Temporal Relations Research Articles

Local spatial and temporal relation discovery model based on attention mechanism for traffic forecasting

Recognizing the evolution pattern of traffic condition and making accurate prediction play a vital role in intelligent transportation systems (ITS). With the massive increase of available traffic data, deep learning-based models have attracted considerable attention for their impressive performance in traffic forecasting. However, the majority of existing approaches neglect to model of asynchronously dynamic spatio-temporal correlation and fail to consider the impact of historical traffic data on future condition. Additionally, the attribute of deep learning method presents challenges in interpreting the explicit spatiotemporal relationships. In order to enhance the accuracy of traffic prediction as well as extract comprehensive and explainable spatial–temporal relevance in traffic networks, we propose a novel attention-based local spatial and temporal relation discovery (ALSTRD) model. Our model firstly implements feature representation learning to effectively express latent input traffic information. Then, a local attention mechanism structure is established to model asynchronous dependencies of historical input data. Finally, another attention network and the Pearson Correlation Coefficient method are introduced to extract the elaborate influence of the historical traffic condition of neighboring roads on the future condition of the target road. The experiment results on several datasets demonstrate that our model achieves significant improvements in prediction accuracy compared to other baseline methods, which can be attributed to its ability to extract the fine-grained correlation among historical traffic data and capture the dynamic association between past and future data. In addition, the incorporation of attention mechanism and Pearson Correlation Coefficient promotes the model’s ability to elucidate spatiotemporal correlations among traffic data, thereby providing a more robust explanation.

The Quaternary sequence of the Nahr el Kebir, NW Syria: An important repository for evidence of Palaeolithic occupation and landscape evolution in the eastern Mediterranean

The third largest river in Syria, the Nahr el Kebir has a well-preserved record of river-terrace deposits that have produced substantial Palaeolithic artefact assemblages both from within the terrace deposits and from the land surfaces above and around them. At the Mediterranean coastline, the fluvial gravels interdigitate with raised shoreline terrace deposits, providing an insight into the temporal and climatic relations of both of these important geomorphological and morphostratigraphical archives, as well as their relationship with each other. New research is reported here on the Pleistocene geology and geomorphology of the Nahr el Kebir and the associated Palaeolithic archaeology, the latter having been reinterpreted based on reassessment of museum collections arising from earlier detailed work. Field visits revealed an additional, hitherto unrecognized low-level river terrace, whereas one of the previously recognized Palaeolithic levels can be shown to coincide with slope deposits that armour hilltops rather than representing a genuine fluvial formation. The new understanding of these geomorphological and sedimentary archives supports ideas that this corner of the Mediterranean has experienced unusually rapid uplift during the recent Quaternary, as a result of which the local rivers, including the Kebir, have deepened their valleys rapidly. Consequently, only the recent part of the Quaternary is recorded in the Kebir system and the ages envisaged previously for the terrace deposits and the Palaeolithic artefact assemblages were considerable overestimates in many cases, a finding that has significance for their correlation with those from the wider region. Reassessment of the Palaeolithic archaeology suggests a settlement history initially dominated by groups using handaxes, alongside simple core working (0.5–0.3 Ma), followed by a major change with the appearance of Levallois core working alongside handaxes, marking the transition to the early Middle Palaeolithic.

Weakly Supervised Video Anomaly Detection via Self-Guided Temporal Discriminative Transformer.

Weakly supervised video anomaly detection is generally formulated as a multiple instance learning (MIL) problem, where an anomaly detector learns to generate frame-level anomaly scores under the supervision of MIL-based video-level classification. However, most previous works suffer from two drawbacks: 1) they lack ability to model temporal relationships between video segments and 2) they cannot extract sufficient discriminative features to separate normal and anomalous snippets. In this article, we develop a weakly supervised temporal discriminative (WSTD) paradigm, that aims to leverage both temporal relation and feature discrimination to mitigate the above drawbacks. To this end, we propose a transformer-styled temporal feature aggregator (TTFA) and a self-guided discriminative feature encoder (SDFE). Specifically, TTFA captures multiple types of temporal relationships between video snippets from different feature subspaces, while SDFE enhances the discriminative powers of features by clustering normal snippets and maximizing the separability between anomalous snippets and normal centers in embedding space. Experimental results on three public benchmarks indicate that WSTD outperforms state-of-the-art unsupervised and weakly supervised methods, which verifies the superiority of the proposed method.

