Using reinforcement learning to find an optimal set of features

Fard, Seyed Mehdi Hazrati; Hamzeh, Ali; Hashemi, Sattar

doi:10.1016/j.camwa.2013.06.031

Cited by 42 publications

(25 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This type of learning has a lot of potential for effective feature selection in the subspace of features. Feature selection can be performed through singleagent 40,41 or multi-agent 42 decision processes. In a single-agent process, only one agent decides on the selection or deselection of features, resulting in a large action space and the risk of getting stuck in a local optimum solution.…”

Section: Reinforcement Learningmentioning

confidence: 99%

Bird's Eye View Feature Selection for High-Dimensional Data

Belhaouari

Shakeel

Erbad

et al. 2023

Preprint

View full text Add to dashboard Cite

In machine learning, an informative dataset is crucial for accurate predictions. However, high dimensional data often contains irrelevant features, outliers, and noise, which can negatively impact model performance and consume computational resources. To tackle this challenge, the Bird's Eye View (BEV) feature selection technique is introduced. This approach is inspired by the natural world, where a bird searches for important features in a sparse dataset, similar to how a bird search for sustenance in a sprawling jungle. BEV incorporates elements of Evolutionary Algorithms with a Genetic Algorithm to maintain a population of top-performing agents, Dynamic Markov Chain to steer the movement of agents in the search space, and Reinforcement Learning to reward and penalize agents based on their progress. The proposed strategy in this paper leads to improved classification performance and a reduced number of features compared to conventional methods, as demonstrated by outperforming state-of-the-art feature selection techniques across multiple benchmark datasets.

show abstract

Section: Reinforcement Learningmentioning

confidence: 99%

Bird's Eye View Feature Selection for High-Dimensional Data

Belhaouari

Shakeel

Erbad

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Some other score functions include statistical scores (variance, t-score, chisquared score, Gini index (Gini, 1912)), similarity scores (Laplacian score (He et al, 2005), SPEC (Zhao and Liu, 2007), Fisher score (Hart et al, 2000), Trace Ratio (Nie et al, 2008)), and information-theoretical-based scores (Mutual Information (Battiti, 1994), MRMR (Peng et al, 2005), CI (Lin and Tang, 2006), JMI (Yang and Moody, 1999), CMI (Vidal-Naquet and Ullman, 2003)). Despite the computation efficiency (Fard et al, 2013), the variables selected by filtering methods are non-optimal since the filtering is done in the preprocessing step and is independent of the main task.…”

Section: Introductionmentioning

confidence: 99%

Reinforced Variable Selection via Natural Policy Gradient

Le,

Zhou,

Bai

2024

Preprint

View full text Add to dashboard Cite

Variable selection identifies the best subset of covariates when building the prediction model, among all possible subsets. In this paper, we propose a novel reinforced variable selection method, called “Actor-Critic-Predictor”. The actor takes an action to choose variables and the predictor evaluates the action based on a well-designed reward function, where the reward baseline is learned by the critic. We model the variable selection process as a multi-armed bandit and update the subset of variables being selected using a natural policy gradient algorithm. We provide an analytical framework on how different errors impact the performance of our method theoretically. Large amounts of experiments on both synthetic and real datasets show that the proposed framework is easy-implemented and outperforms classical variable selection methods in a wide range of scenarios.

show abstract

“…Numerous research have explored ways in which an agent is used to select a subset of data characteristics. 43 Unfortunately, this method needs the agent to decide whether to include or exclude each characteristic. 44 This results in an exponential increase of the explorable interval (2N), similar to evolutionary algorithms 33,35 , and it is challenging to discover the global optimal solution and has a signi cant computing cost.…”

Section: Introductionmentioning

confidence: 99%

MAFSIDS ： A reinforcement learning-based intrusion detection model for multi-intelligence feature selection networks

Ren

Zeng

Zhong

et al. 2023

Preprint

View full text Add to dashboard Cite

In the era of big data, the growing number of cyber assaults poses a significant danger to network services. Intrusion detection systems (IDS) rely on the quality of its features to accurately identify cyber threats. Nowadays prevalent IDS prefer to create intricate neural networks and pay less attention to the problem of feature selection. In this study, we present a multi-intelligence feature selection network intrusion detection model based on reinforcement learning. The model extracts feature information of network traffic by means of a graph convolutional neural network (GCN), using multiple deep Q-network (DQN)-based intelligences to decide whether the corresponding features are selected, and then trains classifiers to identify network attacks by means of deep reinforcement learning (DRL). We examined the model's performance using both the NSL-KDD and CSE-CIC-IDS2018 datasets. The simulation experimental results demonstrate that MAFSIDS is able to extract accurate feature information from the input data via the GCN network, and that the multi-intelligence will then select the optimal feature subset and learn the data via DRL to ultimately enhance the model's cyber attack recognition performance. In the era of big data, the model has vast application potential and provides a solid assurance for network security.

show abstract

Using reinforcement learning to find an optimal set of features

Cited by 42 publications

References 7 publications

Bird's Eye View Feature Selection for High-Dimensional Data

Bird's Eye View Feature Selection for High-Dimensional Data

Reinforced Variable Selection via Natural Policy Gradient

MAFSIDS ： A reinforcement learning-based intrusion detection model for multi-intelligence feature selection networks

Contact Info

Product

Resources

About