Page 410 - Kaleidoscope Academic Conference Proceedings 2024

2024 ITU Kaleidoscope Academic Conference

Applying several machine learning algorithms to classify ventricular activity of the heart with the help of polarization
the abnormal and normal ECG signals that are shown in and depolarization. The early detection of the symptoms
table 3 below. related to A.F. and the prognosis of A.F. activity both need
the use of ML and A.I. techniques. Further identifying
Table 3: Comparison Matrix for ML Algorithms abnormality and correctly classifying the abnormality are
the primary concerns related to capturing ECG signal
SL. Algorithm Accuracy analysis. Below, figure 8 shows the ECG signal
No abnormality detection graph for a better understanding of
the abnormality [27].
1 Naive Bayes 0.6975690152451587

2 SVN 0.8030490317264112
3 Decision Tree 0.8908117016893284

Random
4 Forest 0.9546765554182118
5 KNN 0.6464580152451587

From the above table, it is observed that the random forest
algorithm gives good accuracy as compared to other
machine learning algorithms.

Discussions: The acquisition of polarization working of the
heart is captured by the use of an ECG signal. This paper Figure 8: Abnormality Detection
has tried to find out various health hazards related with the
heart for its irregular functioning. As heart is a major organ CONCLUSION
of our body, any mal functioning of our heart can lead to
mortality. So proper treatment of our heart needs advanced Science and technologies is making a great advancement in
technologies for proper diagnosis and for quick and healthy every sphere of life including the medical field. Now a
recovery. Polarization and depolarization of the heart are day’s health hazards are also increasing at an alarming rate,
essential as it is a continuous process that is responsible for especially heart related diseases. So, to cope up with the
the contraction and de-contraction of the heart for the critical and complicated heart related treatments proper
pumping of blood among several parts of the body. Without utilization of advanced technologies becomes necessary.
polarization and depolarization, it never happened. One Further Random Forest algorithm approach can perform
cycle of an ECG consists of patterns like P, Q, R and S better as compared to other classification techniques.
signals or waves. Examining the chances of several A.I. Diagnosis of heart diseases from a tracing of ECG signals is
techniques for interpreting ECG signals regarding the complex for clinical physicians working at respective
diagnosis of a heart condition is a common activity. The levels. These difficulties give an opportunity for the
difficulty of the ML algorithm model's interpretability has involvement of the ML techniques to analyze the ECG
hindered doctors from having confidence in the diagnosis signals more deeply for better prediction by extracting
results of ML models. Each recording undergoes a important features. The black box nature of these ML
preprocessing stage in order to extract the important feature algorithms and their respective performances helps to detect
vectors from the ECG signals. From 0 Hz range to 85 Hz, abnormality of the heart rhythm more precisely. Random
the power spectrum density and HRV–based characteristics forest algorithm creates different decision trees that help us
are in consideration. The preprocessing activity generates a to find more clarification of heart rhythms. The available
feature vector of several dimensions. Purifying the signal heart ECG data set is preprocessed to remove noise level,
by avoiding the noise level is an important challenge. As then application of several ML methods, their results and
the noise level increases, the performance decreases. complexity can easily choose the best methods. In
Arterial filtration (A.F.) is an atypical electrical activity that conclusion, the respective results achieved help us better
helps to detect the abnormal activity of the heart since diagnose the patients with good physicians. Further new
chambers of the heart or atria cannot pump blood normally. approaches should be discovered, and existing methods
Atrial tachycardia is responsible for the characterization of need to be formalized to achieve physicians ' innovation
the heart rate that is excesses of 100 beats per minute, behind the use of the ML model's decision approach.
which is the presence of an abnormality.
REFERENCES
Moreover, early detection is essential for avoiding fatality.
Predicting mortality demands analyzing the polarization [1] Wu, H., Patel, K.H.K., Li, X. et al. A fully-
and depolarization activity of the cardiovascular system is automated paper ECG digitization algorithm using
essential. The ECG signals help to show the atrial and deep learning. Sci Rep 12, 20963 (2022). Received

– 366 –