Smart Healthcare Systems

Adwitiya Sinha, Megha Rathi, Adwitiya Sinha, Megha Rathi

• Publisher: Chapman and Hall/CRC
• Year: 2019
• ISBN:
  9780429670282
• Edition: 1
• Topic: Informatik
• Subtopic: Informatik Allgemein

About the Book

The book provides details of applying intelligent mining techniques for extracting and pre-processing medical data from various sources, for application-based healthcare research. Moreover, different datasets are used, thereby exploring real-world case studies related to medical informatics. This book would provide insight to the learners about Machine Learning, Data Analytics, and Sustainable Computing.

Salient Features of the Book

• Exhaustive coverage of Data Analysis using R

• 
  
  

• Real-life healthcare models for:

• 
  
  

• Visually Impaired

• 
  
  

• Disease Diagnosisand Treatment options

• 
  
  

• Applications of Big Dataand Deep Learning in Healthcare

• 
  
  

• Drug Discovery

• 
  
  

• Complete guide to learn the knowledge discovery process, build versatile real life healthcare applications

• 
  
  

• Compareand analyze recent healthcare technologies and trends

• 
  
  

Target Audience

This book is mainly targeted at researchers, undergraduate, postgraduate students, academicians, and scholars working in the area of data science and its application to health sciences. Also, the book is beneficial for engineers who are engaged in developing actual healthcare solutions.

1

Big Data Analytics in Healthcare

Priti Bhardwaj

Indira Gandhi Delhi Technical University for Women

Centre for Development of Advanced Computing

Niyati Baliyan

Indira Gandhi Delhi Technical University for Women

CONTENTS

1.1Introduction: Background and Driving Forces

1.2Related Work

1.3Observations

1.4Open Challenges

1.5Proposed Solutions

1.6Conclusion

References

1.1 Introduction: Background and Driving Forces

Today’s world experiences big data from every domain. Thus, big data means lots of data or huge amount of data. When we have business problems, big data does not solve technical problems or give us a technical edge; rather, it solves large and complex business problems that cannot be solved with conventional approaches. So, a note can be taken as follows:

Big data: a problem statement arising out of a business situation

As shown in Figure 1.1, various characteristics of big data may be defined as follows:

• Volume: Data is huge and cannot be stored in a single server.
• Velocity: This indicates the speed at which data enters a system.
• Variety: Data captured is not in a single format. It can be log files, images, audios, or sensors.
• Veracity: Veracity indicates truthfulness. Data comes from multiple sources; so before going into the final picture, it should be free from ambiguities.

Healthcare organizations can experience wonderful benefits through big data analytics (BDA). Existing study has not provided enough awareness to the discovery of big business worth of big data. To enlighten the development of BDA abilities and to obtain profits from these abilities in a healthcare firm, a model has been introduced by Wang and Hajli (2017) using source-based theory and ability structure vision. Sixty-three healthcare bodies were explored to find the underlying associations linking BDA abilities and business value. New visions have to be reflected in healthcare on how to comprise BDA abilities for business revolution and present a practical basis that can motivate a more detailed exploration of BDA execution.

FIGURE 1.1

Characteristics of big data.

One of the most common neurodegenerative anarchies is Parkinson’s disease (PD). To enhance the accomplishment of PD discovery, a source blending procedure containing a combination of gray (GM) and white matter (WM) tissue maps and a decision fusion technique merging all classifiers’ output through correlation-based feature selection (CFS) method by mass selection were used (Cigdem & Demirel, 2018).

For all five classification algorithms, CFS gave the maximum outcome for all five diverse feature selection methods, while Support Vector Machine (SVM) provides the best classification performance, where 95% accuracy has been achieved with this fusion technique of GM and WM.

A significant research topic in healthcare is determining an individual’s risk of developing a certain disorder. An important step in personalized healthcare is precisely recognizing the level of likeness among patients based on their past reports. We need an efficient way to compute individual similarity on electronic health record (EHR) because EHRs are unevenly distributed and have various individual visit lengths and cannot be used to evaluate patient similarity because of inappropriate representation. In Suo et al. (2018), two new deep parallel learning schemes concurrently discover patient exhibitions and compute match up similarity. Convolutional neural network (CNN) has been used to detain neighboring vital statistics in EHRs. Disease predictions and patient clustering were also carried out in experiments using the similarity information. The outcome was that CNN can enhance representations of longitudinal EHR patterns.

There has to be an assurance of improved elasticity in industries along with group personalization, superior quality, and enhanced output such that organizations can deal with threats of progressively manufacturing more individualized products with a short lead time to market and advanced value. Intellectual production has to play an important role. We can convert resources into intellectual things so that they can respond spontaneously within a well-turned-out environment. For this, intellectual production, Internet of Things (IoT) and cloud manufacturing have been reviewed by Zhong et al. (2017). Similarities and dissimilarities have been discussed corresponding to these topics. Some core technologies such as IoT, cyberphysical systems, cloud computing, BDA, and information and communications technology have also been taken into consideration.

Research has shown that big data and prophetic analytics (BDPA) communicate a different governmental ability, and little is known about their performance effects, in particular contextual conditions (inter alia, national context and culture, and firm size). A trial has been done on 205 Indian manufacturing firms. The effects of BDPA have been investigated by Dubey et al. (2017) on public performance (PP) and ecological performance (EP) using variance-based structural equation modeling. It has been found that BDPA has an important bang on PP/EP.

