Cluster Computing

https://doi.org/10.1007/s10586-018-2334-5 (0123456789().,-volV)(0123456789().,-volV)

Chatbot-based heathcare service with a knowledge base for cloud

computing
Kyungyong Chung1 • Roy C. Park2

Received: 31 October 2017 / Revised: 14 February 2018 / Accepted: 1 March 2018

Ó Springer Science+Business Media, LLC, part of Springer Nature 2018

Abstract
With the recent increase in the interest of individuals in health, lifecare, and disease, hospital medical services have been
shifting from a treatment focus to prevention and health management. The medical industry is creating additional services
for health- and life-promotion programs. This change represents a medical-service paradigm shift due to the prolonged life
expectancy, aging, lifestyle changes, and income increases, and consequently, the concept of the smart health service has
emerged as a major issue. Due to smart health, the existing health-promotion medical services that typically have been
operated by large hospitals have been developing into remote medical-treatment services where personal health records are
used in small hospitals; moreover, a further expansion has been occurring in the direction of u-Healthcare in which health
conditions are continuously monitored in the everyday lives of the users. However, as the amount of data is increasing and
the medical-data complexity is intensifying, the limitations of the previous approaches are increasingly problematic;
furthermore, since even the same disease can show different symptoms depending on the personal health conditions,
lifestyle, and genome information, universal healthcare is not effective for some patients, and it can even generate severe
side effects. Thus, research on the AI-based healthcare that is in the form of mining-based smart health, which is a
convergence technology of the 4IR, is actively being carried out. Particularly, the introduction of various smart medical
equipment for which healthcare big data and a running machine have been combined and the expansion of the distribution
of smartphone wearable devices have led to innovations such as personalized diagnostic and treatment services and
chronic-disease management and prevention services. In addition, various already launched applications allow users to
check their own health conditions and receive the corresponding feedback in real time. Based on these innovations, the
preparation of a way to determine a user’s current health conditions, and to respond properly through contextual feedback
in the case of unsound health conditions, is underway. However, since the previously made healthcare-related applications
need to be linked to a wearable device, and they provide medical feedback to users based solely on specific biometric data,
inaccurate information can be provided. In addition, the user interfaces of some healthcare applications are very com-
plicated, causing user inconvenience regarding the attainment of desired information. Therefore, we propose a chatbot-
based healthcare service with a knowledge base for cloud computing. The proposed method is a mobile health service in
the form of a chatbot for the provision of fast treatment in response to accidents that may occur in everyday life, and also in
response to changes of the conditions of patients with chronic diseases. A chatbot is an intelligent conversation platform
that interacts with users via a chatting interface, and since its use can be facilitated by linkages with the major social
network service messengers, general users can easily access and receive various health services. The proposed framework
enables a smooth human–robot interaction that supports the efficient implementation of the chatbot healthcare service. The
design of the framework comprises the following four levels: data level, information level, knowledge level, and service
level.

& Roy C. Park 2

Division of Computer Engineering, Dongseo University, 47,
[email protected] Jurye-ro, Sasang-gu, Busan 617-716, Korea
Kyungyong Chung
[email protected]
1
Division of Computer Science and Engineering, Kyonggi
University, 154-42, Gwanggyosan-ro, Yeongtong-gu,
Suwon-si, Gyeonggi-do 16227, Korea

123
 Cluster Computing

Keywords AI smart health
Chatbot framework
Chatbot health
Medical data mining
Cloud computing

