Natural Language Processing and summarization of medical symptomatic data from geographical diverse locations

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 08 | Aug 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1878
Natural Language Processing and summarization of medical
symptomatic data from geographical diverse locations
Snehal N. Palve1, R.N Awale 2, Vaibhav Awandekar 3, Sunil Lakdawala 4
1 M.Tech Student, Electrical Department, Veermata Jijabai Technological Institute, Mumbai, India
2 Professor, Electrical Department, Veermata Jijabai Technological Institute, Mumbai, India
3 Senior R&D Engineer, A3 Remote Monitoring Technologies Pvt Ltd, India
4 Director, A3 Remote Monitoring Technologies Pvt Ltd, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - This Paper deals with detection of regional
symptomatic diseases at an earlier stage. This early detection
across various geographical locations will be helpful for early
diagnosis and for death prevention on a largerscale. Adopting
random methods for early detection and its respective factors
will be unsystematic. The proposed method is for detecting
disease epidemics/pandemics by considering large
symptomatic data and Natural Language Processing (NLP).
These Symptomatic data is available incommentsmadeby the
doctors during physiological data acquisition from the
database by hospitals. NLP will be used to find the common
diseases over different geographical locations.
Key Words: Regional symptomatic disease; natural
language processing; machine learning; physiological
data acquisition; statistical data analysis.
1. INTRODUCTION
Nowadays, healthcare is taken into account to be a
significant challenge. Infectiousdiseasesareamongthemost
serious health issues in the world. The emergence of these
diseases can be through air, water, direct contact with the
infected person, biologicallyandecological determinants [1].
In the year 2020, the world has witnessed an outbreak of
infectious diseases, whichisCorona Virus/Covid-19.Around
64 L deaths globally are reported till date and it is getting
multiplied at an awfully faster rate. Awareness of such
infectious diseases needs to be spread widely among the
people for the prevention of being infected in prior. These
infectious diseases go on spreadingoverlargerareasleading
to epidemics and pandemics. Also, these outbreaks have
major impacts on the population both socially and
economically.
In the situation of epidemics/pandemics,wheneverythingis
virtual, there are many places in our country which lack
medical facilities. The traditional wayoftreatmenttodisease
may not be enough in the case of serious problems.
Developing a medical diagnosis system based on Natural
Language Processing (NLP) and Machine Learning (ML)
algorithms for prediction of any disease can help in a more
accurate diagnosis and preventingthespreadforpandemics.
Accurate and on-time analysis of any health relatedproblem
is vital for the prevention and treatmentoftheillness.Hence,
detecting the spread of such epidemics / pandemics at an
early stage across various locations is going to be helpful for
early diagnosis and for death prevention on a bigger scale.
After identification of an emerging pandemic, detecting the
disease spread, local and international healthcare
organizations may be notified earlier in order that they will
take steps to halt the disease'sprogress[2].Thus, controlling
the epidemic diseases at the start of its spread may be a vital
solution for epidemics/pandemics.
2. LITERATURE SURVEY
Harini D K, Natesh M [3], In this paper, machine learning
algorithms is used for effective prediction of diseases. It uses
both structured and unstructured data from hospital for
effective prediction of diseases.Latentfactormodelisusedto
overcome the difficulty of missing data. A new convolutional
neural network based multi-modal disease risk prediction
(CNN-MDRP) algorithm is proposed in this paper. The
proposed algorithm accuracyprediction reaches94.8%than
that of the CNN-based unimodal disease risk prediction
(CNN-UDRP) algorithm.
Shratik J. Mishra, Albar M. Vasi , Vinay S. Menon, Prof. K.
Jayamalini [4] , The system implementedhadtheaccuracyof
86.67% on the dataset of 120 patient data. The current
system covered the general diseases or the more commonly
occurring disease, so that early prediction and treatment
could be done, and the fatality rate of deadly diseases
decreases, with the economic benefit.
Minsung Kim, Joon Yeop Lee, Hwangnam Kim [5],This paper
presents an Early Warning System (EWS) which is able to
predict infectious disease outbreaks and detect the sudden
increase of any livestock disease with the potentials to
become epidemic before spreading.
Pahulpreet Singh Kohli, Shriya Arora [6], In this paper,
different classificationalgorithmswereapplied, eachwithits
own advantage on three separate databases of disease
available in UCI repository for disease prediction.

