Graph classification based on graph set reconstruction and graph kernel feature reduction

Ma, Tinghuai; Shao, Wenye; Hao, Yongsheng; Cao, Jie

doi:10.1016/j.neucom.2018.03.029

Cited by 39 publications

(7 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…of the user at a particular point in time. The authors in [11,12] proposed a novel method for human activity recognition that uses a filtering algorithm to pre-process the sensor data before activity classification with multi-strategy feature fusion. Moreover, data collection was done by using the gravity, gyro and altitude sensors on an Android smartphone.…”

Section: Related Workmentioning

confidence: 99%

Smartphone Architecture for Edge-Centric IoT Analytics

Marah

Jing

et al. 2020

Sensors

Self Cite

View full text Add to dashboard Cite

The current baseline architectures in the field of the Internet of Things (IoT) strongly recommends the use of edge computing in the design of the solution applications instead of the traditional approach which solely uses the cloud/core for analysis and data storage. This research, therefore, focuses on formulating an edge-centric IoT architecture for smartphones which are very popular electronic devices that are capable of executing complex computational tasks at the network edge. A novel smartphone IoT architecture (SMIoT) is introduced that supports data capture and preprocessing, model (i.e., machine learning models) deployment, model evaluation and model updating tasks. Moreover, a novel model evaluation and updating scheme is provided which ensures model validation in real-time. This ensures a sustainable and reliable model at the network edge that automatically adjusts to changes in the IoT data subspace. Finally, the proposed architecture is tested and evaluated using an IoT use case.

show abstract

Section: Related Workmentioning

confidence: 99%

Smartphone Architecture for Edge-Centric IoT Analytics

Marah

Jing

et al. 2020

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…The method is easy to implement, and the leakage risk is measurable, so it is widely used [26]. However, because of the increase of the attacker's background knowledge, the effect of privacy protection is getting worse and worse [27]. In order to provide more protection of privacy, the differential privacy has been applied to privacy protection.…”

Section: Introductionmentioning

confidence: 99%

A New Method of Privacy Protection: Random k-Anonymous

Song

Tian

et al. 2019

IEEE Access

Self Cite

View full text Add to dashboard Cite

A new k-anonymous method which is different from traditional k-anonymous was proposed to solve the problem of privacy protection. Specifically, numerical data achieves k-anonymous by adding noises, and categorical data achieves k-anonymous by using randomization. Using the above two methods, the drawback that at least k elements must have the same quasi identifier in the k-anonymous data set has been solved. Since the process of finding anonymous equivalence is very time consuming, a two-step clustering method is used to divide the original data set into equivalence classes. First, the original data set is divided into several different sub-datasets, and then the equivalence classes are formed in the sub-datasets, thus greatly reducing the computational cost of finding anonymous equivalence classes. The experiments are conducted on three different data sets, and the results show that the proposed method is more efficient and the information loss of anonymous dataset is much smaller.

show abstract

“…Feature selection methods are widely used in data mining and machine learning in order to improve the accuracy of classifiers and accelerate the training speed of models with large amounts of input features [Migdady (2013); Rong, Ma, Cao et al (2019)]. From the original feature set, it allows us to eliminate irrelevant and strongly correlated redundant features and so to extract those valuable features [Paisitkriangkrai, Shen and Hengel (2016); Ma, Shao, Hao et al (2018)]. Feature selection is a difficult task due mainly to a large search space, where the total number of possible solutions is 2 n for a dataset with n features [Xue, Zhang, Browne et al (2016)].…”

Section: Introductionmentioning

confidence: 99%

Feature Selection with a Local Search Strategy Based on the Forest Optimization Algorithm

Ma¹,

Zhou²,

Jia³

et al. 2019

Computer Modeling in Engineering &Amp; Sciences

Self Cite

View full text Add to dashboard Cite

Feature selection has been widely used in data mining and machine learning. Its objective is to select a minimal subset of features according to some reasonable criteria so as to solve the original task more quickly. In this article, a feature selection algorithm with local search strategy based on the forest optimization algorithm, namely FSLSFOA, is proposed. The novel local search strategy in local seeding process guarantees the quality of the feature subset in the forest. Next, the fitness function is improved, which not only considers the classification accuracy, but also considers the size of the feature subset. To avoid falling into local optimum, a novel global seeding method is attempted, which selects trees on the bottom of candidate set and gives the algorithm more diversities. Finally, FSLSFOA is compared with four feature selection methods to verify its effectiveness. Most of the results are superior to these comparative methods.

show abstract

Graph classification based on graph set reconstruction and graph kernel feature reduction

Cited by 39 publications

References 23 publications

Smartphone Architecture for Edge-Centric IoT Analytics

Smartphone Architecture for Edge-Centric IoT Analytics

A New Method of Privacy Protection: Random k-Anonymous

Feature Selection with a Local Search Strategy Based on the Forest Optimization Algorithm

Contact Info

Product

Resources

About