Leveraging Flask API and Machine Learning to Forecast Multiple Diseases

Communications on Applied Nonlinear Analysis
ISSN: 1074-133X
Vol 32 No. 1s (2025)
Leveraging Flask API and Machine Learning to Forecast Multiple

Diseases
K Venkateswara Rao1,D Srilatha2,Sridevi Sakhamuri3,Venkata Subbaiah Desanamukula4,M

Asha Priyadarshini5,P Ramya6
1
Professor, Department of CSE, Vignan’s Lara Institute of Technology & Science, Vadlamudi, Guntur(dt), AP,India,
venkat545@gmail.com
2
Professor, Department of EEE, Vasi Reddy Venkatadri Institute of Technology, Namburu-522508, Guntur(dt),AP,India.
srilatha.dande@gmail.com
3
Assistant Professor, Department of IoT, Koneru Lakshmaiah Education Foundation, Vaddeswaram, AP, India, 522302.
srisat1617@gmail.com
4
Professor, Department of Computer science and Engineering, Lakireddy Bali Reddy College of Engineering,
Mylavaram, NTR Dist Andhra Pradesh, INDIA. desanamukula@gmail.com
5
Assistant Professor, Department of CSE, Vignan’s Lara Institute of Technology & Science,
Vadlamudi, India. ashapriyadarshini16@gmail.com
6
Assistant Professor,Department of CSE,Vignan’s Lara Institute of Technology & Science, Vadlamudi, India.
ramya458@gmail.com
Article History: Abstract:

