Optimal Feature Aggregation and Combination for Two-Dimensional Ensemble Feature Selection

Alhamidi, Machmud Roby; Jatmiko, Wisnu

doi:10.3390/info11010038

Cited by 15 publications

(7 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, as mentioned previously, to the best of our knowledge, stability of FS has not been considered in the field of PdM, including the various industrial processes, components, and monitoring tasks existing in this field. A recent topic that has been studied in conjunction with the stability of FS is the homogeneous ensemble of FS [14], [23], [40]- [42]. It has been studied in many domain areas, mainly with the aim of improving the FS stability.…”

Section: Related Workmentioning

confidence: 99%

On the Stability and Homogeneous Ensemble of Feature Selection for Predictive Maintenance: A Classification Application to Tool Condition Monitoring in Milling

Assafo¹,

Städter²,

Meisel³

et al. 2023

Preprint

View full text Add to dashboard Cite

<p>Feature selection (FS) represents an essential step for many machine learning-based predictive maintenance (PdM) applications, including various industrial processes, components, and monitoring tasks. The selected features do not only serve as inputs to the learning models, but also they can influence further decisions and analysis, e.g., sensor selection, understandability of the PdM system. Hence, before deploying the PdM system, it is crucial to examine the reproducibility and robustness of the selected features under variations in the input data. This is particularly critical for the real-world datasets with a low sample-to-dimension ratio (SDR). However, to the best of our knowledge, stability of the FS methods under data variations has not been considered yet in the field of PdM. This paper addresses this issue with an application to tool condition monitoring in milling, where classifiers based on support vector machine and random forest were employed. We used a 5-fold cross-validation to evaluate three popular filter-based FS methods, namely Fisher score, maximum relevance minimum redundancy (mRMR), and ReliefF, in terms of both stability and macro-F1. Further, for each method, we investigated the impact of the homogeneous FS ensemble on both performance indicators. To gain broad insights, we used four (2:2) milling datasets generated from our experiments and NASA’s repository, which differ in the operating conditions, sensors, SDR, number of classes, etc. Among the findings: 1) Different FS methods can yield comparable macro-F1, yet considerably different FS stability values. 2) Fisher score (single and/or ensemble) is the superior in most of the cases. 3) mRMR’s stability is overall the lowest, the most variable over different settings, e.g., sensor(s), subset cardinality, and the one most benefiting from the ensemble. </p>

show abstract

Section: Related Workmentioning

confidence: 99%

On the Stability and Homogeneous Ensemble of Feature Selection for Predictive Maintenance: A Classification Application to Tool Condition Monitoring in Milling

Assafo¹,

Städter²,

Meisel³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Several research contributions have been done to improve the ensemble methods which currently remain a major challenger of deep learning which is most used [21]. Recently, many studies [40][41][42][43] have shown that the combination of ensemble models with preprocessing techniques improves performance in modeling of the unbalanced classification problem. For a CRM dataset such as bank churn data having heterogeneous features and imbalance classes, appropriate data preprocessing is necessary to have the best model performance.…”

Section: The Proposed Approachmentioning

confidence: 99%

Towards Explainable Machine Learning for Bank Churn Prediction Using Data Balancing and Ensemble-Based Methods

et al. 2022

View full text Add to dashboard Cite

The diversity of data collected on both social networks and digital interfaces is extremely increased, raising the problem of heterogeneous variables that are not often favourable to classification algorithms. Despite the significant improvement in machine learning (ML) and predictive analysis efficiency for classification in customer relationship management systems (CRM), their performance remains very limited by heterogeneous data processing, class imbalance, and feature scales. This impact turned out to be more important for simple ML methods which in addition often suffer from over-fitting. This paper proposes a succinct and detailed ML model building process including cross-validation of the combination of SMOTE to balance data and ensemble methods for modelling. From the conducted experiments, the random forest (RF) model yielded the best performance of 0.86 in terms of accuracy and f1-scoreusing balanced data. It confirms the literature summary about this topic which shows that RF was among the most effective algorithms for customer predictive classification issues. The constructed and optimized models were interpreted by Shapley values and feature importance analysis which shows that the “age” feature was the most significant while “HasCrCard” was the less one. This process has proven effective in bridging previously reported research gaps and the resulting model should be used for supporting bank customer loyalty decision-making.

show abstract

“…The idea behind this is that, by flipping the bits, there is a possibility of de-selecting an irrelevant or redundant feature as well as a possibility of selecting a highly relevant feature. The presence of these irrelevant or redundant features causes a degradation in the classification performance (Alhamidi, 2020;Mafarja, 2019). Therefore, performing this perturbance provides room for possible improvement.…”

Section: One-way Local Search (Exploitation Only)mentioning

confidence: 99%

Integration of Specific Local Search Methods in Metaheuristic Algorithms for Optimizing the Feature Selection Process: A Survey Authors

Jeremiah¹

2022

SLUJST

View full text Add to dashboard Cite

Metaheuristic algorithms have proven to be quite effective at solving global optimization issues, particularly feature selection difficulties. This class of algorithms often uses a specialized local search technique as an inner component or as a post-processing mechanism to improve the performance of their search process. This paper presents a comprehensive survey of the use of local search methods integrated into metaheuristic algorithms for optimizing the feature selection process. Based on the manner of operation, the local search methods examined in this study were classed as one-way or two-way. In addition, practical suggestions were also discussed to point out possible future directions.

show abstract

Optimal Feature Aggregation and Combination for Two-Dimensional Ensemble Feature Selection

Cited by 15 publications

References 33 publications

On the Stability and Homogeneous Ensemble of Feature Selection for Predictive Maintenance: A Classification Application to Tool Condition Monitoring in Milling

On the Stability and Homogeneous Ensemble of Feature Selection for Predictive Maintenance: A Classification Application to Tool Condition Monitoring in Milling

Towards Explainable Machine Learning for Bank Churn Prediction Using Data Balancing and Ensemble-Based Methods

Integration of Specific Local Search Methods in Metaheuristic Algorithms for Optimizing the Feature Selection Process: A Survey Authors

Contact Info

Product

Resources

About