1 Introduction through a chatting interface. Since it can connect with the

major social network service (SNS) messengers, general
With the evolvement of wired and wireless communication users are able to access the platform easily and receive a
technologies and the development of multimedia tech- variety of services. Other companies have also provided a
nologies for processing and storage, information and healthcare service using an AI-based chatbot. With a
communication technology (ICT) has gradually become chatbot, it is possible to identify a patient’s conditions,
more complicated and intelligent. With the support of suggest methods for the improvement of his/her conditions,
brain-science and medical brain-science-based technolo- and help in the formation of a connection with a proper
gies, the innovation of AI technology has been triggered. medical service through conversation. Particularly, as a
Until now, healthcare services have been provided for the chatbot is capable of reducing the need for medical staff
diagnostic and treatment processes, and consequently, and assisting patients regardless of time and space, the
research has been conducted on the convergence of chatbot technology has received a lot of attention [21].
advanced medical systems and ICT. In the circumstance Given the previously outlined circumstances, a chatbot-
where the medical data are more complicated, it has been based healthcare service with a knowledge base for cloud
very difficult to control the data using the conventional computing is proposed is this paper. The proposed method
approaches [1–4]. Even in cases where patients are diag- is a chatbot-based mobile healthcare service that can
nosed with the same disease, their symptoms can be dif- immediately respond to the accidents that arise in everyday
ferent depending on their health conditions, living habits, life and to the condition changes of chronic-disease
and dielectric information. Therefore, the average medici- patients. This paper also proposes a framework for the
nal regime is ineffective for some patients, or it may cause human–robot interaction that can ensure an efficient
severe side effects; furthermore, it is difficult to select the implementation of the chatbot service. The design of the
useful data from massive and diverse medical big-data proposed framework consists of the data layer, the infor-
sources during the course of treatment [5–8]. In particular, mation layer, the knowledge layer, and the service layer.
chronic diseases, including diabetes, hypertension, and Based on a variety of collected knowledge, big data are
obesity, require the provision of effective care rather than established and analyzed. For user convenience, a four-
treatment from the time they first occur. For this reason, a layer-based service is offered, and the service of the service
patient’s lifelog is of great significance. To provide cus- layer is offered through a customized UI/UX [6, 22]. In
tomized services for patients on the basis of massive col- addition, for the normal operation of mobile applications,
lections of personal medical data, an AI algorithm is an interaction with the service layer of the chatbot health
essential; therefore, domestic hospitals have established AI framework is required. Accordingly, with the component-
systems that are capable of predicting emergency situa- based design principle, this study applied a service-orien-
tions. Such systems use their AI to analyze a patient’s tation structure to the chatbot mobile-service architecture
biometric information, thereby allowing for the provision for the provision of a mobile service. Regarding the
of emergency-situation notifications to either patients or mobile-service access interface, the clients need to use the
their caregivers and to provide medical staff with an service interface to access all of the application features
awareness of emergencies whereby they can prepare and functions. For this reason, the proposed structure was
accordingly [9–13]. designed to include new components for the improvement
Google and Microsoft have released chatbot health- of the code reuse and maintenance, which are the important
platform services that are usable in everyday life. Baidu requirements of distributed systems. In the proposed plat-
developed the healthcare chatbot called ‘‘Melody,’’ which form, since a user is able to offer his/her everyday-life
communicates with a patient to deliver essential informa- health data in real time, it is possible to efficiently diagnose
tion to medical staff and checks the medical-staff schedule chronic diseases and to provide preventive information
to make doctor appointments [14–16]. A chatbot is a nat- according to observations of the patients’ everyday lives.
ural interactive simulation system with AI-based conver- The rest of this paper is organized as follows. Section 2
sational modeling and knowledge. As the interactive describes the related research of AI-based healthcare and
simulation system is more conversationally capable, it previous healthcare chatbots. Section 3 presents the chat-
learns more and properly responds to situations, making it a bot-based healthcare service with a knowledge base for
sort of advanced AI system [17–20]. A chatbot is an cloud computing. Section 4 describes the implementation,
intelligent interactive platform that can interact with users and Sect. 5 presents the conclusion.

123
Cluster Computing

Fig. 1 UNC (University of North Carolina) Health Care [24]

