Application of Data Mining Techniques for Medical Data Classification

This report reviews the journal called ‘Application of Data Mining Techniques for Medical Data Classification’ which primarily focus on applying the various data mining techniques to healthcare sectors with the intention to classify medical data.

A4

Table of Contents
MAIN BODY..................................................................................................................................3
REFERENCES................................................................................................................................7

MAIN BODY
This report reviews the journal called ‘Application of Data Mining Techniques for
Medical Data Classification’ which primarily focus on applying the various data mining
techniques to healthcare sectors with the intention to classify medical data. This application of
data mining algorithms to healthcare sector is immensely beneficial to improving and enhancing
various operations of the healthcare industries such as the detection of healthcare cases which are
fraudulent or related to abuse or harassment, decision making of the healthcare professionals,
management of patients, etc., as these would take exorbitant amount of time duration if done
through traditional means. The application of data mining techniques to the above mentioned
operations of the healthcare industries is also immensely essential to the growth and
development of healthcare professionals and industries in general as these operations require
professional healthcare expert’s intuition and experiences and even then operational prejudices,
biases and errors can seep through affecting the quality of healthcare that is provided to patients
(Högg and et.al., 2018). Data mining effectively improves the decision making operations of
healthcare professionals as data mining algorithms are able to discover trends and patterns even
amongst data that is humongous in size and complexity which is impossible when done through
traditional means of pattern recognition. The results which are generated by data mining
algorithms can effective increase the productivity, profitability, performance and efficiency of
health care institutions in the industries and allow them to effectively provide relevant and
adequate healthcare to all patients, thus benefiting the society as well.
Based on the journal, it is evident that though data mining algorithms and techniques
have immense potential for the growth and development of healthcare industries, there are still
present various issues for the effective implementation of data mining techniques to operations
of healthcare. The medical data gathered by healthcare institutions is immensely complex and
large in size, which results in failure of traditional systems and software in trying to find patterns
inside, which is also aided by the fact that a large proportion of data is unstructured and highly
complex such as handwritten doctor’s notes, MRI scans, CT scans, radiographic films etc. which
are hard for traditional systems to evaluate. There is a need for systems capable of managing and
evaluating large, unstructured, complex datasets with price and storage restrictions placed by

healthcare institutions considered. This can be done through the categorisation data mining
algorithms for application to clinical imaging. There are primarily two ways through which data
mining algorithms are categorised, supervised or predictive algorithms and unsupervised or
descriptive machine learning algorithms. Predictive mining algorithms make a prediction on the
data set provided, while descriptive mining algorithms provide descriptions and characteristic of
the data that is input into the systems (Pika and et.al., 2019). There are three learning approaches
to data mining algorithms namely, supervised, unsupervised and semi supervised all of which are
used depending on the needs and requirements of the healthcare professional and institution’s
data mining operations. The primary advantage of using data mining algorithms in healthcare is
because they offer increased user interactivity while learning and predicting future needs and
operations on the basis of past operations and they can be updated to learn with the changing
trends of the healthcare industries.
Though the most useful application of data mining algorithms to healthcare operations is
done via the classification mining algorithms, there is still a need for data mining technique
which is flexible enough to be applied to diverse and varied healthcare operations and processes.
The journal effectively provides us with a tabular representation of the various data mining
techniques currently used in healthcare industries for data mining operations with the intention to
extract usable medical information from the huge datasets. As the healthcare operations are quite
diverse form one another, the application of same data mining algorithm to different operations
yields negative results. The journal also provides us with the benefits and drawbacks of using
various different data mining techniques to healthcare operations so that we might get a better
understanding of the issues of using similar data mining techniques for diverse healthcare
operations (Nagarajan and Kumar, 2018). Various different data mining techniques and decision
support systems have been created by professionals in order to effectively apply data mining to
healthcare operations especially to the detection of lung, heart and breast cancers and diseases.
The journal also provides us with a tabular representation of all the medical issues which have
been solved through classification data mining techniques, along with the data sets the technique
requires and results of the application of applying various data mining techniques to distinct
healthcare operational processes. From the journal, it is evident that there still hasn’t been
developed a singular data mining techniques or classifier which can provide best results for
varying data sets and healthcare operations. The performance of each classifier is assesses by

