When machine learning is used for the design of a prediction model in medical science, then higher accuracy is essential. It becomes difficult to achieve higher accuracy due to unavailability of values in certain fields of data set. Therefore, it is necessary to deal with the issue of missing values effectively. This research work focuses on an efficient way to handle missing values. Authors have proposed a systematic methodology for the identification of missing value. Authors have used Cleveland Heart disease dataset from the UCI (University of California, Irvine) repository to test their experiments. Missing values are imparted using three different approaches, namely random, MISSHASH & MISSFIB. Four imputation methods k-nearest neighbor (KNN), multivariate imputation by chained equations (MICE), mean, and mode imputation were analyzed with the help of four classifiers Naive Bayes (NB), support vector machine (SVM), logistic regression (LR), and random forest (RF). Root mean square error (RMSE) of classifiers was compared to find the combination of the best imputation method. It has found that MICE imputation method has performed better related to other imputation methods. Moreover, its accuracy is independent of classifier and missing value distribution.

Multistage Model for Accurate Prediction of Missing Values Using Imputation Methods in Heart Disease Dataset

To introduce the Hemorrhage Intensive Severity and Survivability (HISS) score, based on the fusion of multi-biomarker data; glucose, lactate, pH, potassium, and oxygen tension, to serve as a patient-specific attribute in hemorrhagic trauma. One hundred instances of Sensible Fictitious Rationalized Patient (SFRP) data were synthetically generated and the HISS score assigned by five clinically active physician experts (100 [5]). The HISS score stratifies the criticality of the trauma patient as; low(0), guarded(1), elevated(2), high(3) and severe(4). Standard classifier algorithms; linear support vector machine (SVM-L), multi-class ensemble bagged decision tree (EBDT), artificial neural network with bayesian regularization (ANN:BR) and possibility rule-based using function approximation (PRBF) were evaluated for their potential to similarly classify and predict a HISS score. SVM-L, EBDT, ANN:BR and PRBF generated score predictions with testing accuracies (majority vote) corresponding to 0.91 ± 0.06, 0.93 ± 0.04, 0.92 ± 0.07, and 0.92 ± 0.03, respectively, with no statistically significant difference (p > 0.05). Targeted accuracies of 0.99 and 0.999 could be achieved with SFRP data size and clinical expert scores of 147[7](0.99) and 154[9](0.999), respectively. The predictions of the data-driven model in conjunction with an adjunct multi-analyte biosensor intended for point-of-care continual monitoring of trauma patients, can aid in patient stratification and triage decision-making.

/pdf/toward-a-hemorrhagic-trauma-severity-score-fusing-five-e7iq6eipf0.pdf

Toward a hemorrhagic trauma severity score: fusing five physiological biomarkers

Decision support systems developed using machine learning classifiers have become a valuable tool in predicting various diseases. However, the performance of these systems is adversely affected by the missing values in medical datasets. Imputation methods are used to predict these missing values. In this paper, a new imputation method called hybrid imputation optimized by the classifier (HIOC) is proposed to predict missing values efficiently.,The proposed HIOC is developed by using a classifier to combine multivariate imputation by chained equations (MICE), K nearest neighbor (KNN), mean and mode imputation methods in an optimum way. Performance of HIOC has been compared to MICE, KNN, and mean and mode methods. Four classifiers support vector machine (SVM), naive Bayes (NB), random forest (RF) and decision tree (DT) have been used to evaluate the performance of imputation methods.,The results show that HIOC performed efficiently even with a high rate of missing values. It had reduced root mean square error (RMSE) up to 17.32% in the heart disease dataset and 34.73% in the breast cancer dataset. Correct prediction of missing values improved the accuracy of the classifiers in predicting diseases. It increased classification accuracy up to 18.61% in the heart disease dataset and 6.20% in the breast cancer dataset.,The proposed HIOC is a new hybrid imputation method that can efficiently predict missing values in any medical dataset.

HIOC: a hybrid imputation method to predict missing values in medical datasets

Data usually comes into data warehouses from multiple sources having different formats and are specifically categorized into three groups (ie, structured, semi-structured, and unstructured) Various data mining technologies are used to collect, refine, and analyze the data which further leads to the problem of data quality management Data purgation occurs when the data is subject to ETL methodology in order to maintain and improve the data quality The data may contain unnecessary information and may have inappropriate symbols which can be defined as dummy values, cryptic values, or missing values The present work has improved the expectation-maximization algorithm with dot product to handle cryptic data, DBSCAN method with Gower metrics to ensure dummy values, Wards algorithm with Minkowski distance to improve the results of contradicting data and K-means algorithm along with Euclidean distance metrics to handle missing values in a dataset These distance metrics have improved the data quality and also helped in providing consistent data to be loaded into a data warehouse

Enhancing Data Quality at ETL Stage of Data Warehousing

The performance of machine learning-based decision support systems for predicting any disease is adversely affected by missing values in the dataset. Prediction of these missing values in an accurate manner is a critical issue. In this paper, a new imputation method called hybrid imputation optimized by classifier (HIOC) is proposed to predict missing values in the dataset. The proposed HIOC is developed by using a classifier to combine multivariate imputation by chained equations (MICE), K-nearest neighbor (KNN), mean and mode imputation methods in an optimum way. Proposed HIOC has been validated on the datasets of heart disease and breast cancer. Performance of HIOC is compared to MICE, KNN, mean and mode using root-mean-square error (RMSE). The results suggest that HIOC is the most appropriate method for predicting missing data in medical datasets and provides stable performance even when the rate of missing data is increased from 10 to 30%. It provided a reduction in RMSE up to 18.06% in heart disease dataset and 24.62% in breast cancer dataset. 

A Novel Hybrid Imputation Method to Predict Missing Values in Medical Datasets

Objectives: Exploratory data study is regularly indispensable to evaluate a potential premise for the subsequent studies such as grouping the data in clusters or diversifying the data in classification. Very common incident in the real data is incompleteness. Methods/Statistical Analysis: This problem can result in the biased treatment comparisons and also impacts the overall statistical power of the study. Missing data are proposed in several methods. The central idea of this proposed method is to handle the uncertainty of the missing values due to the vagueness arises in the real world datasets. This research work overcomes the inconsistency of the missing datasets and the proposed method tolerates the missing values using the fuzzy based K-NN. Three different well known dataset are used in this paper. Findings: The results demonstrate that the proposed method is capable of imputing the missing values even with high presence of missing values and overwhelmed the problem of uncertainty precisely. Application/Improvements: As compared to other techniques the proposed fuzzy based K-NN gives high inconsistency. In future it can be improved in concentrating with big dataset and more effectively and efficiently result could be substituted by applying the expected value.

R. Naveen Kumar

Papers

Enhanced Fuzzy K-NN Approach for Handling Missing Values in Medical Data Mining