2 Related research corresponding startups are presently in the growth stage.

Research on the application of AI in the medical field is
2.1 AI-based healthcare actively being conducted. Lunit, for example, is developing
a diagnostic system for lung diseases and breast cancer
A smart medical device that is developed with a combi- using AI image-recognition technology where chest and
nation of big data and machine learning is able to provide breast X-ray photos are utilized. The system, which is
more effective and personalized solutions; therefore, it is based on a deep-learning algorithm, is able to check chest
expected to facilitate innovations in the medical area X-rays and mammographic images; make a diagnosis; and
regarding disease prevention, diagnosis, and treatment. detect the position, size, transformed cells, and specific
Google has presented projections of the use of an AI ser- tissues of a tumor, which are very difficult to observe using
vice for usual health conditions and the corresponding the naked eye [25]. Also, Vuno developed a deep-learning
proper responses through its development of healthcare algorithm that is applicable in the medical field. For its
technologies including a blood-glucose monitor, an anti- algorithm, AI technology is applied in the analysis of
aging treatment agent, and genetic analysis. Through its medical images (X-ray, CT, MRI) and bio-signals. The
DeepMind health AI technology,1 which also developed developed algorithm is used as an aid in the diagnosis of
AlphaGo, Google developed Streams, an AI program for lung diseases, cardiovascular diseases, cerebral aneurysms,
the real-time collection of the information of patients with and the determination of bone age. The company also
severely damaged organs, enabling medical staff to com- developed Vuno-Med, a software application that assists in
plete a rapid diagnosis [23]. the diagnosis of lung cancer. Recently, Apple developed
University of North Carolina (UNC) Health Care2 is a CareKit, a software application for the sharing of treatment
not-for-profit medical institution that provides high-quality plans between patients, caregivers, and medical staff,
medical services for more than 37,000 patients with its whereby medicine regimes can be monitored and patient
cutting-edge facilities. To secure insights that are based on treatments can be delivered more effectively [26]. The
data analysis, UNC Health Care introduced the IBM text- company’s Apple Watch or iPhone technologies provide a
analysis service that is capable of analyzing both structured sensor-based system that allows for the measurement of the
and unstructured data; as a result, it can identify the causes patient health conditions and the visualization of the
of patient hospitalization, thereby reducing rehospitaliza- treatment information.
tion costs [24]. Figure 1 shows a flow of the University of
North Carolina (UNC) Healthcare service. 2.2 Previous healthcare chatbots
Domestically, AI-based healthcare research and devel-
opment (R&D) draws a lot of attention, and therefore, the HealthTap3 is a startup that developed a platform for the
verification of medical information for which medical staff
1
DeepMind Health, https://deepmind.com/applied/deepmind-health/. and patients are connected. More than 100,000 medical top
2
University of North Carolina (UNC) Health Care, http://www.unc.
3
edu/health/. HealthTap, https://www.healthtap.com/.

123
 Cluster Computing

Fig. 2 HealthTap service with a

question and the answers [27]

doctors have registered on the HealthTap platform, and form to assist the medical staff in the formulation of a
more than 6.4 billion medical answers have been offered to potentially more-accurate diagnosis compared with the
patients. HealthTap launched a Facebook-Messenger-based diagnosis that was made during the patients’ hospital visits
healthcare chatbot [27]. Figure 2 shows the HealthTap [29]. AI chatbots that ask patients questions about disease
service with an example of a question and the corre- symptoms are increasingly being tested due to the wide
sponding answers. HealthTap presents answers on the basis usage prevalence. Approximately 1.2 million residents in
of a database that contains similar questions. When a Islington, Camden, Enfield, and Barnet, UK, prefer using
customer asks a question, its chatbot seeks out similar cases the chatbot of Baylon to contacting the National Health
and shows medical-staff answers to the question. Service (NHS). When users enter their symptoms into the
As a startup that is providing a new type of health Baylon app, the given responses are dependent on the input
insurance, the Oscar chatbot4 presently receives a large content. When the users answer the questions of the app,
amount of attention. Oscar quickly offers a health-insur- the provided suggestions regarding their symptoms are in
ance estimate to customers through its Facebook-Messen- reference to a large disease database [30].
ger-based chatbot. Although the Oscar chatbot now
provides a limited service, it is expected that Oscar will
provide a casual healthcare service by utilizing sensitive 3 Chatbot-based heathcare service
user health and disease information to calculate health- with a knowledge base for cloud
insurance estimates [28]. The Molly chatbot service of computing
Sensely,5 which is a virtual-nurse platform that was
developed to save the time of medical staff by 20%. Using 3.1 Chatbot-based healthcare service framework
a patient’s real-time health data, Molly judges a patient’s
health conditions and monitors the patient until his/her next The proposed chatbot-based healthcare-service framework
hospital visit. is composed as follows: data layer, knowledge layer,
The interactive question based medical-diagnosis chat- information layer, and service layer. To establish and
bot of Wisenut was developed with the use of the com- analyze big data on the basis of a variety of collected
pany’s multiyear accumulated medical bigdata analysis knowledge and to provide data for users conveniently, a
technology, machine learning based intelligent technology, service is offered in each of these four layers. In the service
hybrid sentence comprehension technology, and plan based layer, a customized UI/UX offers the service [1, 5, 31].
active conversation management technology. This chatbot Figure 3 shows the chatbot-based healthcare-service
asks daily questions about the health conditions of chronic- framework. The data layer comprises data processing and
disease patients, enabling it to offer suggestions on the storage components that are for the processing and storage
basis of the provided answers. The answers of chronic- of the data that are obtained from multiple sensors
disease patients are also provided to medical staff in a brief including a wireless body area network (WBAN) sensor. In
real time, the data-processing module performs data col-
4 lection, data labeling and analysis, and data filtering
OSCAR, https://www.hioscar.com/ny/.
5 depending on the type of data collection. In the case of
Molly Chatbot Service, http://www.sensely.com/.

