Keratoconus Severity Classification Using Features Selection and Machine Learning Algorithms

Abstract

Keratoconus is a noninflammatory disease characterized by thinning and bulging of the cornea, generally appearing during adolescence and slowly progressing, causing vision impairment. However, the detection of keratoconus remains difficult in the early stages of the disease because the patient does not feel any pain. Therefore, the development of a method for detecting this disease based on machine and deep learning methods is necessary for early detection in order to provide the appropriate treatment as early as possible to patients. Thus, the objective of this work is to determine the most relevant parameters with respect to the different classifiers used for keratoconus classification based on the keratoconus dataset of Harvard Dataverse. A total of 446 parameters are analyzed out of 3162 observations by 11 different feature selection algorithms. Obtained results showed that sequential forward selection (SFS) method provided a subset of 10 most relevant variables, thus, generating the highest classification performance by the application of random forest (RF) classifier, with an accuracy of 98% and 95% considering 2 and 4 keratoconus classes, respectively. Found classification accuracy applying RF classifier on the selected variables using SFS method achieves the accuracy obtained using all features of the original dataset.

Artificial intelligence-assisted diagnosis of ocular surface diseases

“…The epithelial features extracted from the OCT images were the most valuable for the discrimination process. 2021) also found that the RF outperformed the decision tree model (89% accuracy vs. 71%, respectively), while Aatila et al (2021) found the RF model to have the highest accuracy when compared with other ML models in detecting all classes of KC. AI has been used to screen potential candidates for refractive surgery besides detecting KC.…”

Artificial intelligence and corneal diseases

“…Castro-Luna et al [38] found that the random forest outperformed decision tree model (89% accuracy vs. 71%, respectively) based on tomographic and biomechanical variables. Cao et al [39] also found the random forest model outperformed other machine learning algorithms using tomographic and demographic data, while Aatila et al [40] found the random forest model to have the highest accuracy compared to other machine learning models trained on anterior segment (AS)-OCT images in detecting all classes of KCN including ffKCN.…”

Integrating machine learning with electronic health record data to facilitate detection of prolactin level and pharmacovigilance signals in olanzapine-treated patients

“…We used the mean absolute SHAP values (|SHAP value|) to illustrate the global importance of features ( 39 ). The sequential forward feature selection approach was then employed to generate the optimal feature subset ( 40 ). The general practices in this bottom-to-top search method involved starting with an empty feature subset, adding one feature out of the remaining features in each iteration (the order of addition depended on the feature importance calculated by the SHAP values: the more important the feature was, the greater was the precedence it had), and then evaluating the pros and cons of the generated feature subset by using 10-fold cross-validation on the derivation cohort.…”