Characteristics Of Time Series Research Articles

An Efficient Water Quality Prediction and Assessment Method Based on the Improved Deep Belief Network—Long Short-Term Memory Model

The accuracy of water quality prediction and assessment has always been the focus of environmental departments. However, due to the high complexity of water systems, existing methods struggle to capture the future internal dynamic changes in water quality based on current data. In view of this, this paper proposes a data-driven approach to combine an improved deep belief network (DBN) and long short-term memory (LSTM) network model for water quality prediction and assessment, avoiding the complexity of constructing a model of the internal mechanism of water quality. Firstly, using Gaussian Restricted Boltzmann Machines (GRBMs) to construct a DBN, the model has a better ability to extract continuous data features compared to classical DBN. Secondly, the extracted time-series data features are input into the LSTM network to improve predicting accuracy. Finally, due to prediction errors, noise that randomly follows the Gaussian distribution is added to the assessment results based on the predicted values, and the probability of being at the current water quality level in the future is calculated through multiple evolutionary computations to complete the water quality assessment. Numerical experiments have shown that our proposed algorithm has a greater accuracy compared to classical algorithms in challenging scenarios.

Predicting air quality index using attention hybrid deep learning and quantum-inspired particle swarm optimization

Air pollution poses a significant threat to the health of the environment and human well-being. The air quality index (AQI) is an important measure of air pollution that describes the degree of air pollution and its impact on health. Therefore, accurate and reliable prediction of the AQI is critical but challenging due to the non-linearity and stochastic nature of air particles. This research aims to propose an AQI prediction hybrid deep learning model based on the Attention Convolutional Neural Networks (ACNN), Autoregressive Integrated Moving Average (ARIMA), Quantum Particle Swarm Optimization (QPSO)-enhanced-Long Short-Term Memory (LSTM) and XGBoost modelling techniques. Daily air quality data were collected from the official Seoul Air registry for the period 2021 to 2022. The data were first preprocessed through the ARIMA model to capture and fit the linear part of the data and followed by a hybrid deep learning architecture developed in the pretraining–finetuning framework for the non-linear part of the data. This hybrid model first used convolution to extract the deep features of the original air quality data, and then used the QPSO to optimize the hyperparameter for LSTM network for mining the long-terms time series features, and the XGBoost model was adopted to fine-tune the final AQI prediction model. The robustness and reliability of the resulting model were assessed and compared with other widely used models and across meteorological stations. Our proposed model achieves up to 31.13% reduction in MSE, 19.03% reduction in MAE and 2% improvement in R-squared compared to the best appropriate conventional model, indicating a much stronger magnitude of relationships between predicted and actual values. The overall results show that the attentive hybrid deep Quantum inspired Particle Swarm Optimization model is more feasible and efficient in predicting air quality index at both city-wide and station-specific levels.

Fake User Detection Based on Multi-Model Joint Representation

The existing deep learning-based detection of fake information focuses on the transient detection of news itself. Compared to user category profile mining and detection, transient detection is prone to higher misjudgment rates due to the limitations of insufficient temporal information, posing new challenges to social public opinion monitoring tasks such as fake user detection. This paper proposes a multimodal aggregation portrait model (MAPM) based on multi-model joint representation for social media platforms. It constructs a deep learning-based multimodal fake user detection framework by analyzing user behavior datasets within a time retrospective window. It integrates a pre-trained Domain Large Model to represent user behavior data across multiple modalities, thereby constructing a high-generalization implicit behavior feature spectrum for users. In response to the tendency of existing fake user behavior mining to neglect time-series features, this study introduces an improved network called Sequence Interval Detection Net (SIDN) based on Sequence to Sequence (seq2seq) to characterize time interval sequence behaviors, achieving strong expressive capabilities for detecting fake behaviors within the time window. Ultimately, the amalgamation of latent behavioral features and explicit characteristics serves as the input for spectral clustering in detecting fraudulent users. The experimental results on Weibo real dataset demonstrate that the proposed model outperforms the detection utilizing explicit user features, with an improvement of 27.0% in detection accuracy.

Dynamic risk stratification of worsening heart failure using a deep learning-enabled implanted ambulatory single-lead electrocardiogram.

Implantable loop recorders (ILRs) provide continuous single-lead ambulatory electrocardiogram (aECG) monitoring. Whether these aECGs could be used to identify worsening heart failure (HF) is unknown. We linked ILR aECG from Medtronic device database to the left ventricular ejection fraction (LVEF) measurements in Optum® de-identified electronic health record dataset. We trained an artificial intelligence (AI) algorithm [aECG-convolutional neural network (CNN)] on a dataset of 35 741 aECGs from 2247 patients to identify LVEF ≤ 40% and assessed its performance using the area under the receiver operating characteristic curve. Ambulatory electrocardiogram-CNN was then used to identify patients with increasing risk of HF hospitalization in a real-world cohort of 909 patients with prior HF diagnosis. This dataset provided 12 467 follow-up monthly evaluations, with 201 HF hospitalizations. For every month, time-series features from these predictions were used to categorize patients into high- and low-risk groups and predict HF hospitalization in the next month. The risk of HF hospitalization in the next 30 days was significantly higher in the cohort that aECG-CNN identified as high risk [hazard ratio (HR) 1.89; 95% confidence interval (CI) 1.28-2.79; P = 0.001] compared with low risk, even after adjusting patient demographics (HR 1.88; 95% CI 1.27-2.79 P = 0.002). An AI algorithm trained to detect LVEF ≤40% using ILR aECGs can also readily identify patients at increased risk of HF hospitalizations by monitoring changes in the probability of HF over 30 days.