The caloric value of food intake structurally adjusts a neuronal mushroom body circuit mediating olfactory learning in Drosophila.

Associative learning enables the adaptive adjustment of behavioral decisions based on acquired, predicted outcomes. The valence of what is learned is influenced not only by the learned stimuli and their temporal relations, but also by prior experiences and internal states. In this study, we used the fruit fly Drosophila melanogaster to demonstrate that neuronal circuits involved in associative olfactory learning undergo restructuring during extended periods of low-caloric food intake. Specifically, we observed a decrease in the connections between specific dopaminergic neurons (DANs) and Kenyon cells at distinct compartments of the mushroom body. This structural synaptic plasticity was contingent upon the presence of allatostatin A receptors in specific DANs and could be mimicked optogenetically by expressing a light-activated adenylate cyclase in exactly these DANs. Importantly, we found that this rearrangement in synaptic connections influenced aversive, punishment-induced olfactory learning but did not impact appetitive, reward-based learning. Whether induced by prolonged low-caloric conditions or optogenetic manipulation of cAMP levels, this synaptic rearrangement resulted in a reduction of aversive associative learning. Consequently, the balance between positive and negative reinforcing signals shifted, diminishing the ability to learn to avoid odor cues signaling negative outcomes. These results exemplify how a neuronal circuit required for learning and memory undergoes structural plasticity dependent on prior experiences of the nutritional value of food.

Spatio-Temporal Graph Representation Learning for Fraudster Group Detection.

Motivated by potential financial gain, companies may hire fraudster groups to write fake reviews to either demote competitors or promote their own businesses. Such groups are considerably more successful in misleading customers, as people are more likely to be influenced by the opinion of a large group. To detect such groups, a common model is to represent fraudster groups' static networks, consequently overlooking the longitudinal behavior of a reviewer, thus, the dynamics of coreview relations among reviewers in a group. Hence, these approaches are incapable of excluding outlier reviewers, which are fraudsters intentionally camouflaging themselves in a group and genuine reviewers happen to coreview in fraudster groups. To address this issue, we propose "FGDT," a framework for "fraudster group detection through temporal relations." FGDT first capitalizes on the effectiveness of the HIN-recurrent neural network (RNN) in both reviewers' representation learning while capturing the collaboration between reviewers. The HIN-RNN models the coreview relations of reviewers in a group in a fixed time window of 28 days. We refer to this as spatial relation learning representation to signify the generalizability of this work to other networked scenarios. Then, we use an RNN on the spatial relations to predict the spatio-temporal relations of reviewers in the group. In the third step, a graph convolution network (GCN) refines the reviewers' vector representations using these predicted relations. These refined representations are then used to remove outlier reviewers. The average of the remaining reviewers' representation is then fed to a simple fully connected layer to predict if the group is a fraudster group or not. Exhaustive experiments of FGDT showed a 5% (4%), 12% (5%), and 12% (5%) improvement over three of the most recent approaches on precision, recall, and F1-value over the Yelp (Amazon) dataset, respectively.