segmenting the dataset into two sections, testing and training (Chowdhury, Islam and Khan,
2017). This allows for researchers to assess the performance of classifiers by providing it with
test datasets, which if product adequate and relevant results can be trained further. Cross
validation technique is used to further improve the accuracy of these classifiers for both testing
and training datasets.
Despite the growth and development of the application of data mining techniques to the
operations and processes of healthcare institutions and professionals, there hasn’t still been
developed a singular technique which can be extensively used to classify various data sets of
medical institutions pertaining to different operational processes. The primary reason for this is
because there is need for immensely precise data which can be categorised, while providing a
very low rate of false negatives, as the health and safety of patients of healthcare cannot be
jeopardised by faulty results of data mining algorithms. However, there are specialised data
mining techniques which can be applied to diverse datasets, provided accurate prior knowledge
of the healthcare operations and data is provided (Domadiya and Rao, 2019). The main areas in
which the medical data has been improved over the past decades are the innovation and use of
advanced data mining algorithms and techniques in healthcare scenarios, the use of multiple
features of algorithms and the seamless integration of medical datasets into classification
procedures which are used by data mining algorithms. However, despite these innovations and
developments, there is are still multiple challenges to application of data mining algorithms to
medical operations such as the performance and scalability of data mining techniques and
algorithms to suit the diverse healthcare operations and complex data sets. There still remains
great potential for the future applications of data mining techniques and algorithms to the various
healthcare operations and diverse datasets.
There are still many approaches and techniques of data mining operations that can be
applied to healthcare operations to increase their performance, efficiency, productivity so that
better healthcare can be provided to patients in a timely manner. There is still ample room for
future data mining algorithms to improve the performance, privacy, interactivity and timeliness
of data mining algorithms that are applied to the healthcare operations of the industries. It is
evident from the journal that data mining techniques and algorithms cannot revolutionise the way
healthcare is provided in the digital age, yet the benefits that data mining algorithms and

techniques provide are reasons enough to continue to develop better techniques that can mine
data for the various healthcare operations and large, complex data sets through big data analytics
techniques and data mining algorithms (Ukil and et.al., 2016). Developers of the future data
mining techniques need to consider the storage, computational and economical restrictions that
healthcare institutions have towards adopting data mining techniques so that they can be
universally applied to various healthcare operations.

REFERENCES
Books and Journals
Chowdhury, M.H., Islam, M.K. and Khan, S.I., 2017, December. Imputation of missing
healthcare data. In 2017 20th International Conference of Computer and Information
Technology (ICCIT) (pp. 1-6). IEEE.
Domadiya, N. and Rao, U.P., 2019. Privacy preserving distributed association rule mining
approach on vertically partitioned healthcare data. Procedia computer science. 148.
pp.303-312.
Högg, T. and et.al., 2018. Mining healthcare data for markers of the multiple sclerosis prodrome.
Multiple sclerosis and related disorders. 25. pp.232-240.
Nagarajan, V.R. and Kumar, V., 2018. An optimized sub group partition based healthcare data
mining in big data. Int J Innov Res Sci Technol. 4(10). pp.79-85.
Pika, A. and et.al., 2019, September. Towards privacy-preserving process mining in healthcare.
In International Conference on Business Process Management (pp. 483-495). Springer,
Cham.
Ukil, A. and et.al., 2016, March. IoT healthcare analytics: The importance of anomaly detection.
In 2016 IEEE 30th International Conference on Advanced Information Networking and
Applications (AINA) (pp. 994-997). IEEE.