Prediction of Dengue, Diabetes and Swine Flu using Random Forest Classification Algorithm

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 06 | June -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 685
Prediction of Dengue, Diabetes and Swine Flu Using Random Forest
Classification Algorithm
Amit Tate1, Ujwala Gavhane2, Jayanand Pawar3, Bajrang Rajpurohit4 , Gopal B. Deshmukh5
1,2,3,4UG Student, M.E.S. College of Engineering, Pune, SPPU
5Deparent of Computer Engineering, M.E.S. College of Engineering, Pune, SPPU
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract: - In this article we proposed disease prediction
system using Random Forest Algorithm (RFA). Training
dataset is used for prediction of particular disease. The main
aim of this article is that to predict the disease which input
symptoms is taken from patient or user. Recommend a
specialized doctor for particular disease if result positive. Our
algorithms are extendible to dealwithmobile/onlinesolutions
to support patients as well for medical diagnostics. As a first
step, we also developed web-based interfaces to support
patients in calculating risk level for each medical case.
Keywords: - Random Forest Algorithm (RFA), Machine
Learning, Out-Of-Bag (OOB).
1. INTRODUCTION
Supervised learning is part of Machine learning that consist
training dataset which is labelled. In proposed system with
the help of supervised learning we can predicttheclasslabel
which from user input. Disease prediction has become
important in a variety of applications such as health
insurance, tailored health communication and publichealth.
Disease prediction is usually performed using publicly
available datasets.
Disease involved in prediction system as following
A. Swine Flu
B. Diabetes
C. Dengue
Swine flu is a respiratory diseasecausedbyinfluenza viruses
that infect the respiratory tract of pigs and result in a
barking cough, decreased appetite, nasal secretions, and
listless behavior; the virus can be transmitted to humans.
Diabetes is a number of diseases that involve problems with
the hormone insulin. Normally, the pancreas (an organ
behind the stomach) releases insulin to helpyour bodystore
and use the sugar and fat from the food you eat.
Dengue fever is a painful, debilitating mosquito-borne
disease caused by any one of four closely related dengue
viruses. These viruses are related to the viruses that cause
West Nile infection and yellow fever.
Disease prediction using Random forest algorithm is
proposed for Dengue, Diabetes and Swine Flu diseases.
Training dataset is given for predict the particular disease.
Training dataset for each disease is described in III section.
Fig -1: Proposed system for disease prediction system
using Random Forest Algorithm.
In fig 1 contains detailed description of proposed disease
prediction system. Firstly, user log in into our system,ifuser
don’t have login then they need to register itself. After
successfully login user/patient can check their disease
details using prediction system. User/Patient need to enter
or select their symptoms for particular disease (Diabetes,
Swine flu, Dengue). Once user select the disease and their
symptoms then our prediction system will predict the
disease using training dataset. Result should be show
Positive either Negative result. If system gives to user