The study described in this abstract used machine learning to predict several diseases,
Received: 09-08-2024
and it integrated the model into a Flask API. A user-friendly platform for disease
Revised: 25-09-2024 prediction based on patient symptoms and medical history was the goal of the paper.
Accepted: 12-10-2024 A number of strategies were used to optimize the predictive performance once the
machine learning model had been trained on a sizable dataset of patient records. The
model was made accessible as a web service through the usage of the Flask API,
enabling simple deployment and integration into current healthcare systems. The
study's findings demonstrated that the model could accurately forecast diseases with
a high degree of precision, and the Flask API gave healthcare professionals an easy
way to use the model in real-world settings. The study underscores the significance
of developing open and user-friendly platforms for healthcare applications and shows
the potential of machine learning and online APIs in improving disease detection.
Additionally, the study discovered that the model could correctly identify disorders
even in the absence of conclusive diagnostic tests.
Keywords: Flask, ExtraTrees classifier, Grid Search CV.
I. INTRODUCTION
With the development of technology and machine learning, the field of medical science has
experienced enormous progress in recent years. The foretelling of diseases is one such application of
machine learning. Machine learning algorithms can evaluate trends and estimate the likelihood that a
person would contract a specific disease with the aid of massive datasets. In identifying and treating
different disorders, this can be very helpful to medical experts. As fast and precise diagnosis is
essential for successful treatment and positive patient outcomes, disease prediction is a vital
responsibility in healthcare. Recent developments in machine learning have made it possible to
150
https://internationalpubls.com
ISSN: 1074-133X
Vol 32 No. 1s (2025)
create predictive models that can correctly identify diseases based on the symptoms of the patient
and their medical background. The absence of user-friendly interfaces and technical obstacles have
prevented these models from being fully implemented in real-world healthcare settings, despite their
potential. Through the creation of a machine learning model for disease prediction and its integration
into a Flask API, this work tries to address these issues. In order to make machine learning models
easily accessible to healthcare practitioners, Flask is a lightweight web framework for Python that
enables their deployment as web services. It is simpler for healthcare providers to apply the model in
practice since it has been integrated into a Flask API, which enables convenient and user-friendly
access to the model. This study's goals are to show that machine learning and Flask APIs may be
used to improve disease detection and to underline how crucial it is to design accessible and user-
friendly platforms for healthcare apps. To train the machine learning model, the study will use a
sizable dataset of patient records, and it will assess the model's performance in terms of accuracy and
speed. The study's findings will shed light on the potential of machine learning and Flask APIs in the
healthcare sector and guide future work on creating systems similar to this one for disease prediction.
Machine learning can be used to diagnose four important diseases, including pneumonia, malaria,
liver disease, and kidney disease. The algorithms may be taught to spot patterns and symptoms
related to each disease by using large datasets of patient information and medical images for machine
learning training datasets of patient information and medical images for machine learning training. In
impoverished nations with limited access to healthcare specialists, this can significantly increase the
efficacy and accuracy of diagnostics. For instance, machine learning models can use information
about kidney function, such as creatinine levels and many other parameters, to forecast the risk that a
patient would acquire kidney disease. In a similar way, chest X-rays can be used to identify
pneumonia, and liver function tests can be studied to identify liver illness. Machine learning
algorithms can examine microscope images in the instance of malaria to determine whether the
parasite is present. In conclusion, using machine learning to diagnose kidney, liver, malaria, and
pneumonia disorders has the potential to completely transform the healthcare sector by delivering
quicker and more precise diagnoses, particularly in situations with limited resources.
II. LITERATURE REVIEW
A. System for Support Vector Machine-Based Diagnosis of Chronic Kidney Disease:
In this paper, we propose the diagnosing the kidney infection utilizing Support Vector Machine.
SVM is the calculation is employed for grouping and bunching. The essential thought of the
calculation Support Vector Machine is searching for the ideal limit that is utilized to isolate the two
bits of a class called the hyperplane. Scientists have applied SVM to do the arrangement task over
different fields, for example, emotional registering. The motivation behind the improvement of this
framework is a choice emotionally supportive network in diagnosing of kidney illness what's more,
foreseeing whether or not renal disease patients have started to have persistent kidney disease.
B. Liver Disease prediction using Classification Algorithms:
In this exploration work, different characterization calculations in particular Support Vector Machine
and K-Nearest Neighbors have been utilized for liver illness expectation. The correlation of this
multitude of calculations been finished in light of characterization exactness which is viewed as
151
ISSN: 1074-133X
Vol 32 No. 1s (2025)
through confusion grid. From the trial, Logistic Regression also, K-Nearest Neighbors have the most
noteworthy exactness yet Logistic Regression have the most note worthy responsiveness. In this way
it can be reasoned that Logistic Regression is fitting for anticipating liver illness.
C. Malaria Disease Prediction Based on Machine Learning:
As referenced, this paper has effectively satisfied its objectives with a significant commitment over
existing machine learning innovation where ELM beats generally existing models on Malaria RBC
illness expectation with a precision of 97. Furthermore, this outcomes in a high preparation speed. Its
execution is straightforward, and addressing complex can be utilized issues in related region. Inside
the benefits of ELM and its extremely execution contrasted and other ML, its verification that ELM
is strongly suggest being utilized in Malaria fever RBCs prediction.
D. Deep Learning for Pneumonia Diagnosis from Chest X-Ray Images:
The two CNN networks coverage of the pneumonia disease was taken into consideration when
writing this evaluation. Transfer learning and fine-tuning were both used in the development of our
model. They examined the outcomes of two network tests after the stage of preparation. VGG16
network performs worse than exception network for identifying pneumonia patients. At separating
common cases, the VGG16 network is more successful concurrently.
III. PROPOSED WORK AND DESIGN
METHODS FOR MODELS
A. SYSTEM FOR EXISTING APPROACH
A significant number of surviving investigations included examining specific sickness. At the point
when a user needs to examine diabetes necessities to utilize one doctor’s prescription and same user
needs to investigate coronary illness then client needs to utilize another doctor’s prescription. This is
a time taking process. And further more in the event that any client having more than one sickness be
that as it may, in existing framework in the event that anticipating just a single disease is capable
then there is an opportunity of death rate increment due to not ready to foresee the other illness ahead
of time.
B.PROPOSEDAPPROACH
It is possible to predict more than one illness at a time using a multi-sickness model. So, user
compelling reason need to navigate many models to foresee the illnesses. It will lessen time and
further more due to foreseeing various illnesses all at once there is a possibility decreasing death rate.
In multiple diseases prediction, it is feasible to anticipate more than each illness in turn. So, the client
doesn't need to cross various destinations to anticipate the sicknesses. We are taking four sicknesses
that are Kidney, Liver, Malaria and Pneumonia. To execute numerous illness investigations, we will
utilize ML algorithms and deep learning algorithms and Flask. At the point when the client is getting
to this Programming interface, the client needs to send the boundaries of the sickness alongside the
sickness name.
The following stages are part of the suggested method for using machine learning and the Flask API
to predict kidney, liver, malaria, and pneumonia diseases. The initial stage is to acquire sizable
152
ISSN: 1074-133X
Vol 32 No. 1s (2025)
datasets of patient data and medical pictures pertinent to the diseases of interest. Then, any
unnecessary or missing information is removed from this data through preprocessing and cleaning.
After preprocessing the data, supervised learning techniques are used to train machine learning
algorithms. On the basis of the labeled data, the algorithms learn to make predictions (i.e., data
where the diagnosis is known). To reach an acceptable degree of accuracy, the models can be
adjusted and retrained as necessary. After training, numerous metrics, including accuracy, precision,
recall, and F1 score, are used to assess how well the machine learning models performed. The
models are then adjusted or retrained in order to obtain an acceptable degree of accuracy.
The development of a Flask API incorporates the machine learning models after they have been
trained and assessed. The API returns the estimated likelihood that a patient will contract a specific
disease after receiving patient data and medical photos as inputs. The Flask API is then made
available to users over the internet by being deployed to a web server. This makes it simple for
individuals and medical professionals to utilize the disease prediction system and get precise
diagnoses. A user-friendly interface that enables patients to simply input their data and medical
photos and receive forecasts regarding the propensity to develop a specific condition can be created.
The Flask API can be combined with this interface to increase the illness prediction system's
usability and accessibility. By allowing for early intervention to stop the onset of catastrophic
diseases and by offering quick and precise diagnostics, this technology has the potential to
completely transform the healthcare sector.
C.PROPOSED SYSTEM ARCHITECTURE
Fig:1 Proposed System Architecture