The growth of information and communication technology is giving us a flow of new and computerized data associated with how nation and organizations work together. To convert this information into facts, considering the primary behavior of social and monetary mediums, organizations and researchers must tackle large amounts of unstructured and assorted data. To obtain success in task data analysis, the process should be carefully planned and organized considering the characteristics of economic and societal analysis and should include a vast multiplicity of heterogeneous sources of data and a strict supremacy strategy. A big data structural design has been constructed by Blazquez and Domenech (2018), which appropriately combines most of the nonconventional information resources and data analysis methods so as to offer a particularly planned structure for predicting public and profitable behaviors, trends, and changes.

As information technology is growing fast, people can shop online as well as provide feedbacks on social websites. The content that is generated by customers on social media is helpful to understand the shopping pattern of customers and to predict the future purchase. Today, customers are visiting the websites and see a lot of recommendations.

In Table 1.1, there are two customers: customer 1 and customer 2, and their buying patterns on a shopping website have been shown. Like Amazon has a lot of products and users, it needs to find the buying pattern of users and the relation between users. Customer 1 and customer 2 are matched, since their buying patterns are matched. So, the recommendation engine will give a choice of iPhone to customer 2.

TABLE 1.1

Buying Patterns of Customers

Customer 1	Customer 2
Bought items: 1. Titan watch	Bought items: 1. Mac laptop
2. Mac laptop	2. Titan watch
3. iPhone

It is difficult to analyze the customer-generated content on social media. Xu et al. (2017a) examined the characteristics of hotel products and services to observe consumer contentment and disappointment on the basis of online consumer textual feedbacks. Significant attributes motivating contentment and disappointment have been recognized using text mining approach. Latent semantic analysis (LSA) has been applied toward hotel products and service attributes. Business managers are helped to gain a better understanding of consumer’s requests through customer-generated content on websites. We are living in an electronic world, and fresh technologies or implementations are arising with an exponential growth of data. A three-aspect model is proposed by Stylianou and Talias (2017), taking the concern of big data in medical background and to increase the alertness of the considerations that a healthcare person may face in the upcoming time.

Increased higher load on healthcare and the subsequent sprain on funds give rise to the need for ways to increase treatment efficiency. The solution is to use a forecast model to recognize the risk factors for the entire duration of treatment. Execution of a new prediction tool and the first use of an active interpreter in professional therapy practice have been presented by Haraldsson et al. (2017). The tool is rationalized from time to time and enhanced with hereditary upgrading of software. The consequences have shown the consistency and accuracy of predictions for the entire treatment duration. The predictor classified the patients with 100% accuracy and correctness based on formerly hidden records after a learning period of 3 weeks. Professionals are using this predictor to make decisions in complex live systems.

To concentrate on all aspects of care delivery, a standard operating procedure through patient information analysis is medical brainpower. Medical intelligence techniques have been presented through data mining and procedure mining (Giacalone, Cusatelli, & Santarcangelo, 2018). It shows the comparison between these two approaches.

People fitness management (PFM) has been proposed by Wan (2018), and it pays attention on people having risk of chronic disorders who also have the utmost healthcare expenses. People health and PFM have been explained, along with problems in harmonized care. The performance and assessment of People Health Management (PHM) made use of EHRs and other data; and projected PHM practice and exploration.

Cloud computing, BDA, and other rapidly increasing technologies are growing, so the amalgamation of intellectuality and e-commerce systems is required to make a system with improved efficiency, minimized business costs, and groomed information. The expansion of these kinds of systems is still to be achieved (Song et al., 2017). To offer a better understanding of this smart system and assist for upcoming research, Song et al. (2017) explain this type of system in terms of cloud computing, IoT, mobile network, social media, and BDA.

Thus far, we know that big data application functions are the main part of healthcare operations. The study by Van den Broek and van Veenstra (2018) also provided a widespread, detailed, and organized reading of the advanced tools in the big data associated to healthcare uses in five types, including heuristic-based, cloud-based, machine learning, representative-based, and hybrid techniques.

Storing, processing, transporting, mining, and serving the data are big data challenges. The concept of shared computing is being provided by cloud computing. This study by Yang et al. (2017) surveyed the two frontlines—cloud computing and big data. It focuses on the benefits and outcomes of using cloud computing to...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Preface
• Editors
• Contributors
• 1. Big Data Analytics in Healthcare
• 2. Smart Medical Diagnosis
• 3. Lifestyle Application for Visually Impaired
• 4. Classification of Genetic Mutations
• 5. m-Health: Community-Based Android Application for Medical Services
• 6. Nanoemulsions: Status in Antimicrobial Therapy
• 7. Analysis of Air Quality and Impacts on Human Health
• 8. Brain Tumor Detection and Classification in MRI: Technique for Smart Healthcare Adaptation
• 9. Deep Strategies in Computer-Assisted Diagnosis and Classification of Abnormalities in Medical Images
• 10. Major Histocompatibility Complex Binding and Various Health Parameters Analysis
• 11. Partial Digest Problem
• 12. Deep Learning for Next-Generation Healthcare: A Survey of State-of-the-Art and Research Prospects
• 13. Applications of Protein Nanoparticles as Drug Delivery Vehicle
• 14. Exploring Food Domain Using Deep Neural Networks
• Index