Trends in data science

Abstract

This report provides a review of three challenges of Wearable smart devices in Healthcare, and how Big data solves them. (1)The diversity, variety and distribution of wearable data make data processing and analytics more difficult. (2)real-time remote healthcare monitoring data and patients prioritising need decision-making technique. (3)Existing solutions cannot collect long-term health data. Solutions are developing generic semantic Big Data Platforms, researching decision theory and analysing health monitoring data architecture.

The global population growth has increased the demand for technology, computer software algorithms and smart devices that can monitor and help patients anywhere, anytime, so that they can lead independent lives( Kalid, N., Zaidan, A.A., Zaidan, B.B. et al., 2018). By 2020, 40% of IoT-related technologies will be health-related, more than any other category and constituting a $117 billion market (Bauer H, Patel M, Veira J., 2016). New technologies have improved the capacity of home care providers because many chronic diseases that were previously treated in hospitals can now be safely managed at home. ( Kalid, N., Zaidan, A.A., Zaidan, B.B. et al., 2018). In the field of telemedicine, the real-time remote health monitoring system (RTRHMSs) is an essential component.

On the other hand, in the era of big data, with the prominence of data value and the progress of I (Information), data will promote the progress of future technology. Big data not only promotes the comprehensive integration of cloud computing, Internet of things, data center and mobile network but also promotes the cross-integration of multiple disciplines(g. Inf, 2015)

This article gives a review of the challenges of Wearable smart devices in Healthcare, and how Big data solves them.

This section lists three challenges of healthcare and states the opportunities for making use data to address the factors

l The diversity, variety and distribution of wearable data make data processing and analytics more difficult.

l Real-Time remote healthcare monitoring data and patients prioritising need decision-making technique.

l Existing solutions cannot collect long-term health data.

The diversity, variety, and distribution of wearable data make data processing and analysis more difficult. The popularity of wearable technology has created challenges associated with data management and integration. Existing wearable data applications are associated with well-defined data structures that are not scalable enough to support the new devices to integrate. Wearable data generated by multiple sources and arrives in a variety of formats due to the manufacturer and the algorithm used to encode the transmitted data. Embedding new wearable technologies into existing healthcare systems requires the development of appropriate monitoring applications, a challenge for the fast-growing market in the wearables. (Mezghani E. Exposito E., Drira, K. et al. 2015) Besides, the development of Semantic Web technology provides an opportunity to deal with the heterogeneity of semantic data, while semantic data heterogeneity hinders big data analysis when transforming decentralised medical data into valuable information. (Mezghani E. Exposito E., Drira, K. et al. 2015)

They proposed a generic semantic big data architecture that uses the Knowledge as a Service(KaaS) approach to handling data heterogeneity and system scalability challenges. They used the NIST big data reference architecture defined by the NIST working group. Generating heterogeneous data from multiple sources can be stored in a relational database, RDF, NoSQL, etc. or KaaS cluster through an implemented application. Transforming these heterogeneous data sets into understandable and shared knowledge requires the collection, preparation, and processing of these data sets. (Mezghani E. Exposito E. Drira, K. et al. 2015)

They introduce the application of wearable Cass in diabetes:

Suppose there are two wearable devices (W1 and W2) that monitor the same type of parameters (“glucose” class) through two different services S1 and S2, respectively. These services use different units (S1 for “mg/dL”, S2 for “mmol/L”) and different “glucose” parameter classes (S1 for “blood sugar” and S2 for “glucose”) measurement data. By implementing S1 and S2, they simulated the fake data of a virtual patient. The ID of this virtual patient is 09875645.

When KaaS retrieves data from a wearable device, the data is automatically normalised by reference to the “corresponding” attribute and stored for further reuse. Based on the returned wearable ID, the normalised data of the time is extracted from the cluster, and the patient’s blood glucose change is visually displayed. The test results show that two device’s data is integrated.

Real-time telemedicine monitoring data and patient prioritisation require decision-making techniques. According to Kalid, N., Zaidan, AA, Zaidan, BB et al. (2018), big data in real-time telemedicine monitoring includes health collected from individual individuals and large committees about their health at a certain point in time or another. Also, a large number of clinical, clinical, lifestyle, highly diverse, biological and environmental information. In remote monitoring, a variety of sensors can be used, such as ECG, SpO2, glucose sensors. These sensors may contain signals and image formats that increase the size of the data. The increase in the amount of data for the user or patient is related to the use of telemedicine systems within and outside the hospital.

Another challenge is that as the number of patients continues to increase, limited medical professionals should effectively use any developed system to meet growing demand. In the health care system, the ability to provide selective health services to all casualties immediately is insufficient. One way to solve this problem is to give patients priority treatment. (Kalid, N., Zaidan, A.A., Zaidan, B.B. et al. 2018)

One solution to solve these problems was paper triage. There are many problems and deficiencies in the paper triage method in the triaging accuracy and process of prioritisation. Electronic triage development as a problem-solving paper screening accuracy. However, using the START guide to classify remote patients into three categories is not the best solution (Kalid, N., Zaidan, A.A., Zaidan, B.B. et al. 2018).