123
Cluster Computing

Fig. 3 Chatbot-based healthcare-service Framework

various types of collected data and a large data amount, in the lifelog model. In the service layer, the personalized-
however, the data can be damaged by duplication and recommendation information for the final prediction and
noise. Therefore, in the data layer, the labeled data that are recommendation is created.
collected for the real-time filtering are processed first, and As previously indicated, the information layer comprises
this is followed by the storage of the data. Subsequently, both the high-level and low-level context-recognition
the use of multiple analytic filters enables the increasing of models. In this layer, the user’s behavior and context are
the consistency and reliability of the data classification and inferred and modeled. The low-level context recognition
analysis. Besides the real-time data filtering, the batch-fil- recognizes and classifies the user lifestyle including phys-
tering process is performed depending on the amount of the ical activities, emotional status, and position, as well as
collected data and their features. The collected data are not their Internet- and SNS-usage patterns for the time zone
deleted to establish big data. In this process, the security where the medical information is frequently collected.
component in the support layer performs encryption and Based on the classified data, the high-level context recog-
user authentication to enhance the protection of personal nition is able to predict the user contextual and behavior
information [2]. For the ease of the collected-data usage, patterns through convergence and inference. In this case,
lifelog ontology is applied; that is, the data are expressed at the low-level context recognition consists of components
a high level through data-based service provisioning. This such as behavioral recognition, emotional recognition,
information is then saved in the database of a sharing position detection, and SNS-usage analysis. The SNS-
server to smoothly share the data between applications and usage analysis component collects and analyzes multiple
different layers. In this case, the collected informal data is data regarding the user’s medical-information searches and
converted to semistructured data for the lifelog-ontology visits the user’s SNS-usage records. Therefore, it is possi-
mapping. Based on the low-level and high-level context- ble to discern the user’s medical information including the
recognition models that are created in the information layer range and the status and their medical-behavior require-
and the lifelog ontology of the data layer, the collected user ments in a particular situation. At this time, the user’s
data are saved in the form of structured data, and then they physical activities can be recognized by the recognition
will be used in both the service and knowledge layers. In module. The user’s activity and condition can be recog-
the data layer, the personalized prediction performs nized using the data that are collected from the WBAN
behavioral predictions as a detailed module depending on sensor and the multimedia module. The emotion-recogni-
the user’s lifestyle, and the prediction is finally expressed tion module is also able to collect various types of data that

123
 Cluster Computing

are collected from a camera-based user face analysis and A different service is properly called upon depending on
the WBAN sensor whereby the user’s emotional status can the given situation. Therefore, chatbot-type services com-
be drawn. To determine the user conditions accurately, it is prise the following two steps: customized medical-infor-
important to be aware of the user position and movement mation service and personal health service. In the first step,
path, and this means that Global Positioning System (GPS) based on the user data that are offered in the data layer and
information should be used for the purpose of detection. the medical knowledge that is filtered in the knowledge
Among the four components of the low-level context layer, the user lifestyle information and lifelog are ana-
recognition, the Input Adapter component classifies and lyzed for the creation of the user recommendation infor-
selects the priority data and loads the data that are curated mation. In the second step, the medical recommendation
from the data layer. The filtered information in the low- information that is created in the first step is processed on
level context recognition is transmitted to the data layer the basis of the user health conditions, environments, and
through the Output Adapter. The data that are classified in requirements. Personalized recommendation information
the low-level context recognition are used for an accurate includes information that draws the user’s attention, and it
and detailed context-recognition inference regarding the is purposely created so that it can be easily understood by
user that is achieved through the high-level context the user. The recommendation information is created by
recognition. The component of the high-level context the Prediction Manager component. Through the analysis
recognition redesigns and analyzes the user situation that is of the user lifestyle, the situation that is most likely to
recognized by the context-recognition modeling compo- occur is predicted in advance, thereby creating the rec-
nent, and then ontology is used to express the user’s situ- ommendation information and providing a relevant service
ation. In context-recognition modeling, the user’s lifelog is for the users. The last layer is the support layer. In the
saved and all of the user contextual information is saved. support layer, the processes that are performed in the three
To find the user behavior pattern and movement in the other layers, such as the user data collection and analysis,
high-level context, the user contextual modeling and recommendation-information offering, feedback analysis,
analysis are applied. and the generation of security factors for personal-infor-
In the knowledge layer, the user health information that mation protection, are supported for an overall support of
is collected in the data layer is analyzed, and then knowl- the framework. Figure 4 presents the structure of the sup-
edge is created and managed to provide customized port layer with the knowledge base. The analysis module
information through the already established medical-in- finally analyzes the user health conditions and patterns on
formation database; therefore, the expert-based and data- the basis of the collected, processed, and curated data of
based approaches are applied [32]. Through the creation each layer. The collected data from the users includes
component, the expert-based approach creates the rules for sensitive personal information, so an encryption and
the provision of the medical information. For reliability, a security protocol is applied to all of the data. For storage
comparative analysis with the database is conducted to security, the Advanced Encryption Standard (AES) is
remove errors such as the provision of inaccurate infor- applied. A private matching technique and a homomorphic
mation. The data-based approach makes use of the user encryption technique are performed for oblivious
lifelog information that is curated from the data and processing.
information layers. With the use of various data-mining
models in the intermediate database, the data-broker 3.2 Components of chatbot architecture
interface automatically collects the essential information for cloud computing
that is related to the factors and is set by the experts in the
expert-based approach. The created rules are applied in the Chatbot-based mobile personal-life healthcare provides the
authentication process to achieve an increased reliability, user medical patterns that are collected by the chatbot
duplication prevention, and rule consistency; subsequently, framework from a messamge. For a user with a mild dis-
they are offered to the user. ease or mild physical damage in their everyday life who
In the service layer, the medical-knowledge information makes a request, the user’s everyday-life information that
that is newly created and collected in the data, information, has been collected by the chatbot health framework is
and knowledge layers is converted for the user healthcare analyzed, and a proper service is provided in a typed
service. The service layer plays a role in the management message. Through the application programming interfaces
of the user service processes such as the response time to (APIs) of medical institutions, the chatbot-based mobile
the service request for the service loading. Such processes healthcare system helps in the uploading and analysis of
are performed by the service orchestrator. As a service images and data regarding a variety of diseases and trauma
request is necessary in this process, an event request, a that are held at hospitals to compare the pictures of patient
regular time request, and a user direct request are required. conditions and to find the corresponding diseases; the

