A post-selection algorithm for improving dynamic ensemble selection methods

Dynamic Ensemble Selection (DES) is a Multiple Classifier Systems (MCS) approach that aims to select an ensemble for each query sample during the selection phase. Even with the proposal of several DES approaches, no particular DES technique is the best choice for different problems. Thus, we hypothesize that selecting the best DES approach per query instance can lead to better accuracy. To evaluate this idea, we introduce the Post-Selection Dynamic Ensemble Selection (PS-DES) approach, a post-selection scheme that evaluates ensembles selected by several DES techniques using different metrics. Experimental results show that using accuracy as a metric to select the ensembles, PS-DES performs better than individual DES techniques. PS-DES source code is available in a GitHub repository

The choice of scaling technique matters for classification performance

Dataset scaling, also known as normalization, is an essential preprocessing step in a machine learning pipeline. It is aimed at adjusting attributes scales in a way that they all vary within the same range. This transformation is known to improve the performance of classification models, but there are several scaling techniques to choose from, and this choice is not generally done carefully. In this paper, we execute a broad experiment comparing the impact of 5 scaling techniques on the performances of 20 classification algorithms among monolithic and ensemble models, applying them to 82 publicly available datasets with varying imbalance ratios. Results show that the choice of scaling technique matters for classification performance, and the performance difference between the best and the worst scaling technique is relevant and statistically significant in most cases. They also indicate that choosing an inadequate technique can be more detrimental to classification performance than not scaling the data at all. We also show how the performance variation of an ensemble model, considering different scaling techniques, tends to be dictated by that of its base model. Finally, we discuss the relationship between a model's sensitivity to the choice of scaling technique and its performance and provide insights into its applicability on different model deployment scenarios. Full results and source code for the experiments in this paper are available in a GitHub repository.\footnote{https://github.com/amorimlb/scaling\_matters}

Local overlap reduction procedure for dynamic ensemble selection

Class imbalance is a characteristic known for making learning more challenging for classification models as they may end up biased towards the majority class. A promising approach among the ensemble-based methods in the context of imbalance learning is Dynamic Selection (DS). DS techniques single out a subset of the classifiers in the ensemble to label each given unknown sample according to their estimated competence in the area surrounding the query. Because only a small region is taken into account in the selection scheme, the global class disproportion may have less impact over the system's performance. However, the presence of local class overlap may severely hinder the DS techniques' performance over imbalanced distributions as it not only exacerbates the effects of the under-representation but also introduces ambiguous and possibly unreliable samples to the competence estimation process. Thus, in this work, we propose a DS technique which attempts to minimize the effects of the local class overlap during the classifier selection procedure. The proposed method iteratively removes from the target region the instance perceived as the hardest to classify until a classifier is deemed competent to label the query sample. The known samples are characterized using instance hardness measures that quantify the local class overlap. Experimental results show that the proposed technique can significantly outperform the baseline as well as several other DS techniques, suggesting its suitability for dealing with class under-representation and overlap. Furthermore, the proposed technique still yielded competitive results when using an under-sampled, less overlapped version of the labelled sets, specially over the problems with a high proportion of minority class samples in overlap areas. Code available at https://github.com/marianaasouza/lords.

Selecting and combining complementary feature representations and classifiers for hate speech detection

Hate speech is a major issue in social networks due to the high volume of data generated daily. Recent works demonstrate the usefulness of machine learning (ML) in dealing with the nuances required to distinguish between hateful posts from just sarcasm or offensive language. Many ML solutions for hate speech detection have been proposed by either changing how features are extracted from the text or the classification algorithm employed. However, most works consider only one type of feature extraction and classification algorithm. This work argues that a combination of multiple feature extraction techniques and different classification models is needed. We propose a framework to analyze the relationship between multiple feature extraction and classification techniques to understand how they complement each other. The framework is used to select a subset of complementary techniques to compose a robust multiple classifiers system (MCS) for hate speech detection. The experimental study considering four hate speech classification datasets demonstrates that the proposed framework is a promising methodology for analyzing and designing high-performing MCS for this task. MCS system obtained using the proposed framework significantly outperforms the combination of all models and the homogeneous and heterogeneous selection heuristics, demonstrating the importance of having a proper selection scheme. Source code, figures, and dataset splits can be found in the GitHub repository: https://github.com/Menelau/Hate-Speech-MCS.

Label noise detection under the Noise at Random model with ensemble filters

Label noise detection has been widely studied in Machine Learning because of its importance in improving training data quality. Satisfactory noise detection has been achieved by adopting ensembles of classifiers. In this approach, an instance is assigned as mislabeled if a high proportion of members in the pool misclassifies it. Previous authors have empirically evaluated this approach; nevertheless, they mostly assumed that label noise is generated completely at random in a dataset. This is a strong assumption since other types of label noise are feasible in practice and can influence noise detection results. This work investigates the performance of ensemble noise detection under two different noise models: the Noisy at Random (NAR), in which the probability of label noise depends on the instance class, in comparison to the Noisy Completely at Random model, in which the probability of label noise is entirely independent. In this setting, we investigate the effect of class distribution on noise detection performance since it changes the total noise level observed in a dataset under the NAR assumption. Further, an evaluation of the ensemble vote threshold is conducted to contrast with the most common approaches in the literature. In many performed experiments, choosing a noise generation model over another can lead to different results when considering aspects such as class imbalance and noise level ratio among different classes.