Real-time early warning and the prediction of air pollutants for sustainable development in smart cities

Air quality forecasting is vital for sustainable development in smart cities and is essential to protect public health through early warning against air pollutants. This article presents a novel ensemble deep learning method able to capture spatiotemporal features on short-term time intervals and extract high data abstraction levels. This method uses a hierarchical framework with a new activation function called HT-Tanh to improve the prediction of air pollutant emissions in real time. As such, a residual neural network (ResNet) is integrated with a convolutional neural network (CNN) to deeply extract the temporal and spatial features from pollutant and meteorological data. The model network encodes the fully connected layer to fine-tune the CNN output and decodes the temporal prediction relations based on gate recurrent unit (GRU) to obtain more accurate results. The reliability of the proposed method is verified through a case study from Shantou City, China. The model performance is compared with the recently developed and traditional deep learning models. Results reveal that the proposed method has good precision in the multiscale air quality predictions. Compared with CNN-GRU, convolutional long short-term memory, and CNN models, the forecasting results of the proposed method are significantly improved. The presented method in this article can support the early warning systems in multiple regions for air pollution modeling studies.

Using Case Forms for Expressing Temporal Relations in Modern Turkish Language

Many languages, including modern Turkish, employ lexemes with spatial meaning to express temporal relations. However, the Turkish language stands out for its use of case forms of nouns both to describe the spatial position and movement of objects and to establish the temporal reference of actions and events. These differences are systemic, suggest that native Turkish speakers perceive spatial and temporal phenomena in a unique way, and determine the historical change in their models of time perception. They challenge the belief that spatial relations prevail over temporal ones in the perception of the surrounding world by peoples with nomadic roots but highlight the necessity of considering materials from other Turkic languages. The results obtained show that facts from different linguistic levels should be involved to analyze how temporal relations are conveyed.

Spatial and temporal properties of intra-operatively recorded spikes and high frequency oscillations in focal cortical dysplasia

ObjectiveFocal cortical dysplasias (FCD) are characterized by distinct interictal spike patterns and high frequency oscillations (HFOs; ripples: 80–250 Hz; fast ripples: 250–500 Hz) in the intra-operative electrocorticogram (ioECoG). We studied the temporal relation between intra-operative spikes and HFOs and their relation to resected tissue in people with FCD with a favorable outcome. MethodsWe included patients who underwent ioECoG-tailored epilepsy surgery with pathology confirmed FCD and long-term Engel 1A outcome. Spikes and HFOs were automatically detected and visually checked in 1-minute pre-resection-ioECoG. Channels covering resected and non-resected tissue were compared using a logistic mixed model, assessing event numbers, co-occurrence ratios, and time-based properties. ResultsWe found pre-resection spikes, ripples in respectively 21 and 20 out of 22 patients. Channels covering resected tissue showed high numbers of spikes and HFOs, and high ratios of co-occurring events. Spikes, especially with ripples, have a relatively sharp rising flank with a long descending flank and early ripple onset over resected tissue. ConclusionsA combined analysis of event numbers, ratios, and temporal relationships between spikes and HFOs may aid identifying epileptic tissue in epilepsy surgery. SignificanceThis study shows a promising method for clinically relevant properties of events, closely associated with FCD.

Language, responsibility, and agency: How UK's immigrant‐background young adults navigate linguistic heritage

AbstractThis study explores the intersection of agency and responsibility in navigating linguistic heritage among immigrant‐background young adults (18–29 years old) in the UK. Drawing on Judith Butler's theorisation, it provides a nuanced understanding of how individuals engage with and respond to the burdening aspects of heritage language maintenance within the complex context of globalised power structures and privileged dynamics. The research employs a qualitative approach, integrating journal entries, interview transcripts, and co‐analysis to elucidate participants' experiences. Through grounded theory development, the study identifies key themes such as: negotiating responsibility across temporalities; the sense of guilt, shame and indebtedness embedded in the process of language maintenance; and finding agency through renarration. The paper contributes to geographical thought on migration and language by going beyond discussions of family language policy and proficiencies and drawing attention to the multifaceted implications of navigating linguistic heritage within immigrant families and monolingual Britain across the instabilities of social and temporal relations and power‐laden positionings. It also highlights the repertoires of narrative and interrelational approaches to reframing heritage that immigrant‐background young adults may access as they formulate their identities with and against language expectations, norms, proficiencies and hegemonies.

