Predicting Nosocomial Infection by Using Data Mining Technologies - New Contributions in Information Systems and Technologies - page 177

during this period. Data Mining (DM) technologies can be used to create predictive
models and apply them to other patient data in order to make predictions.
This work is also related to the healthcare professional's need to take fast, reasoned
and accurate decisions to improve the efficiency as well as the productivity of the
healthcare organization and, consequently, the quality of the delivered care. Through
prediction this work is capable to help healthcare professionals in the nosocomial
infection related decision-making process.
Therefore, the main goal is to study the applicability of DM techniques to perform
clinical predictions related to the occurrence of nosocomial infections in the Medicine
Units of Centro Hospitalar do Porto (CHP). To achieve this goal the design of
classification DM models was needed. These models can be used to identify the
possible occurrence of nosocomial infections, allowing healthcare professionals to
plan and implement specific and efficient infection prevention measures. This way, it
is possible to prevent these infections and, thus, to improve the patients' safety and
well-being and, consequently, the healthcare quality.
Besides the introduction, this article includes four more sections. The first is
related to the background and provides an overview of Knowledge Discovery in
Databases and DM. It also refers to the implementation of DM in healthcare. The
second section presents the DM study for nosocomial infection prediction performed
in this work and its main results. The third section suggests some future work
measures and the last section presents the main conclusions of the work.

2 Background

2.1 Knowledge Discovery in Databases and Data Mining

Data Mining (DM) refers to the process of finding unknown patterns in huge amounts
of complex data and use that information to build predictive models capable to help
decision-making [4].
According to Fayyad [5], DM is just one stage of the Knowledge Discovery in
Databases (KDD) process. The KDD process refers to the discovery of useful
knowledge on data, while DM refers only to the analysis of data and application of
algorithms to extract patterns from them. The remaining four steps of the KDD
process, such as data selection, data pre-processing, data transformation and the
interpretation and evaluation of the DM stage results, are essential to ensure the
extraction of useful knowledge from data [5]. Data selection consists of choosing
the data useful to solve the problem in study. The pre-processing stage is responsible
for cleaning the data in order to make them consistent. The transformation stage
manipulates data in order to make them suitable for the DM models. The last stage of
the KDD is the interpretation and evaluation of the mined patterns and their
application in the decision-making process [5] [6].
The quality of the DM stage results is strictly related with the quality of the data
used in the DM stage, therefore, the quality of the extracted knowledge depends on
the quality of the data used and that is a crucial factor to take good decisions [4].