The development of malign cells that can grow in any part of the stomach, known as gastric cancer, is one of the most common causes of death worldwide. In order to increase the survival rate in patients with this condition, it is essential to improve the decision-making process leading to a better and more efficient selection of treatment strategies. Nowadays, with the large amount of information present in hospital institutions, it is possible to use data mining algorithms to improve the healthcare delivery. Thus, this study, using the CRISP methodology, aims to predict not only the mortality associated with this disease, but also the occurrence of any complication following surgery. A set of classification models were tested and compared in order to improve the prediction accuracy. The study showed that, on one hand, the J48 algorithm using oversampling is the best technique to predict the mortality in gastric cancer patients, with an accuracy of approximately 74%. On the other hand, the rain forest algorithm using oversampling presents the best results when predicting the possible occurrence of complications among gastric cancer patients after their in-hospital stays, with an accuracy of approximately 83%.

Strokes are neurological events that affect a certain area of the brain. Since brain controls fundamental body activities, brain cell deterioration and dead can lead to serious disabilities and poor life quality. This makes strokes the leading cause of disabilities and mortality worldwide. Patients that suffer strokes are hospitalized in order to be submitted to surgery and receive recovery therapies. Thus, it's important to predict the length of stay for these patients, since it can be costly to them and their family, as well as to the medical institutions. The aim of this study is to make a prediction on the number of days of patients' hospital stays based on information available about the neurological event that happened, the patient's health status and surgery details. A neural network was put to test with three attribute subsets with different sizes. The best result was obtained with the subset with fewer features obtaining a RMSE and a MAE of 5.9451 and 4.6354, respectively.

Gastric cancer is one of the most prevalent types of cancer in the whole world, affecting millions of people over the last decades. Its symptoms are ambiguous, which leads to late diagnoses, reducing the patients' chances of survival. In most countries, routine screenings are not usual, which also contributes to the detection of this gastric malignancy in later and more dangerous (and often fatal) stages. One of the main focus of improving healthcare services related to gastric cancer relies on increasing the survival rates. This and predicting if a patient will suffer from any complication following the surgery can aid the healthcare professionals in selecting better and more efficient treatment strategies. Thus, this constitutes as the aims of this study which will test and compare a set of classification models in order to improve the prediction accuracy. Data mining techniques will be put into use, since it's been proved they are one of the best ways of producing useful information for many businesses, including healthcare.