D.DATASET PREPARATION
For a framework to anticipate legitimate outcomes, first it ought to be prepared appropriately with
existing information. Pre-Handling the information is significant so great quality information is
153
ISSN: 1074-133X
Vol 32 No. 1s (2025)
utilized for preparing the model. Information cleaning and evacuation of missing values is a portion
of the cycles engaged with Pre-processing. We utilized the Datasets of diseases from Kaggle
Information from different sources has been gathered and collected
E.PREDICTION USING VARIOUS DEEP LEARNING AND MACHINE LEARNING
ALGORIHTMS
Article primary center is to construct a multi disease prediction model so ML and profound learning
strategies utilized are momentarily summed up here. Kidney disease, Liver sickness prediction and
malaria and pneumonia discovery are broken down by different profound learning methods have
been utilized to view the status of the patient. For Kidney disease prediction we have utilized many
ML algorithms like Gradient Boosting classifier, Stochastic Gradient Boosting, XG Boost, Ada
Boost, Cat Boost classifier, Gradient Boosting classifier etc. Among these Extra Trees classifier
results about 99% exactness, for Liver disease characterization Random forest yield 77% exactness
among all other algorithms such as Ada Boost, Gradient Boosting, Randomized search CV, Grid
search CV. Malaria and Pneumonia examination contains red blood cell pictures and chest X ray
pictures. So utilized the python’s tensor flow library for investigating the pictures. Convolution
neural networks and deep learning techniques are used in the model's construction and testing on the
test set. For Malaria prediction we have used Inception v3 algorithm. Inceptionv3is a convolution
neural network for aiding picture investigation and item discovery. The discussed algorithm gave
96% accuracy. For Pneumonia prediction we have utilized VGG 19
algorithmwhichgaveabout92%accuracy.
F.DATA FLOW
Fig:2 Working Model of the proposed system
154
ISSN: 1074-133X
Vol 32 No. 1s (2025)
We conducted experiments on these four diseases in the above figure. The Chronic Kidney disease
dataset, Indian Liver dataset, Malaria cells Images dataset and chest X-Rays Image dataset
respectively, have been imported as the first stage for the datasets for Kidney disease, Liver disease,
Malaria and pneumonia disease. The depiction of each inputted value occurs once the dataset has
been imported. The data is split into training as well as testing on the updated dataset following pre-
processing, which includes checking for outliers, missing values, and scaling the dataset. Next, we
applied various learning algorithms to the training dataset, then we used the testing dataset to apply
what we knew about the classified algorithm. Knowledge will be put to use, and the best algorithm
will be chosen. The model behavior of the best algorithm for each disease is saved using pickle
processing, which includes checking for outliers, missing values, and scaling the dataset. Next, we
applied various learning algorithms to the training dataset, then we used the testing dataset to apply
what we knew about the classified algorithm. Knowledge will be put to use and the best algorithm
will be chosen. The model behavior of the best algorithm for each disease is saved using pickle.
IV. DATASETS USED:
KIDNEY DISEASE DATASET:
Table:1 Sample Kidney Disease Dataset.
155
ISSN: 1074-133X
Vol 32 No. 1s (2025)
LIVER DISEASE DATASET:

Table:3 Sample Liver Disease Dataset.
MALARIA DISEASE DATASET:

Table:4 Sample Malaria Disease Dataset.
156
ISSN: 1074-133X
Vol 32 No. 1s (2025)
PNEUMONIA DISEASE DATASET:

Table:5 Sample Pneumonia Disease Dataset.
V. RESULTS
LOGIN AND REGISTRATION PAGE:
INTERFACE FOR KIDNEY DISEASE PREDICTION:
157
ISSN: 1074-133X
Vol 32 No. 1s (2025)
RESULT FOR KIDNEY DISEASE PREDICTION:
INTERFACE FOR LIVER DISEASE PREDICTION:
RESULT FOR LIVER DISEASE PREDICTION:
158
ISSN: 1074-133X
Vol 32 No. 1s (2025)
INTERFACE FOR MALARIA DISEASE PREDICTION:
RESULT AFTER UPLOADING THE RBC PICTURE:
INTERFACE FOR PNEUMONIA DISEASE PREDICTOR:
159
ISSN: 1074-133X
Vol 32 No. 1s (2025)
DISPLAYING THE STATUS OF PATIENT AFTER UPLOADING CHEST X-RAY

PICTURE:
ACCURACIES:
DISEASE BEST ALGORITHM DETECTED RESULTED ACCURACY

KIDNEY DISEASE Extra trees classifier 99%
LIVER DISEASE Random Forest 77%
MALARIA Inceptionv3 96%
PNEUMONIA VGG-19algorithm 92%
VI.CONCLUSION
To sum up, the application of machine learning and the Flask API to the multi-illness prediction
process has the potential to significantly raise the precision and effectiveness of disease detection in
the healthcare sector. Machine learning algorithms can learn to spot patterns and symptoms
associated with different diseases by being trained on vast datasets of patient information and
medical pictures. This allows for more precisediagnosis. Healthcare professionals and patients can
readily access the disease prediction system online thanks to the machine learning models'
deployment as a Flask API. The creation of a user-friendly interface can improve the disease
prediction system's usability and accessibility by allowing patients to enter their data and get
estimates of how likely they are to contract a specific condition. Overall, the application of machine
learning and the Flask API for multi-disease prediction has the potential to significantly raise the
standard of healthcare, particularly in places with limited resources, and to save lives by enabling
early intervention and the avoidance of dangerous diseases.
VII. FUTURE ENHANCEMENTS
The future enhancements for multi-disease prediction using machine learning and Flask API could
include
160
ISSN: 1074-133X
Vol 32 No. 1s (2025)
Improved data collection: As more patient information and medical images become available, the
size and quality of the datasets used to train the machine learning algorithms can be improved,
leading to more accurate diagnoses
Integration with electronic health records (EHRs):
The disease prediction system could be integrated with EHRs, allowing healthcare professionals to
quickly and easily access patient information and receive predictions about the likelihood of
developing a particular disease
Real-time predictions:
The Flask API could be optimized to provide real-time predictions, enabling healthcare professionals
to make fast and accurate diagnoses and enabling early intervention to prevent the onset of serious
diseases.
Incorporation of additional data sources: The machine learning algorithms could be trained on
additional data sources, such as genetic information or environmental data,to provide even more
accurate predictions.
Improved interpretability:
The machine learning algorithms could be developed to provide interpretable predictions, allowing
healthcare professionals to understand the reasoning behind a particular prediction and to make more
informed decisions.
Personalized predictions:
The disease prediction system could be developed to provide personalized predictions based on
individual patient information, leading to more accurate diagnoses and better patient outcomes.
VIII.REFERENCES
[1] E.E.Shair, M.M.Mehdyand P.Y.Ng, "Artificial Neural Networks in Image Processing for Earlier Detection of Breast
Cancer".
[2] Jwan K. Alwan,Mohammad B. Naeem Mohamed Ibrahim, DhafarHamed,"The Utilisationof Machine Learning
Approaches for Medical Data Classification
[3] Rakesh Kumar,Archana Singh “Heart Disease Prediction Using Machine Learning Algorithms”
[4] Baddigam Jaya Krishna Reddy,A.Sivasangari, Annamareddy Kiran, P.Ajitha,” Diagnosis of Liver Disease using
Machine Learning Models”
[5] Priyanka Sonar, Prof. K. JayaMalini,” Diabetes Prediction Using Different Machine Learning Approaches.
[6] ”Support vector machine based disease classification model employing hasten eagle cuculidae search optimization”,
concurrency and computation: practice and experience, ISSN : 1532-0626, Vol-34, Issue-25 (Nov-2022), Wiley
Publisher.
[7] “Regression based price prediction of staple food materials using multivariate Models”, Scientific Programming,
ISSN : 1058-9244, Vol-2022(June), Hindawi Publisher.
[8] “Parkinson’s Disease Prediction and Classification Using Machine Learning Techniques”, Journal of
Optoelectronics Laser, ISSN : 1005-0086, Vol-41, Issue-5, 2022, Page No: 690-698.
[9] “A Study on Defensive Issues and Challenges in Internet of Things”, Lecture Notes in Electrical [10] ”Disease
Prediction and Diagnosis Implementing Fuzzy Neural Classifier based on IoT and Cloud”, International Journal of
Advanced Science and Technology (IJAST), ISSN : 2005-4238, Vol-29 Issue-5, May 2020, Page No: 737-745.
161
ISSN: 1074-133X
Vol 32 No. 1s (2025)
[11] " Rotating Solar Trees ", Lecture Notes in Electrical Engineering 601, Springer Nature Singapore Pte Ltd. 2020,
Page No: 482-487.
[12] " Wireless-Sensor-Network with Mobile Sink Using Energy Efficient Clustering ", Lecture Notes on Data
Engineering and Communications Technologies , Springer Nature Switzerland AG. 2020, Page No: 582-589.
[13] "Research of Feature Selection Methods to Predict Breast Cancer", International Journal of Recent Technology and
Engineering(IJRTE), ISSN : 2277-3878, Vol-8 Issue-2s11, Sep 2019, Page No: 2353-2355.
[14] ”Suicide Prediction on Social Media by Implementing Sentimental Analysis along with Machine Learning”,
International Journal of Recent Technology and Engineering(IJRTE), ISSN : 2277-3878, Vol-8 Issue-2, July 2019,
Page No: 4833-4837.
[15] “Issues for building an Artificial Intelligent System” JARDCS, Vol-7 Issue-13, Dec 2018, ISSN : 1943-023X, Page
No: 1447-1451.
[16] “Late Patterns In Chart Model For Content Examination And Content Mining” IJPT, ISSN: 0975-766X, Volume-8,
Issue-2, June-2016, page no: 14729-14736.
[17] Priyadarshini, M. Asha, et al. "A Multi-Feature Approach with Data Augmentation for Speech Emotion
Recognition using Deep Learning." ICCIET-2024. Atlantis Press, 2024 https://doi.org/10.2991/978-94-6463-471-
6_80
[18] Patro, Pramoda, et al. "A hybrid approach estimates the real-time health state of a bearing by accelerated
degradation tests, Machine learning." 2021 Second International Conference on Smart Technologies in Computing,
Electrical and Electronics (ICSTCEE). IEEE, 2021
[19] Priyadarshini, M. Asha, et al. "A Visionary Approach to Anemia Detection: Integrating Eye Condition Data and
Machine Learning." ICCIET-2024. Atlantis Press, 2024. https://doi.org/10.2991/978-94-6463-471-6_75 .
162