123
Cluster Computing

Fig. 4 Structure of the support layer with a knowledge base

relevant information can then be provided. In addition, by medical-information systems. Figure 5 shows the service
obtaining and processing the biometric data that are environment to provide chatbot-based health.
occurring in real time in the everyday lives of chronic- The chatbot-based mobile healthcare system manages
disease patients, this system provides an additional service the personal everyday-life area of patients; that is, the user
for the storage of personal-life health records in hospital obtains his/her health information that has been collected
from a WBAN in the mobile environment via the Bluetooth

Fig. 5 Service environment to provide chatbot-based health

123
 Cluster Computing

Fig. 6 Component of chatbot architecture for cloud computing

technology, and then he/she can make additions and time stamp for the sequential processing of each input data,
changes in the user app. Subsequently, the information is and this is followed by the processing of the waiting
transmitted to a hospital information system for the queued data in their input sequence [22]. Figure 6 shows
establishment of a personal health-information record, the structure of chatbot-architecture components for cloud
thereby allowing the information to be offered to the user computing.
in a typed message. To run a mobile application normally, The data-standby component guarantees real-time data
it is necessary to interact with the service layer of the processing; that is, it offers a queuing service to resolve the
chatbot health framework, and this is why the service- data loss and error that are caused by service delays,
oriented structure of the chatbot mobile-service architec- thereby ensuring an asynchronous data process. So when a
ture follows the component-based design principle. That is, user uses the chatbot service, the management of the data
the system was designed to ensure the maintenance of the queue is necessary to solve the overall problems that are
following factors that maximize the service or component caused by different service speeds, thereby providing a
reusability and the scalability: abstraction, loose coupling, queuing service. In addition, when a massive data input
autonomy, and statelessness [33]. leads to the exceeding of a service process capacity, rather
The service layer of the chatbot healthcare service than the occurrence of an external environmental problem
receives the personal-condition data that are sent by the in a network, it is possible to guarantee the data processing
WBAN sensor in real-time, and these data are then trans- through a delay process. The data-conversion component
mitted to the server of the chatbot health framework that guarantees a compliance with the medical record and
sends information to an integrated system. The data-ac- transmission standard HL7 [34] in the service layer. The
quisition component interacts with the data-processing personal health data that are collected in real time are
module of the chatbot framework to filter the collected data transmitted as the IEEE 11073 PHD data type; as a result, it
and obtains only the data that needs to be observed by the is necessary to follow the HL7 standard [34, 35]. For this
user and the medical staff. The reason for this requisite reason, a server module is used to observe the security
filtering function is its use in the solving of the overload guide of the received data standard and to protect the
problem that occurs, where unnecessary information is medical information. The data-encryption component pro-
collected, and for the provision of accurate and simple vides an encryption service. So, after the data is received
information in accordance with the principle of exception stably, the component accumulates the personal health
management. As for the data that are collected by the fil- information from the personal-life health record into a
tering function, its process is guaranteed by the data- database using a DBMS. The data-representation compo-
identification component. In this case, to prepare the nent guarantees a user-friendly output in the service layer.
queuing service, a time stamp is attached to assign a pri- Based on the already analyzed and encrypted information,
ority to the input data and to identify the real-time data. In the chronic-disease monitoring and feedback services are
brief, the key to the queuing service is the addition of a offered to the user through the mobile-application UI [22].