Positive result, then system should shows recommend
doctor for those disease and user details sent to specialized
doctor for that particular disease. Patients can make
appointment directly through recommendation system.
Doctor will response back to patient quickly. If system
gives to user negative result, then system should show
precautions for those disease.
2. RELATED WORK
Author introduced classification algorithmsusedfordisease
prediction and comparison of each algorithm is described in
details with graph analysis such as Prediction time, Recall,
Precision, TP (True Positive) rate, FP (False Positive) rate.
[1]. Here our system is using Random forest algorithm
because author introduced this algorithm is best for disease
prediction. [1]
The RandomForestapproachoffersthehighestaccurateness
rate reaching 99.9% with 9 measured factors after the
reduction step. In this paper, author describes four case
studies fromtheLebanesehealthdomain,AcuteAppendicitis
(AP), Premature Birth (PB), Coronary Heart Disease (CHD),
and Osteoporosis Disease (OD). For these applications,
prediction systems were developed for decision support
using data mining techniques. [2]
In this paper, author proposed a model based on four years’
medical analysis data to predict the change possibility ofthe
coming year’s FBG (Fasting blood glucose). Based on four
years’ historic medical examination data, a predictionmodel
of coming year’s FBG is offered using traditional data mining
techniques with Random forest algorithmandSVM(Support
Vector Algorithm). [3]
3. PROPOSED SYSTEM
Proposed system is Disease prediction system which can
predict the disease with the help of training datasetanduser
input.
Proposed system includes the Patient healthcare, Disease
prediction (Dengue, Diabetes, and Swine flu), Doctor
Recommendation for particular disease, Precaution for
disease, Doctor details for when patients need to contact
them directly, Make an appointment withspecializeddoctor.
Disease prediction means that user will gives some input to
our system in the form of yes or no for symptoms of
particular disease and this system will proceedonthatgiven
symptoms. Finally predict the accurate disease using
training dataset and given input.Randomforestwill proceed
on that given symptoms and predict the accurate output.
4. TECHNOLOGY OVERVIEW
Random Forest algorithm are an ensemble supervised
learning method which is used as predictor of data for
classification and regression. In the classification process
algorithm build a number of decision trees at training time
and construct the class that is the mode of the classes output
by using each single tree. (Random Forests is introduced by
Leo Breiman and Adele Cutler for an ensemble of decision
trees). [5]
Random Forest algorithm is a grouping of tree predictors
where each tree based on the values of a random vector
experimented independently with the equal distribution for
all trees in the forest. The basic principle is that a group of
“weak learners” can come together to form a “strong
learner”. Random Forests are a perfect tool for making
predictions considering they do not over fit. Presenting the
accurate kind of randomness makes them accurate
classifiers and regression. [5]
Single decision trees often have high variance or high bias.
Random Forests trying to moderate the high variance
problems and high bias by averaging to find a natural
balance between the two extremes. Considering that
Random Forests have few parameters to tune and can be
used simply with default parameter settings, they are a
simple tool to use without having a model or to produce a
reasonable model fast and efficiently.
Random Forests are easy to learn and use for both
professionals and lay people - with little research and
programming required and may be used by folks without a
strong statistical background. Simply put, you can safely
make more accurate predictions without most basic
mistakes common to other methods.
Random Forests produces several classifications for given
trees. Each tree is grown as follows:
1. If number of circumstances in the training data set is D,
sample D cases at random state but with replacement, from
the original dataset. This sample testing set will be the
training set for increasing the tree.
2. If there are input variables from training dataset, a
number is indicated such that at each node of the tree, m

variables are selected at random available for the and the
best splitting on these is used to splitting the node. The
value of is used as constant during entire the forest
growing.
3. Each tree is grown to the largest size as possible. There
is no pruning an overall grownup tree.
The random forest algorithm is an ensemble classifier
algorithm based on the decision tree model. It generates k
different training data subsets fromanoriginal datasetusing
a bootstrap sampling approach, and then, k decision trees
are built by training these subsets. A random forest is finally
constructed from these decision trees. Each sample of the
testing dataset is predicted by all decisiontrees,andthefinal
classification result is returned depending on the votes of
these trees. [11]
Fig 2. Process of the construction of the RF algorithm
The original training dataset is formalized as
S = {(ai, bj), i = 1, 2... D; j = 1, 2... },
Where A is a sample and b is a feature variable of S. Namely,
the original training dataset contains D samples, and there
are feature variables in each sample.
The main process of the construction of the RF algorithm is
presented in Fig. 2[1]
The steps of the construction of the random forestalgorithm
are as follows:
Step 1. Sampling k training subsets.
In the first step, k training datasets are experimented from
the original training dataset S in a bootstrap selection
manner. Namely, N records are selected from S by a random
sampling and replacement method in each sampling time.
After the current step, k trainingsubsetsareconstructedasa
collection of training subsets S Train:
S Train = {S1, S2…….Sk}.
At the same time, the records that are not to be selected in
each sampling period are composed as an Out-Of-Bag(OOB)
dataset.
In this way, k OOB sets are constructed as a collection of
SOOB:
SOOB = {OOB1, OOB2... OOBk},
Where k ≪ N, Si ∩ OOBi = ϕ and Si ∪ OOBi = S.
To obtain the classification accuracy of each tree model,
these OOB sets are used as testing sets after the training
process. [1]
Step 2. Constructing each decision tree model.
In an RF model, each Meta decision tree is created by a C4.5
or CART algorithm from each training subset Si. In the
growth process of each tree, m feature variablesofdatasetSi
are randomly selected from M variables. In each tree node’s
dividing process is done, then gain ratio of each feature
variable is computed, and the best one or most priority node
is chosen as the splitting node. This splitting process is
repeated until a leaf node is generated. Finally, k decision
trees are trained from k training subsetsinthesameway. [1]
Step 3. Collecting k trees into an RF model.
The k trained trees are collected into an RF model, which is
defined in Eq. (1):
where hi(x,Θj) is a meta decision tree classifier, X are the
input feature vectors of the training dataset, and Θj is an
independent and identically distributed random vector that
determines the growth process of the tree. [1]