Syntax, vocabulary, phraseology: Aspects of correlation and interaction

This article raises the issues of using an integrative approach with respect to the study of language units, the problems of interaction of syntax, vocabulary and phraseology, and their interpenetration and contact within the framework of the semantic structure of the language as a functioning system. We also consider syntactic units that are in the zone of abstract meanings that ensure the integrity and vital activity of the language. The diachronic aspect of the coverage of the questions posed contributed to the identification of two processes – grammaticalization and lexicalization, which stimulate mutual transitions and borderline positions between principal and auxiliary parts, predicative and nominative parts, mono and polysemantic parts, as well as stable and instable parts in the sentence. The syntactic phraseomodels, analyzed in the article, are frequent in the sphere of hypotaxis when expressing adversarial-concessional, temporal and causal relations. It is especially important that the language has the integrativity of abstract vocabulary and relational semantics, designed to cement and structure speech activity. The syntactic phraseomodels under consideration complicate relational semantics with modal or expressive meanings. The modal component is usually realized through the vocabulary of anthropocentric content, covering the entire spectrum of human speech-psychomental activity, it is represented by separate semantic fields and lexical-semantic groups. Within the framework of expressive contexts, as essential elements are the lexical content of pretexts and the presence of emotive components in their composition. In these phrases, one can see a cause-and-effect relationship between two propositions, the real and potential ones in the subordinate part, although such models often move into the sphere of intensity semantics and the zone of emotiveness. The syntactic phraseomodels under consideration and those of them that have acquired a stable appearance are prone to parceling, budding from the base part and functioning more autonomously in the text. Thus, semantic-syntactic relations interact with abstract vocabulary and have a certain range of emotional-evaluative and modal meanings.

Parts of Night and Day in the Quranic Context

This study examines the parts of night and day in the Quranic context, discussing the Quranic division of temporal relations associated with night and day and the specific portions thereof mentioned in the Quran. It reveals that time, in the Quranic understanding, warrants attention and organization to ensure that social and human relations are based on respecting and managing time, avoiding its wastage on fruitless pursuits. The study also illustrates that as hours and times pass, with their consecutive regularity, they alert to the stages of night and day and their constituent parts. These parts gradually unfold the span of life, fostering perpetual awareness that enables individuals to sense the passage of time. It encourages individuals to hold themselves accountable for their life relationships and engagements, understanding how they spend their time and to what extent they overcome the constraints of time. This is delineated through the detailed exploration and focal points addressed in this research. 
 Originality/value: The study focuses on a deep analysis of the parts of night and day in the Quranic context, shedding light on the importance of respecting and organizing time in individual and social life. Additionally, it offers a fresh perspective on understanding time in Islam and how Quranic concepts can effectively guide human behavior in investing and managing their time.

Multi-Granularity Aggregation with Spatiotemporal Consistency for Video-Based Person Re-Identification.

Video-based person re-identification (ReID) aims to exploit relevant features from spatial and temporal knowledge. Widely used methods include the part- and attention-based approaches for suppressing irrelevant spatial-temporal features. However, it is still challenging to overcome inconsistencies across video frames due to occlusion and imperfect detection. These mismatches make temporal processing ineffective and create an imbalance of crucial spatial information. To address these problems, we propose the Spatiotemporal Multi-Granularity Aggregation (ST-MGA) method, which is specifically designed to accumulate relevant features with spatiotemporally consistent cues. The proposed framework consists of three main stages: extraction, which extracts spatiotemporally consistent partial information; augmentation, which augments the partial information with different granularity levels; and aggregation, which effectively aggregates the augmented spatiotemporal information. We first introduce the consistent part-attention (CPA) module, which extracts spatiotemporally consistent and well-aligned attentive parts. Sub-parts derived from CPA provide temporally consistent semantic information, solving misalignment problems in videos due to occlusion or inaccurate detection, and maximize the efficiency of aggregation through uniform partial information. To enhance the diversity of spatial and temporal cues, we introduce the Multi-Attention Part Augmentation (MA-PA) block, which incorporates fine parts at various granular levels, and the Long-/Short-term Temporal Augmentation (LS-TA) block, designed to capture both long- and short-term temporal relations. Using densely separated part cues, ST-MGA fully exploits and aggregates the spatiotemporal multi-granular patterns by comparing relations between parts and scales. In the experiments, the proposed ST-MGA renders state-of-the-art performance on several video-based ReID benchmarks (i.e., MARS, DukeMTMC-VideoReID, and LS-VID).