123
Cluster Computing

Fig. 7 Sequence diagram of

chatbot healthcare platform
between mobile clients

Accordingly, the user is able to receive a variety of cus- In the health-service component, the user-measured data
tomized services by searching for his/her own health are saved in a database. When a user checks his/her health
information directly. Figure 7 shows a sequence diagram history including the measured data and makes a data
of the chatbot healthcare platform between mobile clients. request through a variety of devices, the relay model
The data-sharing process of the healthcare platform that accesses the database to retrieve the data, converts the
occurs through the mobile-phone app is presented as fol- measured data into a proper data format, and the data are
lows: A personal-health data manager of the framework finally parsed to meet the request. As for the data format, a
sends a request asking for the transmission of personal measurement data, a measured item, measured data, and
health data to a mobile device, receives the personal health other essential data are extracted and then filtered to pro-
data in response, and then transmits the data to a mapping vide the measured data to a user efficiently. In the case of
manager. The mapping manager converts the data into a the mobile-service access interface, to enable browser and
health-information message and then sends it to a trans- device-client access to all of the application functions and
mission manager. A platform manager requests a platform features, it is necessary to use the same service interface
to perform an authentication. The platform collects the [36]. For an immediate treatment and transfer of an
user-requested data and then finally transmits it to the emergency patient, the chatbot-based healthcare service
mobile device. delivers the information and conditions of the patient and
suggestions regarding the conditions through the chatting
function, and it also communicates with the chatbots and
4 Implementation relevant organizations. The user sends his/her personal
information, including name, gender, and resident regis-
For the mobile service that has been designed for this tration number, and the image information of the patient
study, an application was developed using computer H/W condition using the messaging function. The sent data are
with the Intel(R) Core(TM) i7-4770 central processing unit compared with the medical data of the chatbot, and then the
(CPU) 3.40 GHz and a 16.00-GB RAM, and for S/W, the relevant emergency information is sent back to the user.
64-bit Windows 8.1 Enterprise K and the Android Studio Figure 9 shows the initial screen of the developing chatbot
5.1.1 (KITKAT) were used. Further, as a relay model is health service. When an emergency occurs or when a
required for the chat-based healthcare service, this study patient or their caregiver submits a report, the report is
made use of the simple object access protocol (SOAP)- accepted and the chatbot service is provided concurrently.
based Web service that is mainly applied in health-appli- The chatbot sends the reporter a message with the infor-
cation services to create a relay model. Figure 8 shows the mation of the acceptance of the report as well as other
cloud-computing environments for mobile healthcare. messages.

123
 Cluster Computing

Fig. 8 Cloud-computing environments for mobile healthcare

Fig. 9 Initial screen of developing chatbot health service

As shown in Fig. 10, a patient takes an image of their In addition, all of the data that are related to the user’s
injury or disease region and sends the information to a health profile and a health-portal service for statistics and
chatbot. The chatbot receives the image and compares it other services are offered. Figure 11 shows the life-health
with its own health images or a text-based medical status-confirm service. The measured values and informa-
knowledge base to make a decision regarding the patient’s tion of the most recently collected physical activities of the
condition. Once the determination is complete, information user are presented, and these data items can be easily
on the proper emergency treatment is sent back; with this monitored with graphs and tables. Based on the provided
process, first aid can be implemented before an ambulance information, the user is able to set up and change his/her
officer arrives. Further, the process that is shown in Fig. 10 treatment and prevention plans within his/her controllable
is also performed. competency. The collected and analyzed information is
used to provide the chronic-disease monitoring and

