Deep Cascade of Extra Trees

Berrouachedi, Abdelkader; Jaziri, Rakia; Bernard, Gilles

doi:10.1007/978-3-030-26142-9_11

Cited by 7 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The simplification can also occur once the DF architecture is trained, as in [11] selecting in each forest the most important paths to reduce the network time-and memory-complexity. Approaches to increase the approximation capacity of DF have also been proposed by adjoining weights to trees or to forests in each layer [20,21], replacing the forest by more complex estimators (cascade of ExtraTrees) [2], or by combining several of the previous modifications notably incorporating data preprocessing [9]. Overall, the related works on DF exclusively represent algorithmic contributions without a formal understanding of the driving mechanisms at work inside the forest cascade.…”

Section: Introductionmentioning

confidence: 99%

Analyzing the tree-layer structure of Deep Forests

Ludovic¹,

Boyer²,

Scornet³

et al. 2020

Preprint

View full text Add to dashboard Cite

Random forests on the one hand, and neural networks on the other hand, have met great success in the machine learning community for their predictive performance. Combinations of both have been proposed in the literature, notably leading to the so-called deep forests (DF) [25]. In this paper, we investigate the mechanisms at work in DF and outline that DF architecture can generally be simplified into more simple and computationally efficient shallow forests networks. Despite some instability, the latter may outperform standard predictive tree-based methods. In order to precisely quantify the improvement achieved by these light network configurations over standard tree learners, we theoretically study the performance of a shallow tree network made of two layers, each one composed of a single centered tree. We provide tight theoretical lower and upper bounds on its excess risk. These theoretical results show the interest of tree-network architectures for well-structured data provided that the first layer, acting as a data encoder, is rich enough.

show abstract

Section: Introductionmentioning

confidence: 99%

Analyzing the tree-layer structure of Deep Forests

Ludovic¹,

Boyer²,

Scornet³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The analysis process involved the selection of features that were theoretically linked to gas-sensitive properties, and the data set was divided into a training set (70%) and a test set (30%), both sets underwent normalization. Ten algorithms were adopted to identify the most suitable machine learning model, which are Multilayer Perceptron (MLP), Logistic Regression, Random Forest, K-Nearest Neighbors (KNNs), Support Vector Machine (SVM), Decision Tree, Extra Trees, AdaBoost, Bagging, and Voting . To mitigate overfitting in the context of the logistic regression algorithm, we employed L2 regularization as a method to reduce the risk of overfitting.…”

mentioning

confidence: 99%

General Model for Predicting Response of Gas-Sensitive Materials to Target Gas Based on Machine Learning

Yang,

Sun,

Gao

et al. 2024

ACS Sens.

View full text Add to dashboard Cite

Gas sensors play a crucial role in various industries and applications. In recent years, there has been an increasing demand for gas sensors in society. However, the current method for screening gas-sensitive materials is time-, energy-, and cost-consuming. Consequently, an imperative exists to enhance the screening efficiency. In this study, we proposed a collaborative screening strategy through integration of density functional theory and machine learning. Taking zinc oxide (ZnO) as an example, the responsiveness of ZnO to the target gas was determined quickly on the basis of the changes in the electronic state and structure before and after gas adsorption. In this work, the adsorption energy and electronic and structural characteristics of ZnO after adsorbing 24 kinds of gases were calculated. These computed features served as the basis for training a machine learning model. Subsequently, various machine learning and evaluation algorithms were utilized to train the fast screening model. The importance of feature values was evaluated by the AdaBoost, Random Forest, and Extra Trees models. Specifically, charge transfer was assigned importance values of 0.160, 0.127, and 0.122, respectively, ranking as the highest among the 11 features. Following closely was the d-band center, which was presumed to exert influence on electrical conductivity and, consequently, adsorption properties. With 5-fold cross-validation using the Extra Tree accuracy, the 24-sample data set achieved an accuracy of 88%. The 72-sample data set achieved an accuracy of 78% using multilayer perceptron after 5-fold cross-validation, with both data sets exhibiting low standard deviations. This verified the accuracy and reliability of the strategy, showcasing its potential for rapidly screening a material's responsiveness to the target gas.

show abstract

Rapid Discovery of Gas Response in Materials Via Density Functional Theory and Machine Learning

Gao,

Cheng,

Chen

et al. 2024

Energy & Environ Materials

View full text Add to dashboard Cite

In this study, a framework for predicting the gas‐sensitive properties of gas‐sensitive materials by combining machine learning and density functional theory (DFT) has been proposed. The framework rapidly predicts the gas response of materials by establishing relationships between multisource physical parameters and gas‐sensitive properties. In order to prove its effectiveness, the perovskite Cs3Cu2I5 has been selected as the representative material. The physical parameters before and after the adsorption of various gases have been calculated using DFT, and then a machine learning model has been trained based on these parameters. Previous studies have shown that a single physical parameter alone is not enough to accurately predict the gas sensitivity of materials. Therefore, a variety of physical parameters have been selected for machine learning, and the final machine learning model achieved 92% accuracy in predicting gas sensitivity. It is important to note that although there have been no previous reports on the response of Cs3Cu2I5 to hydrogen sulfide, the resulting model predicts the gas response of H2S; it is subsequently confirmed experimentally. This method not only enhances the understanding of the gas sensing mechanism, but also has a universal nature, making it suitable for the development of various new gas‐sensitive materials.

show abstract

Deep Cascade of Extra Trees

Cited by 7 publications

References 13 publications

Analyzing the tree-layer structure of Deep Forests

Analyzing the tree-layer structure of Deep Forests

General Model for Predicting Response of Gas-Sensitive Materials to Target Gas Based on Machine Learning

Rapid Discovery of Gas Response in Materials Via Density Functional Theory and Machine Learning

Contact Info

Product

Resources

About