Multi-label learning for dynamic model type recommendation

Dynamic selection techniques aim at selecting the local experts around each test sample in particular for performing its classification. While generating the classifier on a local scope may make it easier for singling out the locally competent ones, as in the online local pool (OLP) technique, using the same base-classifier model in uneven distributions may restrict the local level of competence, since each region may have a data distribution that favors one model over the others. Thus, we propose in this work a problem-independent dynamic base-classifier model recommendation for the OLP technique, which uses information regarding the behavior of a portfolio of models over the samples of different problems to recommend one (or several) of them on a per-instance manner. Our proposed framework builds a multi-label meta-classifier responsible for recommending a set of relevant model types based on the local data complexity of the region surrounding each test sample. The OLP technique then produces a local pool with the model that yields the highest probability score of the meta-classifier. Experimental results show that different data distributions favored different model types on a local scope. Moreover, based on the performance of an ideal model type selector, it was observed that there is a clear advantage in choosing a relevant model type for each test instance. Overall, the proposed model type recommender system yielded a statistically similar performance to the original OLP with fixed base-classifier model. Given the novelty of the approach and the gap in performance between the proposed framework and the ideal selector, we regard this as a promising research direction. Code available at github.com/marianaasouza/dynamic-model-recommender.

Evaluating Competence Measures for Dynamic Regressor Selection

Dynamic regressor selection (DRS) systems work by selecting the most competent regressors from an ensemble to estimate the target value of a given test pattern. This competence is usually quantified using the performance of the regressors in local regions of the feature space around the test pattern. However, choosing the best measure to calculate the level of competence correctly is not straightforward. The literature of dynamic classifier selection presents a wide variety of competence measures, which cannot be used or adapted for DRS. In this paper, we review eight measures used with regression problems, and adapt them to test the performance of the DRS algorithms found in the literature. Such measures are extracted from a local region of the feature space around the test pattern, called region of competence, therefore competence measures.To better compare the competence measures, we perform a set of comprehensive experiments of 15 regression datasets. Three DRS systems were compared against individual regressor and static systems that use the Mean and the Median to combine the outputs of the regressors from the ensemble. The DRS systems were assessed varying the competence measures. Our results show that DRS systems outperform individual regressors and static systems but the choice of the competence measure is problem-dependent.

ICPRAI 2018 SI: On dynamic ensemble selection and data preprocessing for multi-class imbalance learning

Class-imbalance refers to classification problems in which many more instances are available for certain classes than for others. Such imbalanced datasets require special attention because traditional classifiers generally favor the majority class which has a large number of instances. Ensemble of classifiers have been reported to yield promising results. However, the majority of ensemble methods applied to imbalanced learning are static ones. Moreover, they only deal with binary imbalanced problems. Hence, this paper presents an empirical analysis of dynamic selection techniques and data preprocessing methods for dealing with multi-class imbalanced problems. We considered five variations of preprocessing methods and fourteen dynamic selection schemes. Our experiments conducted on 26 multi-class imbalanced problems show that the dynamic ensemble improves the AUC and the G-mean as compared to the static ensemble. Moreover, data preprocessing plays an important role in such cases.

Analyzing different prototype selection techniques for dynamic classifier and ensemble selection

In dynamic selection (DS) techniques, only the most competent classifiers, for the classification of a specific test sample are selected to predict the sample's class labels. The more important step in DES techniques is estimating the competence of the base classifiers for the classification of each specific test sample. The classifiers' competence is usually estimated using the neighborhood of the test sample defined on the validation samples, called the region of competence. Thus, the performance of DS techniques is sensitive to the distribution of the validation set. In this paper, we evaluate six prototype selection techniques that work by editing the validation data in order to remove noise and redundant instances. Experiments conducted using several state-of-the-art DS techniques over 30 classification problems demonstrate that by using prototype selection techniques we can improve the classification accuracy of DS techniques and also significantly reduce the computational cost involved.

META-DES.H: a dynamic ensemble selection technique using meta-learning and a dynamic weighting approach

In Dynamic Ensemble Selection (DES) techniques, only the most competent classifiers are selected to classify a given query sample. Hence, the key issue in DES is how to estimate the competence of each classifier in a pool to select the most competent ones. In order to deal with this issue, we proposed a novel dynamic ensemble selection framework using meta-learning, called META-DES. The framework is divided into three steps. In the first step, the pool of classifiers is generated from the training data. In the second phase the meta-features are computed using the training data and used to train a meta-classifier that is able to predict whether or not a base classifier from the pool is competent enough to classify an input instance. In this paper, we propose improvements to the training and generalization phase of the META-DES framework. In the training phase, we evaluate four different algorithms for the training of the meta-classifier. For the generalization phase, three combination approaches are evaluated: Dynamic selection, where only the classifiers that attain a certain competence level are selected; Dynamic weighting, where the meta-classifier estimates the competence of each classifier in the pool, and the outputs of all classifiers in the pool are weighted based on their level of competence; and a hybrid approach, in which first an ensemble with the most competent classifiers is selected, after which the weights of the selected classifiers are estimated in order to be used in a weighted majority voting scheme. Experiments are carried out on 30 classification datasets. Experimental results demonstrate that the changes proposed in this paper significantly improve the recognition accuracy of the system in several datasets.