Chronic Renal Disease Prediction (CRD).pptx

An approach for prediction CRD is proposed. Kidneys are the prominent organs which help in removal of waste and toxic material from the body.
Kidney malfunctioning occurs due to various reasons but if certain symptoms are ignored and not treated on time, then it may lead to persistent malfunctioning leading to Chronic Renal Disease (CRD). It is a progressive and long-term condition that affects millions of people worldwide.
This work depicts the appropriate, relevant and correlated attributes among all the attributes and reduction of features in the dataset using Chi-squared test on to the patients’ dataset for better detection and prediction of CRD.
The CRDP algorithm is implemented, and the results are predominantly used in Logistic Regression and K-Nearest Neighbour classification techniques to enhance and improve their prediction accuracy on CRD.
The system architecture diagram shows how the dataset considered from University of California, Irvine for analysis is preprocessed to reduce the appropriate and relevant correlated attributes using chi-square testing.
10 attributes are selected and reduced from 24 attributes which are relevant and significant for further testing.
70% of data is trained and the remaining 30% of data is used for testing the Logistic Regression and K-Nearest Neighbour and the Random Forest Machine Learning techniques for evaluation and their performance in predicting the CRD.
From the results obtained, it is observed that the accuracy obtained for predicting CRD through KNN is improved and better than Logistic Regression after reduction of attributes using Chi-squared test.
These results of proposed algorithm are better when compared to that of Random forest algorithm accuracy which is 97.12%, while ANN accuracy is 94.5% [9].
CONCLUSIONS:
The main objective of this work, i.e to accurately identify CRD utilizing a number of tests and indicators is achieved. In this work, 10 most relevant out of 24 attributes are taken into consideration, which are used to conduct testing and produce reliable findings.
The results revealed that the K-Nearest Neighbour classifier achieved an improved accuracy of 99.17% after the reduction of attributes using Chi-squared test when compared to 98.33% for Logistic Regression.
The work can be extended to compute the detection and prediction accuracy for various other machine learning algorithms and the best can be chosen.
The research work is presented at International Conference on
Evolutionary Artificial Intelligence (ICEAI 2023).
The CRDP algorithm is implemented, and the results are predominantly used in Logistic Regression and K-Nearest Neighbour classification techniques to enhance and improve their prediction accuracy on CRD.
This work depicts the appropriate, relevant and correlated attributes among all the attributes and reduction of features in the dataset using Chi-squared test on to the patients’ dataset for better detection and prediction of CRD.
Kidney malfunctioning is

1. International Conference on Evolutionary Artificial Intelligence (ICEAI 2023) 13/09/2023 ICEAI-028 1 INTRODUCTION: Kidneys are the prominent organs which help in removal of waste and toxic material from the body. Kidney malfunctioning occurs due to various reasons but if certain symptoms are ignored and not treated on time, then it may lead to persistent malfunctioning leading to Chronic Renal Disease (CRD). It is a progressive and long-term condition that affects millions of people worldwide. This work depicts the appropriate, relevant and correlated attributes among all the attributes and reduction of features in the dataset using Chi-squared test on to the patients’ dataset for better detection and prediction of CRD. The CRDP algorithm is implemented, and the results are predominantly used in Logistic Regression and K-Nearest Neighbour classification techniques to enhance and improve their prediction accuracy on CRD.

2. International Conference on Evolutionary Artificial Intelligence (ICEAI 2023) 13/09/2023 ICEAI-028 2 SYSTEM ARCHITECTURE: The system architecture diagram shows how the dataset considered from University of California, Irvine for analysis is preprocessed to reduce the appropriate and relevant correlated attributes using chi-square testing. 10 attributes are selected and reduced from 24 attributes which are relevant and significant for further testing. 70% of data is trained and the remaining 30% of data is used for testing the Logistic Regression and K-Nearest Neighbour and the Random Forest Machine Learning techniques for evaluation and their performance in predicting the CRD.

3. International Conference on Evolutionary Artificial Intelligence (ICEAI 2023) 13/09/2023 ICEAI-028 3 RESULTS & ANALYSIS: From the results obtained, it is observed that the accuracy obtained for predicting CRD through KNN is improved and better than Logistic Regression after reduction of attributes using Chi-squared test. These results of proposed algorithm are better when compared to that of Random forest algorithm accuracy which is 97.12%, while ANN accuracy is 94.5% [9]. Evaluation Parameters Logistic Regression K-Nearest Neighbour Accuracy 98.33 % 99.17 % Precision 0.9833 0.9778 Recall 0.9833 1.0000

4. International Conference on Evolutionary Artificial Intelligence (ICEAI 2023) 13/09/2023 ICEAI-028 4 CONCLUSIONS: The main objective of this work, i.e to accurately identify CRD utilizing a number of tests and indicators is achieved. In this work, 10 most relevant out of 24 attributes are taken into consideration, which are used to conduct testing and produce reliable findings. The results revealed that the K-Nearest Neighbour classifier achieved an improved accuracy of 99.17% after the reduction of attributes using Chi-squared test when compared to 98.33% for Logistic Regression. The work can be extended to compute the detection and prediction accuracy for various other machine learning algorithms and the best can be chosen.