	References
Chapter 14-Telehealth: Former, Today, and Later
	Introduction
	Process of Evolution
	History of Medicine
	Pre-Historic ERA
	Early Civilization
	Modern History
	Evolution of Telehealth
	What is Telehealth?
	What is Telemedicine?
	Business Models of Telehealth
	Optimized Delivery in Telehealth Care
	Barriers to Telehealth
		Literature Review
	Methodology
	Results
	References
Chapter 15-Predictive Modeling in Health Care  Data Analytics: A Sustainable Supervised Learning Technique
	1. Introduction
	2. Predictive Analytics in Health Care
	3. Techniques for Predictive Modeling
		3.2. k-Nearest Neighbors
		3.3. Naïve Bayes Classification Modeling
		3.4. Decision Trees
		3.5. Linear Regression
		3.6. Logistic Regression
	4. Applications of Predictive Modeling in Health Care
		4.1. Disease Diagnosis and Treatment Selection
		4.2. Health Care Management
		4.3. Reducing Health Care Costs
	5. Conclusions
	References
Index