MCDM is the most famous decision-making technique, and it is a branch of Operations Research (OR) that deals with decision-making issues related to decision-making standards (Malczewski, J., 1999, Petrovic-Lazarevic, S., & Abraham, A., 2004, cited by Kalid, N., Zaidan, A.A., Zaidan, B.B. et al. 2018). MCDM involves the use of multiple standards to construct, plan, and solve decision problems (Malczewski, J., 1999, cited by Kalid, N., Zaidan, A.A., Zaidan, B.B. et al. 2018). Its purpose is to help decision makers solve these problems [289]. Keeney and Raiffa (Keeney, R. L., & Raiffa, H., 1993, cited by Chen, m., Ma, y., Song, J. Et al., 2016) define MCDM as “an extension of decision theory that covers any decision with multiple goals.” A method of assessing individual choices, usually conflicting standards, and combining them into a single assessment…”

The prioritisation process considers multiple attributes (vital signs and complaints) simultaneously and scores the patient’s big data based on the most urgent cases. In the real world, useful methods of dealing with MCDM challenges are introduced as suggested solutions to help decision makers organise the issues to be solved, analyse, evaluate and rank (Kalid, N., Zaidan, AA, Zaidan, BB et al. 2018).

Existing solutions cannot collect long-term health data:

To achieve emotional care, the system needs to collect data on human emotions. Because emotional care requires more quantity and quality of body signals than traditional health care(Chen M, Gonzalez S, Leung V, Zhang Q, Li M 2010, cited by Chen, m., Ma, y., Song, J. Et al., 2016), and emotional interaction with users is also higher. Unfortunately, although there are many research results in the field of health monitoring, existing solutions have different shortcomings and cannot achieve long-term health monitoring. Traditional health monitoring methods have little effect on the sustainable health monitoring of chronic diseases.

Furthermore, many chronic diseases are caused by emotional and psychological problems. Furthermore, long-term unhealthy emotions may lead to the deterioration of chronic diseases. Therefore, the service level of the health monitoring system needs to be further improved. New solutions should be proposed for chronic disease and emotional care programs (Chen, M., Ma, Y., Song, J. et al. 2016).

Chen, M., Ma, Y., Song, J. et al. (2016) propose a sustainable health monitoring architecture based on smart clothing. It connects human and cloud in a quite natural way. Smart clothing can be defined as a new type of system that integrates various micro-sensors for physical signal acquisition. smart clothing can be applied to health monitoring, gaming, entertainment, military and other fields. The intelligence of smart clothing comes from sensor intelligence and cloud terminal intelligence. (Chen, M., Ma, Y., Song, J. et al. 2016) The system includes the following three design issue:

l Intra-SmartClothing System.

l Communications for Inter-SmartClothing.

l Beyond-SmartClothing (BSC) on Clouds.

Sustainable health monitoring has a close relationship with health big data, which can be observed from the following two aspects:

l From the data collection point of view: Without long term of physiological data collection supported by sustainable health monitoring, the data volume cannot reach the level of big data.

l From cloud intelligence point of view: Health big data analytics on clouds provide intelligence for more efficient health monitoring and make it more sustainable.

Big data is crucial for sustainable health monitoring (Zhang Y, Chen M, Mao S, Hu L, Leung V 2014, cited by Chen, m., Ma, y., Song, J. Et al., 2016). It can optimize public and private health systems. Big data on health can promote healthy lifestyles and activities, avoid the occurrence of chronic diseases (such as hypertension), slow down chronic diseases and transfer dependent patients to monitoring centers. At present, based on the BAN, and the application of A large amount of body area network business platform, can collect A large number of medical and health data (Chen M, 2014, Chen, m., Ma, Y reference) by big data technology research and the study of human activity recognition has become an important research direction of the BAN (Poon, CC, Lo BP, Yuce MR, Alomainy A, Hao Y 2015 cited by Chen, m., Ma, y., Song, J. Et al., 2016).

This article examines three health monitoring issues and the application of big data. It appears that developing a generic semantic Big Data Platform, research decision theory and developing health monitoring data architecture seems to be contemporary trends of big data in the field of health monitoring.

1. Chen, M., Ma, Y., Song, J. et al. ‘Smart Clothing: Connecting Human with Clouds and Big Data for Sustainable Health Monitoring’, Mobile Netw Appl (2016) 21: 825. <https://doi.org/10.1007/s11036-016-0745-1>

2. Healthc Inform Res. 2016 ‘Medical Internet of Things and Big Data in Healthcare’, viewed 13 April 2018, <https://synapse.koreamed.org/DOIx.php?id=10.4258/hir.2016.22.3.156>

3. Kalid, N., Zaidan, A.A., Zaidan, B.B. et al. ‘Based Real-Time Remote Health Monitoring Systems: A Review on Patients Prioritization and Related “Big Data” Using Body Sensors information and Communication Technology’, J Med Syst (2018) 42: 30. <https://doi.org/10.1007/s10916-017-0883-4>

4. Li, G. Inf, ‘Big Data Related Technologies, Challenges and Future Prospects’, Technol Tourism (2015) 15: 283. <https://doi.org/10.1007/s40558-015-0027-y>

5. Mezghani, E., Exposito, E., Drira, K. et al. 2015, ‘A Semantic Big Data Platform for Integrating Heterogeneous Wearable Data in Healthcare’, J Med Syst (2015) 39: 185. <https://doi.org/10.1007/s10916-015-0344-x>

6. Srivathsan Ma, Yogesh Arjun K, ‘Health Monitoring System by Prognotive Computing using Big Data Analytics’, Procedia Computer Science

Volume 50, 2015, Pages 602–609

7. Zhang Y, Chen M, Mao S, Hu L, Leung V (2014) ‘Community activity prediction based on big data analysis’. IEEE Netw 28(4):52–57

Related posts:

Entrepreneurship seems to be a scary word, because we usually imagine a daddy’s son, who with his large amount of money established a company that just occurred to him and is now richer. But, thanks…

I recently wanted to have shortcut to my favorite files — all the org files under my ~/org directory. Being not familiar enough with Emacs to write my own file chooser library, I Googled for help and…

Tight like the rubber band in her wrist. “Tenacious” is published by sheray johnson in El Sereno Community Garden.