A Machine Learning Methodology for Diagnosing Chronic Kidney Disease

Abstract

Chronic kidney disease (CKD) is a global health problem with high morbidity and mortality rate, and it induces other diseases. Since there are no obvious symptoms during the early stages of CKD, patients often fail to notice the disease. Early detection of CKD enables patients to receive timely treatment to ameliorate the progression of this disease. Machine learning models can effectively aid clinicians achieve this goal due to their fast and accurate recognition performance. In this study, we propose a machine learning methodology for diagnosing CKD. The CKD data set was obtained from the University of California Irvine (UCI) machine learning repository, which has a large number of missing values. KNN imputation was used to fill in the missing values, which selects several complete samples with the most similar measurements to process the missing data for each incomplete sample. Missing values are usually seen in real-life medical situations because patients may miss some measurements for various reasons. After effectively filling out the incomplete data set, six machine learning algorithms (logistic regression, random forest, support vector machine, k-nearest neighbor, naive Bayes classifier and feed forward neural network) were used to establish models. Among these machine learning models, random forest achieved the best performance with 99.75% diagnosis accuracy. By analyzing the misjudgments generated by the established models, we proposed an integrated model that combines logistic regression and random forest by using perceptron, which could achieve an average accuracy of 99.83% after ten times of simulation. Hence, we speculated that this methodology could be applicable to more complicated clinical data for disease diagnosis. INDEX TERMS Chronic kidney disease, machine learning, KNN imputation, integrated model.

Ensemble Feature Ranking for Cost-Based Non-Overlapping Groups: A Case Study of Chronic Kidney Disease Diagnosis in Developing Countries

“…The resultant model achieved a slightly higher predictive accuracy than that of the individual models of RF and LG. It can be observed that decision treebased ensemble modeling approaches consistently produce models with high predictive accuracy for the CKD problem as reported in [10], [12].…”

“…In this regard, the selected problem is of high interest as it blights the developing countries over a long period. As machine learning-based screening and diagnosis applications provide a workable solution to the many problems faced by the healthcare sector, therefore, recently a lot of scholarly work is reported on developing disease diagnosis and management systems for the CKD [9], [10], [12], [13], [21], [22]. The current study is in the continuation of the scholarly work performed on the CKD diagnosis problem.…”

“…40 cases. Qin et al, [12] investigated the effects of missing values substitution for the CKD diagnosis problem. In this regard, the authors have taken into account a number of classifiers such as logistic regression (LG), Random Forest (RF), SVM, k-nearest neighbor (KNN), Naïve Bayes (NB), and ANN.…”

“…The proposed feature selection technique is simple to implement and would incur less computational time A large feature set is selected for the CKD problem than other comparative techniques for the same dataset [10] Based on a set-theoretic heuristic a set of features are selected and subsequently, an XGBoost model is built XGBoost classifier is demonstrated to produce a highly accurate model based on the selected features A large feature set is selected for the CKD problem than other comparative techniques for the same dataset [12] LG [28] proposed an ensemble-based feature selection technique. The proposed technique is based on a set of filter methods such as information gain, gain ratio, chi-square, and ReliefF to comprehensively assign a score to the features.…”

“…Statistical Difference Chen et al, [29] * Ali et al, [30] * Salekin et al, [22] * Polat et al, [13] * Almansour et al, [21] * Ogunley et al, [10] † Qin et al, [12] * Osanaiye et al, [28] * * Proposed solution is not statistically better † Proposed solution is statistically better solution-2 but the overall cost is also low. It is important to note that all the results reported in Table 18 are based on the CKD diagnosis dataset.…”