3. PROPOSED WORK
The already existing system, for disease prediction, uses
various data processing types, which is the actual basic
foundation of Artificial Intelligence (AI) and Machine
Learning (ML). Natural Language Processing (NLP) refersto
one of the method of AI, which is concerned with giving
computers a potential to understand text and spoken words
in the same way humans can.
Figure 1: Proposed System Block Diagram
The Figure 1 shows the Block Diagram of the proposed
System. It has mainly two parts. The first part is the ‘disease
prediction’, that is using NLP algorithm to extract the
symptoms from complaints given by user and then
predicting the corresponding diseases. The second part is
the ‘plotting over geographical regions’ which uses the
latitude and longitude from differentstates,inordertoknow
the spread of diseases.
The main aim of the proposed system is to develop an
artificial intelligent system which detects the spread of
diseases over geographical locations. This can be done by
extracting the regional symptomatic data that is the
symptoms, in the form of complaints, given by the
paramedic/doctor during the physiological data acquisition.
The database contains tables, which has the information
about the patient details like name, age, gender, contact
details, visit date, complaints/symptoms the patient having
and location. Second table consisting of symptoms, disease
lists and mapping of symptoms to disease prepared by the
pre-registered doctors. Third table consists of city with
latitude and longitude of location. The patient
data will be stored on the web server so that the doctor can
access the information whenever required from
anywhere and does not have to be physically present.
Figure 2: Symptom Extraction Process
The Figure 2 shows the symptom extraction process. So, the
user enters the input in the form of complaints,whichcanbe
a sentence or a paragraph. Then the entered input will go
through the tokenization process, where it breaks down the
sentence or the paragraph into smaller chunks of words.
This smaller chunks will be processed with ‘stop word
removal’, to remove the stop words like ‘is, this, the, are, a,
an, etc.’ After, the stop words are removed, the result is
passed through removal of delimiters that is ‘ , and .’ . Later
when the delimiters as well as stop words areremovedfrom
the sentence/paragraph, the remaining words are referred
to as ‘candidate keywords’. Sometimes, the candidate
keyword can also be the extracted symptom from the
sentence/paragraph. The next is the segmentation process
where meaningful phrases are created from the data
obtained after the entire tokenization process.
And finally, the system will cross check the obtained
segmented data and the symptom list in the database. Each
time a match is found, will be considered as a symptomfrom
the user’s entered input. Later, these symptoms will be used
for further disease prediction.
For the disease prediction, the database will have the list of
symptoms and diseases which are prepared by the pre-
registered doctors. After the process of symptom extraction,
the system will make cross-matching with the database
which contains both ‘symptom and disease list’. Finally, the
system will then predict the disease.
4. PLOTTING OVER GEOGRAPHICAL REGIONS
AND GENERATING REPORTS
The final output is more beneficial when the system
generates report in the form of graph. The graph indicates
the rise of symptoms and disease location wise as well as
month wise. The Figure 3 shows the month-wise symptom
rise and the Figure 4 shows the state-wise symptom rise.
The system also generatesriseofepidemics/pandemicsover
geographical regions. The geographical region on map uses

latitude and longitude from the database for mapping the
symptoms over thestates.Afterselectingmultiplesymptoms
from the checkbox of the user interface, the map-view will
show the number of patient spread across different regions
having the selected symptom. Figure 5 shows the
geographical view of the symptom “Cold and cough”, ”Skin
problem” and “Body pain”, in the states of India. This
geographical view shows the total count of patients having
particular symptom in the form of bubble map. Multiple
symptoms on map can be distinguished by different color
code. When the cursor moves over to the coordinates of any
particular state or symptom, cursor will show the count of
the patients in that state for that particular symptom, as
displayed the count of patient for skin problem in the state
Gujarat. The dataset used for geographical plotting consists
of more than 7k entries.
Figure 3: Month wise symptoms rise
Figure 4: State wise Symptoms rise
Figure 5: Symptom rise over geographical regions
5. CONCLUSION
The proposed system aims at detecting the spread of
epidemics /pandemics over geographical map plot. The
reports generated in the results shown can be used by
medical or health organizations for analysis at national or
international level. The future work is to visualize the data
more prominently and add animations to the graphs. These
graphs can consider many other attributes like year, age,
gender, etc. Mapping over geographical regions can further
be done at district level as well as village level.
6. REFERENCES
[1] Inayatulloh, Selvyna Theresia, “Early Warning System
for Infectious Diseases”, DOI:
10.1109/TSSA.2015.7440435 ,IEEE 2015.
[2] Khanita Duangchaemkarn, Varin Chaovatut, Phongtape
Wiwatanadate, and Ekkarat Boon chieng, “Symptom-
based Data Preprocessing for the Detection of Disease
Outbreak”, DOI:10.1109/EMBC.2017.8037393 ,pp.
2614-2617, EEE, 2017.
[3] Harini D K, Natesh M , “Prediction Of Probability Of
Disease Based On Symptoms Using Machine Learning
Algorithm” ,International Research Journal of
Engineering and Technology (IRJET), Volume: 05 Issue:
05 | May-2018.

[4] Shratik J. Mishra, Albar M. Vasi , Vinay S. Menon, Prof. K.
Jayamalini, “GDPS - General Disease PredictionSystem”,
International Research Journal of Engineering and
Technology (IRJET), Volume: 05 Issue: 03 | Mar-2018.
[5] Minsung Kim, Joon Yeop Lee, HwangnamKim, “Warning
and Detection System for Epidemic Disease”,
DOI: 10.1109/ICTC.2016.7763517 ,pp. 478-483 ,IEEE,
2018.
[6] P. S. Kohli and S. Arora, “Application of Machine
Learning in Disease Prediction”, 4th International
Conference on Computing Communication and
Automation(ICCCA),DOI:10.1109/CCAA.2018.8777449
,pp. 1-4 , 2018.
[7] Fariz Bramasta Putra et. al., “IdentificationofSymptoms
Based on Natural Language Processing (NLP) for
Disease Diagnosis Based on International Classification
of Diseases and Related Health Problems”,International
Electronics Symposium (IES), DOI:
10.1109/ELECSYM.2019.8901644, pp. 1-5 ,IEEE, 2019.
[8] Aswathy K P, Rathi R, Shyam Shankar E P, “NLP based
Segmentation Protocol for Predicting Diseases and
Finding Doctors”, International Research Journal of
Engineering and Technology (IRJET) , Volume:06Issue:
02 | Feb 2019.
[9] PM. Lavanya , E. Sasikala, “Deep LearningTechniqueson
Text Classification Using Natural Language Processing
(NLP) In Social Healthcare Network: A Comprehensive
Survey”, International Conference on Signal Processing
and Communication, DOI: 10.1109/
ICSPC51351.2021.9451752, pp. 603-609,IEEE, 2021.

Natural Language Processing and summarization of medical symptomatic data from geographical diverse locations

More Related Content

Similar to Natural Language Processing and summarization of medical symptomatic data from geographical diverse locations (20)

More from IRJET Journal (20)

Recently uploaded (20)

Natural Language Processing and summarization of medical symptomatic data from geographical diverse locations