Adaptive time series prediction and recommendation

作者：

Highlights：

• In terms of methodology, this paper offers new insight into the trend of prediction and temporal recommendation.

• From a theoretical point of view, we propose a family of time-series predictive models to explore the temporal patterns hidden behind the chronological order of data occurrence.

• We investigate an adaptive parameter optimization strategy based on the BFGS optimization algorithm for the proposed family of time-series predictive models.

• From a practical viewpoint, a novel hybrid Top-N recommendation framework (HNATS) is proposed to find out the influence and mechanism of structural and temporal information on recommendation, which synchronously improves the accuracy of recommendation.

摘要

•In terms of methodology, this paper offers new insight into the trend of prediction and temporal recommendation.•From a theoretical point of view, we propose a family of time-series predictive models to explore the temporal patterns hidden behind the chronological order of data occurrence.•We investigate an adaptive parameter optimization strategy based on the BFGS optimization algorithm for the proposed family of time-series predictive models.•From a practical viewpoint, a novel hybrid Top-N recommendation framework (HNATS) is proposed to find out the influence and mechanism of structural and temporal information on recommendation, which synchronously improves the accuracy of recommendation.

论文关键词：Time series prediction,Adaptive parameter optimization,Temporal recommendation,Hybrid recommendation

论文评审过程：Received 21 April 2020, Revised 26 December 2020, Accepted 4 January 2021, Available online 15 January 2021, Version of Record 15 January 2021.

论文官网地址：https://doi.org/10.1016/j.ipm.2021.102494