5. HOW IT WORK
Proposed system disease prediction system fully based on
training dataset. If dataset incorrect then output could be
wrong. In table no 1, described symptoms of Swine flu. In
table no 3, described symptoms of Diabetes. In table no 5,
described symptoms of Dengue.
E.g. In table no 2 user inputs are all yes for all symptoms or
any input some of them is yes and some of them is no. If user
select the yes for all symptoms,thenresultshouldbepositive
and if user select the no for all symptoms then result should
be negative. Input from user can multiple possibilities such
as some input could be yes and some input could be no. After
that Random forest algorithm will proceedonuserinput and
using training dataset will predict the output as Positive or
Negative.
If output is positive, then specialized doctor will display on
same page. Patient can make an appointment directly with
specialized doctor for particular disease. Predicted result is
positive then Patients details will have sent to particular
doctor which is specialize in that disease.
If output is negative, then precaution of that disease will
display on same page. Patient can make use of online
healthcare for another disease also. Healthcare include the
main disease which spread over population quickly.
Table -1: Symptoms of swine flu disease
Symptoms
for Swine
Flu Disease
1. Chills
2. Fever
3. Coughing
4. Sore Throat
5. Fatigue
6. Nausea
7. Vomiting
8. Diarrhea
9. Runny Nose
10. Stuffy Nose
11. Body Aches
Table - 2: Default user input for positive or negative
output for swine flu
Input from user Result
yes, yes, yes, yes, yes, yes,
yes, yes, yes, yes, yes
Positive
no, no, no, no, no, no
no, no, no, no, no
Negative
In table no 2, user input can be all yes and can be no
for each symptom or can be some Yes and can be some No
depend on what patient givesinput.Wementionedhereonly
two test cases for predict result first one is if user select yes
for all symptoms and second is user select no for all
symptoms. We need more training dataset to predict
accurate output which is not mentioned in this article but
another training dataset is used in proposed system.
Table - 3: Symptoms of diabetes disease
Symptoms
for Diabetes
Disease
1. Heavy thirst
2. Increased hunger
3. Dry mouth
4. Pain in belly
5. Fatigue
6. Nausea
7. Vomiting
8. Frequent urination
9. Unexplained weight
loss
10. Blurred vision
11. Heavy laboured
breathing
Table – 4: Default user input for positive or negative
output for diabetes disease
yes, yes, yes, yes, yes, yes,
yes, yes, yes, yes, yes
Positive
no, no, no, no, no
Negative
Table -5: Symptoms of dengue disease
Symptoms
for Dengue
1. Sudden high fever
2. Severe headaches
3. Severe joint and
muscle pain
4. Pain behind the eye
5. Fatigue
6. Nausea
7. Vomiting
8. Skin rash which
appears two to five
days after the onset
of fever
9. Mild bleeding

