Incremental Gaussian Discriminant Analysis based on Graybill and Deal weighted combination of estimators for brain tumour diagnosis

Abstract

In the last decade, machine learning (ML) techniques have been used for developing classifiers for automatic brain tumour diagnosis. However, the development of these ML models rely on a unique training set and learning stops once this set has been processed. Training these classifiers requires a representative amount of data, but the gathering, preprocess, and validation of samples is expensive and time-consuming. Therefore, for a classical, non-incremental approach to ML, it is necessary to wait long enough to collect all the required data. In contrast, an incremental learning approach may allow us to build an initial classifier with a smaller number of samples and update it incrementally when new data are collected. In this study, an incremental learning algorithm for Gaussian Discriminant Analysis (iGDA) based on the Graybill and Deal weighted combination of estimators is introduced. Each time a new set of data becomes available, a new estimation is carried out and a combination with a previous estimation is performed. iGDA does not require access to the previously used data and is able to include new classes that were not in the original analysis, thus allowing the customization of the models to the distribution of data at a particular clinical center. An evaluation using five benchmark databases has been used to evaluate the behaviour of the iGDA algorithm in terms of stability-plasticity, class inclusion and order effect. Finally, the iGDA algorithm has been applied to automatic brain tumour classification with magnetic resonance spectroscopy, and compared with two state-of-the-art incremental algorithms. The empirical results obtained show the ability of the algorithm to learn in an incremental fashion, improving the performance of the models when new information is available, and converging in the course of time. Furthermore, the algorithm shows a negligible instance and concept order effect, avoiding the bias that such effects could introduce.

Clinical decision support systems for brain tumor characterization using advanced magnetic resonance imaging techniques

“…In such cases, the implementation of incremental learning algorithms is a promising solution for clinical environments. Tortajada et al [53] evaluated the performance of an incremental classifier based on single voxel Short TE spectra in comparison to static classifiers. The results revealed that the classification performance was improved when the incremental classifiers were used comparing to performance of the static classifies.…”

Clinical Decision Support Systems for Brain Tumour Diagnosis: Classification and Evaluation Approaches

“…This is highly convenient since there exists large 1.5T databases compiled throughout several years and the prediction models based on 1.5T acquisitions can be used for diagnosis of cases acquired with 3T instruments. In addition, the result reported in [125] can also be applied to the work presented by Tortajada et al [126], where they presented an incremental learning algorithm and successfully applied it to several diagnosis classification problems using MR spectra. Tortajada et al realized that in a clinical or research setting, the gathering, pre-process, and validation of samples is expensive and time-consuming.…”

Incremental Learning approaches to Biomedical decision problems

“…Unlike many of the state-of-the-art incremental learning algorithms, we assume that, once the datasets are used for fitting the models, they will not be available again. The first incremental learning algorithm is based on maximum-likelihood parameter estimation and a weighted combination of the old model parameters with the new dataset to develop a new model [28]. The second incremental learning algorithm is based on the Bayesian inference paradigm.…”

“…However, there are real scenarios with data protection policies or possible conflict of interests where the previous data are not available if a model trained with data from one organization is moved to another organization. For instance, when a model is moved from one hospital to another, patient consent and ethical approval needs to be obtained to send and store data [28]. In order to take into account this condition, we will assume that previous data are not accessible at all.…”

“…Furthermore, the algorithm shows a negligible instance and concept order effect, avoiding the bias that such effects could introduce. [28].…”