123
Cluster Computing

Fig. 10 Result of sending emergency-patient information

transmitted to medical services like the personal health

record (PHR) and the health information system (HIS),
thereby facilitating the offering of consistent information.
In most of these studies, however, the architecture plays the
simple role of data transmission. The system that is pro-
posed in this study is able to perform data filtering and
processing through the previously defined chatbot health
framework to transmit a variety of information to the user
and the HIS, and to also provide an accurate medical ser-
vice. Unlike the existing mobile health systems, the chat-
bot-based mobile health system provides not only the given
information, but it also interactively provides the infor-
mation that is necessary for the user. In addition, the user-
life medical information that is used in the chatbot con-
versation is saved in the PHR. Through an interactive
process with the HIS, it is possible for patients to receive a
more accurate diagnosis, an improved treatment, and
Fig. 11 Life-health status-confirm service superior preventive information. As a result of the ability to
manage their health conditions in real time, chronic-disease
feedback services to the user through Web and mobile patients expect that their behaviors will change and their
applications. Therefore, the user is able to search for his/ QOL will be improved. Due to the interactions between the
her own health information and receive various customized proposed system and various types of smartphone health
services. application and data, it is now possible to expand medical-
service applications. Due to the extension of the value
chain of the management of the diagnostic, treatment, and
5 Conclusions follow-up processes, which are typically applied only in
medical institutions, it is now expected that the entire
According to an analysis of the studies on recent mobile consumer life cycle will be medically managed.
health services, a number of studies considered the archi-
tecture in which physical body data, which are sent by a Acknowledgements This research was supported by Basic Science
WBAN sensor with the use of a mobile device, are Research Program through the National Research Foundation of