Automatic Identification of Pangolin Behavior Using Deep Learning Based on Temporal Relative Attention Mechanism.

With declining populations in the wild, captive rescue and breeding have become one of the most important ways to protect pangolins from extinction. At present, the success rate of artificial breeding is low, due to the insufficient understanding of the breeding behavior characteristics of pangolins. The automatic recognition method based on machine vision not only monitors for 24 h but also reduces the stress response of pangolins. This paper aimed to establish a temporal relation and attention mechanism network (Pangolin breeding attention and transfer network, PBATn) to monitor and recognize pangolin behaviors, including breeding and daily behavior. There were 11,476 videos including breeding behavior and daily behavior that were divided into training, validation, and test sets. For the training set and validation set, the PBATn network model had an accuracy of 98.95% and 96.11%, and a loss function value of 0.1531 and 0.1852. The model is suitable for a 2.40 m × 2.20 m (length × width) pangolin cage area, with a nest box measuring 40 cm × 30 cm × 30 cm (length × width × height) positioned either on the left or right side inside the cage. A spherical night-vision monitoring camera was installed on the cage wall at a height of 2.50 m above the ground. For the test set, the mean Average Precision (mAP), average accuracy, average recall, average specificity, and average F1 score were found to be higher than SlowFast, X3D, TANet, TSN, etc., with values of 97.50%, 99.17%, 97.55%, 99.53%, and 97.48%, respectively. The recognition accuracies of PBATn were 94.00% and 98.50% for the chasing and mounting breeding behaviors, respectively. The results showed that PBATn outperformed the baseline methods in all aspects. This study shows that the deep learning system can accurately observe pangolin breeding behavior and it will be useful for analyzing the behavior of these animals.

Identifying temporal pathways using biomarkers in the presence of latent non-Gaussian components.

Time-series data collected from a network of random variables are useful for identifying temporal pathways among the network nodes. Observed measurements may contain multiple sources of signals and noises, including Gaussian signals of interest and non-Gaussian noises, including artifacts, structured noise, and other unobserved factors (eg, genetic risk factors, disease susceptibility). Existing methods, including vector autoregression (VAR) and dynamic causal modeling do not account for unobserved non-Gaussian components. Furthermore, existing methods cannot effectively distinguish contemporaneous relationships from temporal relations. In this work, we propose a novel method to identify latent temporal pathways using time-series biomarker data collected from multiple subjects. The model adjusts for the non-Gaussian components and separates the temporal network from the contemporaneous network. Specifically, an independent component analysis (ICA) is used to extract the unobserved non-Gaussian components, and residuals are used to estimate the contemporaneous and temporal networks among the node variables based on method of moments. The algorithm is fast and can easily scale up. We derive the identifiability and the asymptotic properties of the temporal and contemporaneous networks. We demonstrate superior performance of our method by extensive simulations and an application to a study of attention-deficit/hyperactivity disorder (ADHD), where we analyze the temporal relationships between brain regional biomarkers. We find that temporal network edges were across different brain regions, while most contemporaneous network edges were bilateral between the same regions and belong to a subset of the functional connectivity network.

Multimodal learning for temporal relation extraction in clinical texts.

