A carregar...
Publicação
Data Mining Application for Healthcare Sector: Predictive Analysis of Heart Attacks
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 4.04 MB | Adobe PDF |
Orientador(es)
Resumo(s)
Cardiovascular diseases are the main cause of the number of deaths in the world, being the heart
disease the most killing one affecting more than 75% of individuals living in countries of low and middle
earnings. Considering all the consequences, firstly for the individual’s health, but also for the health
system and the cost of healthcare (for instance, treatments and medication), specifically for
cardiovascular diseases treatment, it has become extremely important the provision of quality services
by making use of preventive medicine, whose focus is identifying the disease risk, and then, applying
the right action in case of early signs. Therefore, by resorting to DM (Data Mining) and its techniques,
there is the ability to uncover patterns and relationships amongst the objects in healthcare data, giving
the potential to use it more efficiently, and to produce business intelligence and extract knowledge
that will be crucial for future answers about possible diseases and treatments on patients. Nowadays,
the concept of DM is already applied in medical information systems for clinical purposes such as
diagnosis and treatments, that by making use of predictive models can diagnose some group of
diseases, in this case, heart attacks.
The focus of this project consists on applying machine learning techniques to develop a predictive
model based on a real dataset, in order to detect through the analysis of patient’s data whether a
person can have a heart attack or not. At the end, the best model is found by comparing the different
algorithms used and assessing its results, and then, selecting the one with the best measures.
The correct identification of early cardiovascular problems signs through the analysis of patient data
can lead to the possible prevention of heart attacks, to the consequent reduction of complications and
secondary effects that the disease may bring, and most importantly, to the decrease on the number
of deaths in the future. Making use of Data Mining and analytics in healthcare will allow the analysis
of high volumes of data, the development of new predictive models, and the understanding of the
factors and variables that have the most influence and contribution for this disease, which people
should pay attention. Hence, this practical approach is an example of how predictive analytics can have
an important impact in the healthcare sector: through the collection of patient’s data, models learn
from it so that in the future they can predict new unknown cases of heart attacks with better
accuracies. In this way, it contributes to the creation of new models, to the tracking of patient’s health
data, to the improvement of medical decisions, to efficient and faster responses, and to the wellbeing
of the population that can be improved if diseases like this can be predicted and avoided. To conclude, this project aims to present and show how Data Mining techniques are applied in
healthcare and medicine, and how they contribute for the better knowledge of cardiovascular diseases
and for the support of important decisions that will influence the patient’s quality of life.
Descrição
Project Work presented as the partial requirement for obtaining a Master's degree in Information Management, specialization in Knowledge Management and Business Intelligence
Palavras-chave
Cardiovascular Disease Heart Attack Data Mining Machine Learning Predictive Analysis Neural Networks Logistic Regression Decision Trees SDG 3 - Good health and well-being