International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
MACHINE LEARNING APPROACHES TO MULTI-DISEASE PROGANASTICATION 1G K Karthik, 2S Sahishnu Nag, 3K Jwalitha 1,2,3 Student, GITAM (Deemed to be University), Visakhapatnam, Andhra Pradesh, India.
---------------------------------------------------------------------***---------------------------------------------------------------------
1. Abstract ‐
customers with accurate and real-time disease predictions, along with corresponding information about the associated symptoms.
Progress in machine learning in the biomedical and healthcare domains has made accurate analysis of medical data better for early sickness detection, patient treatment, and community services. Insufficient quality of medical data results in a decline in the study's accuracy. Furthermore, distinct geographical areas display distinct manifestations of certain localised illnesses, thus undermining the forecasting of disease epidemics. The suggested approach offers machine learning algorithms for accurate forecasting of different illness incidences in communities where diseases are common. It tests the modified estimate models using actual hospital data that has been gathered. It uses a latent factor model to reconstruct the missing data in order to get over the challenge of incomplete data. It tests a localised chronic form of cerebral infarction. It makes use of both structured and unstructured hospital data. By mining data sets for conditions like diabetes, breast cancer, and heart disease, it makes predictions about likely ailments. To the best of our knowledge, no previous effort in the field of medical big data analytics has taken into account both forms of data. Our suggested technique outperforms numerous common estimate algorithms in terms of calculation accuracy, reaching 94.8%, and has a faster rate of convergence than the machine learning disease risk prediction algorithm.
To address this challenge, we are introducing a Djangobased system designed for predicting specific medical conditions. In our initial implementation, we will focus on analyzing malaria, heart disease, and diabetes. However, it's important to note that we have the flexibility to expand the range of diseases in the future. Our approach combines Django, a robust web framework, with machine learning techniques to develop multiple disease prediction models. One notable advantage of our system is its comprehensive consideration of all relevant factors contributing to each disease during the analysis, which enables more precise and effective disease identification. To ensure the preservation of the model's behavior, we utilize Python pickling, allowing us to save and load the trained model as a pickle file in Python. This approach ensures that our system maintains its predictive accuracy and functionality. Many existing healthcare systems have been primarily designed to assess individual diseases separately. For instance, one system might be dedicated to diabetes, another to diabetic retinopathy, and yet another to heart disease prediction. This fragmented approach often requires organizations to deploy a variety of models to evaluate patient health information. These systems are tailored for analyzing specific ailments in isolation.
Key Words: Heart Disease prediction, Diabetes, Breast Cancer.
In contrast, our proposed system offers a more versatile approach. Users of our multi-disease prediction system can conveniently assess several diseases on a single webpage. They no longer need to navigate multiple platforms to determine whether they might be affected by a particular illness. To utilize this comprehensive disease prediction system, users simply need to select the name of the disease of interest, input the relevant parameters, and click the "submit" button. The system will then invoke the appropriate machine learning model to forecast the outcome and display the results on the screen. This integrated approach streamlines the process for users, providing a more user-friendly and efficient experience.
2. Introduction In today's digital age, data has become a valuable asset, with vast amounts being generated across various industries. The healthcare sector, in particular, relies on patient-related information, collectively referred to as healthcare data. In this context, we propose a general architecture for predicting illnesses in the healthcare sector. Many existing models tend to focus on analyzing one disease at a time, such as diabetes, cancer, or skin conditions. However, there is a noticeable absence of a comprehensive system that can simultaneously assess multiple diseases. Our primary objective is to provide
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 625