TY  - JOUR
T1  - A Framework for Accurate Disease Diagnosis using Cover Data Mining Rule on Homogenous
Data
AU - Gayathri, K. AU - Chitra, M. 
JO  - Research Journal of Agronomy
VL  - 12
IS  - 2
SP  - 8
EP  - 16
PY  - 2018
DA  - 2001/08/19
SN  - 1815-9354
DO  - rjagr.2018.8.16
UR  - https://makhillpublications.co/view-article.php?doi=rjagr.2018.8.16
KW  - Sequential class covering
KW  -disease diagnosis
KW  -first order rule
KW  -homogenous data classifier
KW  -boosting
KW  -dimensionality
KW  -machine learning
AB  - Knowledgeable data is the fundamental step for
discovering different types of patterns from large
database. The pattern to be discovered from vast amount
of data employs classification technique. Classification
(i.e., classifier) builds a model with the relationship
between the attribute set, class set and input data.
However, most of the classification techniques do not fit
with a good starting point on classifying multiple data
sources class patterns. Even if it works on multiple data
sources class patterns, it produces both the best and worst
cases of result set. On occurrence of worst case result,
patterns are not nested properly resulting in the tradeoff
while fetching high class accuracy result. These
drawbacks in the current work are overcome in our
research work by working with sample of large quantities
of information about patients and their medical
conditions. In this research, an efficient framework for
accurate disease diagnosis, Sequential Class Covering
Rule based Homogeneous Data Classifier (SCCR-HDC)
is proposed. Initially, SCCR-HDC framework uses the
classifier tree to analyze medical information about
patients from different dimensional level. For analyzing
this classifier tree, a modern boosting based machine
learning concept is introduced. The analyzed results of the
tree are used for rule formation in the second step for
efficient diagnosis of the disease patterns. The rule
formed is applied on the training and test sample
homogenous data to easily diagnosis the disease class
accuracy. A sequential class covering rule is formed to
extract the best result patterns in sequential manner
from the current set of training data instances. Similarly,
to diagnosis the normal, abnormal, critical disease
patterns from the test samples, a searching process
called, first order rule based general to precise
searching process is performed in SCCR-HDC
framework. Experiment is conducted on the factors
such as class accuracy rate on disease diagnosis,
classifier tree based time rate on predicting disease pattern
and precision rate on categorizing disease patterns.
ER  -