123
 Cluster Computing

Korea (NRF) funded by the Ministry of Education IoT-enabled Healthcare and Wellness Technologies and Systems,
(2016R1C1B1014849). pp. 23–28 (2016)
19. Oh, K.J., Lee, D.K., Ko, B.S., Choi, H.J.: A chatbot for psychi-
atric counseling in mental healthcare service based on emotional
dialogue analysis and sentence generation. In: Proceedings of the
References 18th IEEE International Conference on Mobile Data Manage-
ment, pp. 371–375 (2017)
1. Chung, K., Park, R.C.: PHR open platform based smart health 20. Yoo, H., Chung, K.: Heart rate variability based stress index
service using distributed object group framework. Clust. Comput. service model using bio-sensor. Clust. Comput. (2017). https://
19(1), 505–517 (2016) doi.org/10.1007/s10586-017-0879-3
2. Chung, K., Kim, J.C., Park, R.C.: Knowledge-based health ser- 21. Fryer, L.K., Ainley, M., Thompson, A., Gibson, A., Sherlock, Z.:
vice considering user convenience using hybrid Wi-Fi P2P. Inf. Stimulating and sustaining interest in a language course: an
Technol. Manag. 17(1), 67–80 (2016) experimental comparison of chatbot and human task partners.
3. Jung, H., Yoo, H., Chung, K.: Associative context mining for Comput. Hum. Behav. 75, 461–468 (2017)
ontology-driven hidden knowledge discovery. Clust. Comput. 22. Park, R.C., Jung, H., Chung, K., Yoon, K.H.: Picocell based
19(4), 2261–2271 (2016) telemedicine health service for human UX/UI. Multimed. Tools
4. Yoo, H., Chung, K.: PHR based diabetes index service model Appl. 74(7), 2519–2534 (2015)
using life behavior analysis. Wirel. Pers. Commun. 93(1), 23. DeepMind Health. https://deepmind.com/applied/deepmind-
161–174 (2017) health/
5. Chung, K., Park, R.C.: Cloud based U-healthcare network with 24. University of North Carolina Health Care. http://www.unc.edu/
QoS guarantee for mobile health service. Clust. Comput. (2017). health/
https://doi.org/10.1007/s10586-017-1120-0 25. Lunit Healthcare. https://lunit.io/
6. Park, R.C., Jung, H., Shin, D.K., Kim, G.J., Yoon, K.H.: M2 M- 26. Vuno Healthcare. https://www.vuno.co/
based smart health service for human UI/UX. Clust. Comput. 27. HealthTap. https://www.healthtap.com/
18(1), 221–232 (2015) 28. OSCAR. https://www.hioscar.com/ny/
7. Park, R.C., Jung, H., Jo, S.M.: ABS scheduling technique for 29. Wisenut. http://www.wisenut.com/
interference mitigation of M2 M based medical WBAN service. 30. UK National Health Service (NHS). https://www.nhs.uk/
Wirel. Pers. Commun. 79(4), 2685–2700 (2014) 31. Song, C.W., Jung, H., Chung, K.: Development of a medical big-
8. Kang, S.K., Chung, K.Y., Ryu, J.K., Rim, K.W., Lee, J.H.: Bio- data mining process using topic modeling. Clust. Comput. (2017).
interactive healthcare service system using lifelog based context https://doi.org/10.1007/s10586-017-0942-0
computing. Wirel. Pers. Commun. 73(2), 341–351 (2013) 32. Jung, H., Chung, K.: Sequential pattern profiling based bio-de-
9. Kim, J.C., Chung, K.: Emerging risk forecast system using tection for smart health service. Clust. Comput. 18(1), 209–219
associative index mining analysis. Clust. Comput. 20(1), 547–558 (2015)
(2017) 33. Yoo, H., Chung, K.: Mining-based lifecare recommendation
10. Jung, H., Chung, K.: PHR based life health index mobile service using peer-to-peer dataset and adaptive decision feedback. Peer-
using decision support model. Wirel. Pers. Commun. 86(1), to-Peer Netw. Appl. (2017). https://doi.org/10.1007/s12083-017-
315–332 (2016) 0620-2
11. Jung, H., Chung, K.: Knowledge-based dietary nutrition recom- 34. HL7. Health Level Seven International. http://www.hl7.org/
mendation for obese management. Inf. Technol. Manag. 17(1), 35. Kim, J.C., Chung, K.: Mining health-risk factors using PHR
29–42 (2016) similarity in a hybrid P2P network. Peer-to-Peer Netw. Appl.
12. Kim, J.C., Yoo, H., Kang, M.A., Ghose, S., Chung, K.: Big (2018). https://doi.org/10.1007/s12083-018-0631-7
lifelog-based management using similarity for health platform. 36. Liu, K.H., Chen, T.Y., Chen, C.S.: MVC: a dataset for view-
In: Proceedings of the 7th International Conference on Conver- invariant clothing retrieval and attribute prediction. In Proceed-
gence Technology, pp. 627–628 (2017) ings of the 2016 ACM on International Conference on Multi-
13. Sheth, A., Jaimini, U., Thiruarayn, K., Banerjee, T.: Augmented media Retrieval, pp. 313–316 (2016)
personalized health: how smart data with IoTs and AI is about to
change healthcare. In: Proceedings of the 2017 IEEE Interna-
tional Forum on Research and Technologies for Society and Kyungyong Chung has received
Industry (RTSI) (2017) B.S., M.S., and Ph.D. degrees in
14. Schaeffer, D.M., Olson, P.C.: Information and communications 2000, 2002, and 2005, respec-
technology in Africa: enabling big data, to enable development. tively, all from the Department
J. Comput. Sci. Colleges 33(3), 71 (2018) of Computer Information Engi-
15. Kar, D.M., Lazring, I., Ray, I.: POSTER: PriReMat: a distributed neering, Inha University, Korea.
tool for privacy preserving record linking in healthcare. In: Pro- He has worked for Software
ceedings of the ACM SIGSAC Conference on Computer and Technology Leading Depart-
Communications Security, pp. 207–2509 (2017) ment, Korea IT Industry Pro-
16. Kim, J.C., Chung, K.: Exploring design opportunities for a con- motion Agency (KIPA). From
text-adaptive medical checklist through technology probe 2006 to 2016, he was a profes-
approach. In: Proceedings of the 2017 Conference on Designing sor in the School of Computer
Interactive Systems, pp. 57–68 (2017) Information Engineering, Sangji
17. Kim, J.C., Chung, K.: Depression index service using knowledge University, Korea. Since 2017,
based crowdsourcing in smart health. Wirel. Pers. Commun. he is currently a professor in the
93(1), 255–268 (2017) Division of Computer Science and Engineering, Kyonggi University,
18. Fischer, M., Lam, M.: From books to bots: using medical liter- Korea. His research interests include Data Mining, Artificial Intelli-
ature to create a chat bot. In: Proceedings of the Workshop on gent, Healthcare, Knowledge System, HCI, and Recommendation
System.

123
Cluster Computing

Roy C. Park received the B.S.

degrees from Dept. of Industry
Engineering, and M.S., Ph.D.
degrees from Dept. of Computer
Information Engineering, Sangji
University, Korea, in 2010 and
2015. He has worked for Tech-
nical Research Center IBS
Management Team, Cosmotier
Co., Ltd. He is currently a pro-
fessor in the Division of Com-
puting Engineering, Dongseo
University, Korea. His research
interests include WLAN Sys-
tem, Heterogeneous Network,
Ubiquitous Network Service, Health Informatics, Knowledge System,
Peer-to-Peer, and Cloud Network.

123