Leveraging Flask API and Machine Learning to Forecast Multiple Diseases

Uploaded by

Leveraging Flask API and Machine Learning to Forecast Multiple Diseases

Uploaded by

Communications on Applied Nonlinear Analysis

Leveraging Flask API and Machine Learning to Forecast Multiple

K Venkateswara Rao1,D Srilatha2,Sridevi Sakhamuri3,Venkata Subbaiah Desanamukula4,M

Article History: Abstract:

Fig:1 Proposed System Architecture

Fig:2 Working Model of the proposed system

LIVER DISEASE DATASET:

MALARIA DISEASE DATASET:

PNEUMONIA DISEASE DATASET:

INTERFACE FOR KIDNEY DISEASE PREDICTION:

RESULT FOR KIDNEY DISEASE PREDICTION:

INTERFACE FOR LIVER DISEASE PREDICTION:

RESULT FOR LIVER DISEASE PREDICTION:

INTERFACE FOR MALARIA DISEASE PREDICTION:

RESULT AFTER UPLOADING THE RBC PICTURE:

INTERFACE FOR PNEUMONIA DISEASE PREDICTOR:

DISPLAYING THE STATUS OF PATIENT AFTER UPLOADING CHEST X-RAY

DISEASE BEST ALGORITHM DETECTED RESULTED ACCURACY

You might also like