We propose Deep Dict, a deep learning-based lossy time series compressor designed to achieve a high compression ratio while maintaining decompression error within a predefined range. Deep Dict incorporates two essenti...
详细信息
ISBN:
(纸本)9781728190549
We propose Deep Dict, a deep learning-based lossy time series compressor designed to achieve a high compression ratio while maintaining decompression error within a predefined range. Deep Dict incorporates two essential components: the Bernoulli transformer autoencoder (BTAE) and a distortion constraint. BTAE extracts Bernoulli representations from time series data, reducing the size of the representations compared to conventional autoencoders. The distortion constraint limits the prediction error of BTAE to the desired range. Moreover, in order to address the limitations of common regression losses such as L1/L2, we introduce a novel loss function called quantized entropy loss (QEL). QEL takes into account the specific characteristics of the problem, enhancing robustness to outliers and alleviating optimization challenges. Our evaluation of Deep Dict across diverse time series datasets from various iot domains reveals that Deep Dict outperforms state-of-the-art lossy compressors in terms of compression ratio by a significant margin by up to 53.66%.
The paper presents a multi-layered datacompression framework that reduces the amount of data before being stored in cloud. At present, Internet of Things (iot) has gained noticeable attention due to the approaches an...
详细信息
ISBN:
(纸本)9781538692424
The paper presents a multi-layered datacompression framework that reduces the amount of data before being stored in cloud. At present, Internet of Things (iot) has gained noticeable attention due to the approaches and advancements towards smart city aspects. With increasing number of devices and sensors connected to the Internet, tremendous amount of data is being generated at every moment which requires volumes of storage space to be stored. However, datacompression techniques can reduce the size of the data and the storage requirement by compressing the data more efficiently. In this article we introduced a two layered compression framework for iotdata that reduces the amount of data with maintaining minimum error rate as well as avoiding bandwidth wastage. In our proposed datacompression scheme, we got an initial compression at the fog nodes by 50% compression ratio and in the Cloud storage we have compressed the data up to 90%. We also showed that the error is varied from the original data by 0% to 1.5% after decompression.
暂无评论