Web Objects Based Contextual Data Quality Assessment Model for Semantic Data Application

Jarwar, Muhammad Aslam; Chong, Ilyoung

doi:10.3390/app10062181

Cited by 7 publications

(5 citation statements)

References 44 publications

(65 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Fine-grained emotion recognition requires algorithms to predict multiple emotion states by relying on signals within one certain time interval. To train such recognition models, previous works [1], [3], [13], [14] need large amounts of data which are annotated in fine-level of granularity. Specifically, they usually require more than 90% of the annotated data in the datasets (e.g., CASE [16], RECOLA [17], K-EmoCon [19], MERCA [1]) to train an accurate recognition model.…”

Section: A Emotion Recognition On Small Datamentioning

confidence: 99%

“…A growing number of emotion recognition algorithms were developed in recent years [1]- [3] to model the temporal dynamics of emotion states of users. The accurate recognition of emotions while users consume different types of media content (e.g., videos, music, movies) can help content providers to better understand users' emotions towards the media content they provide and adjust it accordingly [4].…”

Section: Introductionmentioning

confidence: 99%

“…Previous works [1], [3], [13], [14] on fine-grained emotion recognition rely on large amounts of training data with fine-grained emotion labels. These labels are required to be collected in a fine level of granularity (normally the same or similar frequency as the input signal) to train the recognition algorithms [15].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Few-Shot Learning for Fine-Grained Emotion Recognition Using Physiological Signals

Zhang

Ali

Hanjalic

et al. 2023

IEEE Trans. Multimedia

View full text Add to dashboard Cite

Fine-grained emotion recognition can model the temporal dynamics of emotions, which is more precise than predicting one emotion retrospectively for an activity (e.g., video clip watching). Previous works require large amounts of continuously annotated data to train an accurate recognition model, however experiments to collect such large amounts of continuously annotated physiological signals are costly and time-consuming. To overcome this challenge, we propose an Emotion recognition algorithm based on Deep Siamese Networks (EmoDSN) which can rapidly converge on a small amount of training data, typically less than 10 samples per class (i.e., < 10 shot). EmoDSN recognizes fine-grained valence and arousal (V-A) labels by maximizing the distance metric between signal segments with different V-A labels. We tested EmoDSN on three different datasets collected in three different environments: desktop, mobile and HMD-based virtual reality, respectively. The results from our experiments show that EmoDSN achieves promising results for both one-dimension binary (high/low V-A, 1D-2C) and two-dimensional 5-class (four quadrants of V-A space + neutral, 2D-5C) classification. We get an averaged accuracy of 76.04%, 76.62% and 57.62% for 1D-2C valence, 1D-2C arousal, and 2D-5C, respectively, by using only 5 shots of training data. Our experiments show that EmoDSN can achieve better results if we select training samples from the changing points of emotion or the ending moments of video watching.

show abstract

Section: A Emotion Recognition On Small Datamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Few-Shot Learning for Fine-Grained Emotion Recognition Using Physiological Signals

Zhang

Ali

Hanjalic

et al. 2023

IEEE Trans. Multimedia

View full text Add to dashboard Cite

show abstract

“…Data quality assessments were performed several times based on the application context and requirement [20]. In addition, the quality of the same data can differ for different applications and users [25]. Thus, this is a time-consuming process.…”

Section: Qualitative Findingsmentioning

confidence: 99%

A Thematic Review on Data Quality Challenges and Dimension in the Era of Big Data

Ridzuan

Zainon

Zairul

2021

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

Data quality is the primary concern faced by most of the organizations due to improper maintenance in the database. Data obtained from the various resources are dirty, affecting the accuracy of predicted results. There are a lot of challenges when handling Big Data because it requires well-defined and precise measurement processes. The challenges are in the characteristics of big data itself where the V's play an important role in measuring and determining data quality. Although the issue has been discussed over 20 years, there is no guideline in identifying the important dimension of data quality being proposed to adhere with the context of Big Data. Therefore, the purpose of this systematic review is to review literature on the issue, challenges, and dimension of data quality in the era of Big Data using thematic review. This review included journal and conference proceeding papers from ACM Digital Library, Scopus, and Science Direct published between 2016 until 2020. Inclusion and exclusion processes have filtered out 21 final articles for the review. A systematic review on these 21 articles focuses on the issue, challenges, and dimension of data quality. The results of this study benefit the future study on the development of data quality dimensions and can be a guideline for the researcher to design the data quality assessment framework.

show abstract

“…Data quality has different meanings in different contexts. For example, data quality can be about measuring defective or outlier data in a general context [25][26][27], or describing whether the data meet the expected purpose in a specific context [28]. In this paper, we define data quality as a measure of data suitability for constructing a DL training set.…”

Section: Introductionmentioning

confidence: 99%

ADQE: Obtain Better Deep Learning Models by Evaluating the Augmented Data Quality Using Information Entropy

Cui,

Li,

Xie

et al. 2023

Electronics

View full text Add to dashboard Cite

Data augmentation, as a common technique in deep learning training, is primarily used to mitigate overfitting problems, especially with small-scale datasets. However, it is difficult for us to evaluate whether the augmented dataset truly benefits the performance of the model. If the training model is relied upon in each case to validate the quality of the data augmentation and the dataset, it will take a lot of time and resources. This article proposes a simple and practical approach to evaluate the quality of data augmentation for image classification tasks, enriching the theoretical research on data augmentation quality evaluation. Based on the information entropy, multiple dimensional metrics for data quality augmentation are established, including diversity, class balance, and task relevance. Additionally, a comprehensive data augmentation quality fusion metric is proposed. Experimental results on the CIFAR-10 and CUB-200 datasets show that our method maintains optimal performance in a variety of scenarios. The cosine similarity between the score of our method and the precision of model is up to 99.9%. A rigorous evaluation of data augmentation quality is necessary to guide the improvement of DL model performance. The quality standards and evaluation defined in this article can be utilized by researchers to train high-performance DL models in situations where data are limited.

show abstract

Web Objects Based Contextual Data Quality Assessment Model for Semantic Data Application

Cited by 7 publications

References 44 publications

Few-Shot Learning for Fine-Grained Emotion Recognition Using Physiological Signals

Few-Shot Learning for Fine-Grained Emotion Recognition Using Physiological Signals

A Thematic Review on Data Quality Challenges and Dimension in the Era of Big Data

ADQE: Obtain Better Deep Learning Models by Evaluating the Augmented Data Quality Using Information Entropy

Contact Info

Product

Resources

About