This study focuses on refining temporal relation extraction within medical documents by introducing an innovative bimodal architecture. The overarching goal is to enhance our understanding of narrative processes in the medical domain, particularly through the analysis of extensive reports and notes concerning patient experiences. Our approach involves the development of a bimodal architecture that seamlessly integrates information from both text documents and knowledge graphs. This integration serves to infuse common knowledge about events into the temporal relation extraction process. Rigorous testing was conducted on diverse clinical datasets, emulating real-world scenarios where the extraction of temporal relationships is paramount. The performance of our proposed bimodal architecture was thoroughly evaluated across multiple clinical datasets. Comparative analyses demonstrated its superiority over existing methods reliant solely on textual information for temporal relation extraction. Notably, the model showcased its effectiveness even in scenarios where not provided with additional information. The amalgamation of textual data and knowledge graph information in our bimodal architecture signifies a notable advancement in the field of temporal relation extraction. This approach addresses the critical need for a more profound understanding of narrative processes in medical contexts. In conclusion, our study introduces a pioneering bimodal architecture that harnesses the synergy of text and knowledge graph data, exhibiting superior performance in temporal relation extraction from medical documents. This advancement holds significant promise for improving the comprehension of patients' healthcare journeys and enhancing the overall effectiveness of extracting temporal relationships in complex medical narratives.

READ-PVLA: Recurrent Adapter with Partial Video-Language Alignment for Parameter-Efficient Transfer Learning in Low-Resource Video-Language Modeling

Fully fine-tuning pretrained large-scale transformer models has become a popular paradigm for video-language modeling tasks, such as temporal language grounding and video-language summarization. With a growing number of tasks and limited training data, such full fine-tuning approach leads to costly model storage and unstable training. To overcome these shortcomings, we introduce lightweight adapters to the pre-trained model and only update them at fine-tuning time. However, existing adapters fail to capture intrinsic temporal relations among video frames or textual words. Moreover, they neglect the preservation of critical task-related information that flows from the raw video-language input into the adapter’s low-dimensional space. To address these issues, we first propose a novel REcurrent ADapter (READ) that employs recurrent computation to enable temporal modeling capability. Second, we propose Partial Video-Language Alignment (PVLA) objective via the use of partial optimal transport to maintain task-related information flowing into our READ modules. We validate our READ-PVLA framework through extensive experiments where READ-PVLA significantly outperforms all existing fine-tuning strategies on multiple low-resource temporal language grounding and video-language summarization benchmarks.

F³-Pruning: A Training-Free and Generalized Pruning Strategy towards Faster and Finer Text-to-Video Synthesis

Recently Text-to-Video (T2V) synthesis has undergone a breakthrough by training transformers or diffusion models on large-scale datasets. Nevertheless, inferring such large models incurs huge costs. Previous inference acceleration works either require costly retraining or are model-specific. To address this issue, instead of retraining we explore the inference process of two mainstream T2V models using transformers and diffusion models. The exploration reveals the redundancy in temporal attention modules of both models, which are commonly utilized to establish temporal relations among frames. Consequently, we propose a training-free and generalized pruning strategy called F3-Pruning to prune redundant temporal attention weights. Specifically, when aggregate temporal attention values are ranked below a certain ratio, corresponding weights will be pruned. Extensive experiments on three datasets using a classic transformer-based model CogVideo and a typical diffusion-based model Tune-A-Video verify the effectiveness of F3-Pruning in inference acceleration, quality assurance and broad applicability.

Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentation

Recently, video object segmentation (VOS) referred by multi-modal signals, e.g., language and audio, has evoked increasing attention in both industry and academia. It is challenging for exploring the semantic alignment within modalities and the visual correspondence across frames. However, existing methods adopt separate network architectures for different modalities, and neglect the inter-frame temporal interaction with references. In this paper, we propose MUTR, a Multi-modal Unified Temporal transformer for Referring video object segmentation. With a unified framework for the first time, MUTR adopts a DETR-style transformer and is capable of segmenting video objects designated by either text or audio reference. Specifically, we introduce two strategies to fully explore the temporal relations between videos and multi-modal signals. Firstly, for low-level temporal aggregation before the transformer, we enable the multi-modal references to capture multi-scale visual cues from consecutive video frames. This effectively endows the text or audio signals with temporal knowledge and boosts the semantic alignment between modalities. Secondly, for high-level temporal interaction after the transformer, we conduct inter-frame feature communication for different object embeddings, contributing to better object-wise correspondence for tracking along the video. On Ref-YouTube-VOS and AVSBench datasets with respective text and audio references, MUTR achieves +4.2% and +8.7% J&F improvements to state-of-the-art methods, demonstrating our significance for unified multi-modal VOS. Code is released at https://github.com/OpenGVLab/MUTR.