Table -6: Default user input for positive or negative
output for dengue
yes, yes, yes, yes, yes, yes, yes, yes,
yes
Positive
no, no, no
Negative
System architecture of disease prediction system using
Random Forest Algorithm described in fig 1. Classification
process of disease prediction described below fig no 1. In
this process user input and training dataset is compared.
In Fig, no 1 User input will proceed using Random forest
algorithm. RFA will generates the maximum numberoftrees
with help of given training dataset and the output is in the
form of Yes (Positive prediction) or No (Negative
Prediction). Each tree would have single output. Fig no 3
Contains tree 1, tree 2…. tree N will have separate result.
Result 1, Result 2…. Result N shows you result of given
symptoms which is calculated by tree’s using training
dataset in the form of Yes or No. Finally, Result is combined
in two categories Yes or No, which tree has maximum
number of Yes or No this result will show the final output as
if Maximum tree is Yes then result should be Positive and if
Maximum tree is No then result should be Negative.
Fig -3: System architecture for disease prediction system
6. RESULTS
In table no 1, symptoms of swine flu are given. Using these
symptoms disease prediction system will predict result with
help of user input. Swine flu output in case of positive result.
In table no 2, two cases of swine flu are given for positive
and negative result. Positive result should display in case of
only input is Yes from user probably greater than No. Fig 4
and fig 5 contains positive and negative result respectively.
Fig -4: swine flu positive result
Negative result should display in case of only input is No
from user probably greater than Yes.
Fig -5: swine flu negative result
In table no 3, symptoms of diabetes are given. In table no 4,
two cases of diabetes are given for positive and negative
result. Fig 6 and fig 7 contains positive and negative result of
diabetes respectively.
Fig -6: Diabetes Positive Result

Fig -7: Diabetes Negative Result
In table no 5, symptoms of dengue are given. In table no 6,
two cases of dengue are given for positive and negative
result. Fig 6 and fig 7 contains positive and negative result of
dengue respectively.
Fig -8: Dengue Positive Result
Fig -9: Dengue Negative Result
7. CONCLUSION
In proposed Diseasepredictionsystemcanpredictparticular
disease using training dataset. In this article, we proposed
disease prediction system as web/mobile based online
applicationfor patient’shealthcare.Randomforestalgorithm
maintains best accuracy as compare to others classification
system. After result predict the disease, then
recommendation system will work on their predicted
disease. If positive result show to user, then recommended
doctor will display on same page. If negative result, then
precautions for same disease will display.
Reference
[1] Amit Tate, Bajrangsingh Rajpurohit, Jayanand
Pawar, Ujwala Gavhane,Gopal B.
Deshmukh."Comparative Analysis of Classification
Algorithms Used for Disease Prediction in Data
Minin" Vol. 2 - Issue 6 (Nov - Dec 2016),
International Journal of Engineering and
Techniques (IJET), ISSN: 2395 - 1303,
www.ijetjournal.org.
[2] Ahmad Shahin, Walid Moudani, Fadi Chakik,
Mohamad Khalil “Data Mining in Healthcare
Information Systems: Case Studies in Northern
Lebanon” Doctoral School for Science and
Technology The Lebanese University Tripoli,
Lebanon.
[3] Wenxiang Xiao, Jun Ji, Fengjing Shao* , Rencheng
Sun, Chunxiao Xing on “Fasting Blood Glucose
Change Prediction Model Based on Medical
Examination Data and Data Mining Techniques”
2015 IEEE International Conference on Smart City/
Social Com /Sustain Com together with Data Com
2015 and SC2 2015
[4] Mihail Popescu ; Mohammad Khalilia “Improving
disease prediction using ICD-9ontological features”
Fuzzy Systems (FUZZ), 2011 IEEE International
Conference on June 2011.
[5] http://www.datasciencecentral.com/profiles/blo
gs/random-forests-algorithm.
[6] Asmaa S. Hussein; Wail M. Omar; Xue Li ; Modafar
Ati “Efficient Chronic Disease Diagnosis prediction
and recommendation system. Biomedical
Engineering and Sciences (IECBES), 2012 IEEE
EMBS Conference.
[7] April Morton ; Eman Marzban ; Georgios
Giannoulis ; Ayush Patel ; Rajender Aparasu ;
Ioannis A. Kakadiaris “A Comparison of Supervised
Machine Learning Techniques for Predicting Short-
Term In-Hospital Length of Stay among Diabetic
Patients” Machine Learning and Applications
(ICMLA), 2014 13th International Conference on
Dec. 2014(IEEE).
[8] http://www.webmd.com.

Prediction of Dengue, Diabetes and Swine Flu using Random Forest Classification Algorithm

More Related Content

What's hot (20)

Similar to Prediction of Dengue, Diabetes and Swine Flu using Random Forest Classification Algorithm (20)

More from IRJET Journal (20)

Recently uploaded (20)

Prediction of Dengue, Diabetes and Swine Flu using Random Forest Classification Algorithm