MCN4Rec: Multi-level Collaborative Neural Network for Next Location Recommendation

Next location recommendation plays an important role in various location-based services, yielding great value for both users and service providers. Existing methods usually model temporal dependencies with explicit time intervals or learn representation from customized point of interest (POI) graphs with rich context information to capture the sequential patterns among POIs. However, this problem is perceptibly complex, because various factors, e.g., users’ preferences, spatial locations, time contexts, activity category semantics, and temporal relations, need to be considered together, while most studies lack sufficient consideration of the collaborative signals. Toward this goal, we propose a novel M ulti-Level C ollaborative Neural N etwork for next location Rec ommendation (MCN4Rec). Specifically, we design a multi-level view representation learning with level-wise contrastive learning to collaboratively learn representation from local and global perspectives to capture complex heterogeneous relationships among user, POI, time, and activity categories. Then, a causal encoder-decoder is applied to the learned representations of check-in sequences to recommend the next location. Extensive experiments on four real-world check-in mobility datasets demonstrate that our model significantly outperforms the existing state-of-the-art baselines for the next location recommendation. Ablation study further validates the benefits of the collaboration of the designed sub-modules. The source code is available at https://github.com/quai-mengxiang/MCN4Rec .

Developing a multivariate time series forecasting framework based on stacked autoencoders and multi-phase feature

Time series forecasting across different domains has received massive attention as it eases intelligent decision-making activities. Recurrent neural networks and various deep learning algorithms have been applied to modeling and forecasting multivariate time series data. Due to intricate non-linear patterns and significant variations in the randomness of characteristics across various categories of real-world time series data, achieving effectiveness and robustness simultaneously poses a considerable challenge for specific deep-learning models. We have proposed a novel prediction framework with a multi-phase feature selection technique, a long short-term memory-based autoencoder, and a temporal convolution-based autoencoder to fill this gap. The multi-phase feature selection is applied to retrieve the optimal feature selection and optimal lag window length for different features. Moreover, the customized stacked autoencoder strategy is employed in the model. The first autoencoder is used to resolve the random weight initialization problem. Additionally, the second autoencoder models the temporal relation between non-linear correlated features with convolution networks and recurrent neural networks.Finally, the model's ability to generalize, predict accurately, and perform effectively is validated through experimentation with three distinct real-world time series datasets. In this study, we conducted experiments on three real-world datasets: Energy Appliances, Beijing PM2.5 Concentration, and Solar Radiation. The Energy Appliances dataset consists of 29 attributes with a training size of 15,464 instances and a testing size of 4239 instances. For the Beijing PM2.5 Concentration dataset, there are 18 attributes, with 34,952 instances in the training set and 8760 instances in the testing set. The Solar Radiation dataset comprises 11 attributes, with 22,857 instances in the training set and 9797 instances in the testing set. The experimental setup involved evaluating the performance of forecasting models using two distinct error measures: root mean square error and mean absolute error. To ensure robust evaluation, the errors were calculated at the identical scale of the data. The results of the experiments demonstrate the superiority of the proposed model compared to existing models, as evidenced by significant advantages in various metrics such as mean squared error and mean absolute error. For PM2.5 air quality data, the proposed model's mean absolute error is 7.51 over 12.45, about ∼40% improvement. Similarly, the mean square error for the dataset is improved from 23.75 to 11.62, which is ∼51%of improvement. For the solar radiation dataset, the proposed model resulted in ∼34.7% improvement in means squared error and ∼75% in mean absolute error. The recommended framework demonstrates outstanding capabilities in generalization and outperforms datasets spanning multiple indigenous domains.