Evaluation is key: a survey on evaluation measures for synthetic time series

Synthetic data generation describes the process of learning the underlying distribution of a given real dataset in a model, which is, in turn, sampled to produce new data objects still adhering to the original distribution. This approach often finds application where circumstances limit the availability or usability of real-world datasets, for instance, in health care due to privacy concerns. While image synthesis has received much attention in the past, time series are key for many practical (e.g., industrial) applications. To date, numerous different generative models and measures to evaluate time series syntheses have been proposed. However, regarding the defining features of high-quality synthetic time series and how to quantify quality, no consensus has yet been reached among researchers. Hence, we propose a comprehensive survey on evaluation measures for time series generation to assist users in evaluating synthetic time series. For one, we provide brief descriptions or - where applicable - precise definitions. Further, we order the measures in a taxonomy and examine applicability and usage. To assist in the selection of the most appropriate measures, we provide a concise guide for fast lookup. Notably, our findings reveal a lack of a universally accepted approach for an evaluation procedure, including the selection of appropriate measures. We believe this situation hinders progress and may even erode evaluation standards to a “do as you like”-approach to synthetic data evaluation. Therefore, this survey is a preliminary step to advance the field of synthetic data evaluation.

Time and frequency-domain feature fusion network for multivariate time series classification

Multivariate time series classification is a significant research topic in the realm of data mining, which encompasses a wide array of practical applications in domains such as healthcare, energy systems, and traffic. The complex temporal and spatial dependencies inherent in multivariate time series pose challenges to classification tasks. Previous studies have usually focused on the time-domain information of multivariate time series. However, achieving accurate classification using only the time-domain information may be difficult. To address this challenge, a time and frequency-domain feature fusion network (TF-Net) for multivariate time series classification is proposed in this paper. Our model contains two modules, the time-domain module and the frequency-domain module. The time-domain module is used to capture the time-domain features of multivariate time series. It is constructed using CNNs and GCNs, enabling the capture of both temporal and spatial dependencies within the time-domain. The frequency-domain module is used to capture the frequency-domain features of multivariate time series data. In this module, we treat the frequency-domain features as images and innovatively transform the multivariate time series classification task into an image classification task. Our method is able to classify multivariate time series from both the time-domain and frequency-domain, which provides a new perspective for multivariate time series analysis. We conduct extensive experiments on the UAE archive, and the experimental results show that our model achieves the best performance compared to the ten state-of-the-art methods.

A Bayesian adversarial probsparse Transformer model for long-term remaining useful life prediction

Long-term remaining useful life (RUL) prediction is essential for the maintenance of safety-crucial engineering assets. Deep learning (DL) models, especially Transformer-based models have achieved outstanding performance in long-term RUL prediction. However, existing Transformer models neglect the impact of discrepancy loss in model training. The accumulation of the discrepancy loss during the inference will hamper the generalization of prediction model, resulting in an overfitting problem. To address the problem, this paper proposes a Bayesian Adversarial Probsparse Transformer (BAPT) model for long-term RUL prediction. Firstly, the adversarial learning method is leveraged to mitigate the impact of accumulated discrepancy loss caused by varying working conditions in long-term prediction, thus diminishing the error accumulation. Secondly, the Probsparse multi-head attention is adopted to enhance the efficiency of feature extraction. The Probsparse multi-head attention focuses on the significant degradation features in long time-series to reduce the computation complexity. Lastly, the Bayesian neural network is introduced to quantify the uncertainty in RUL prediction. The effectiveness of the proposed model is verified using two commercial aircraft turbofan engine datasets. The results indicate that BAPT model for long-term RUL prediction demonstrates better performance than the existing state-of-the-art models.

Power Load Forecast Based on CS-LSTM Neural Network

Load forecast is the foundation of power system operation and planning. The forecast results can guide the power system economic dispatch and security analysis. In order to improve the accuracy of load forecast, this paper proposes a forecasting model based on the combination of the cuckoo search (CS) algorithm and the long short-term memory (LSTM) neural network. Load data are specific data with time series characteristics and periodicity, and the LSTM algorithm can control the information added or discarded through the forgetting gate, so as to realize the function of forgetting or memorizing. Therefore, the use of the LSTM algorithm for load forecast is more effective. The CS algorithm can perform global search better and does not easily fall into local optima. The CS-LSTM forecasting model, where CS algorithm is used to optimize the hyper-parameters of the LSTM model, has a better forecasting effect and is more feasible. Simulation results show that the CS-LSTM model has higher forecasting accuracy than the standard LSTM model, the PSO-LSTM model, and the GA-LSTM model.

Spatio-temporal prediction of total energy consumption in multiple regions using explainable deep neural network

Energy consumption forecasting is essential for energy system integration and management. However, existing studies mainly focus on temporal features of energy consumption, which neglects the spatial correlation of variables with time information. Capturing the spatio-temporal relationships helps to improve forecasting accuracy and further promote energy dispatch. To tackle this problem, an explainable Convolutional Neural Network-Long Short Term Memory forecasting model is employed to effectively predict the total energy consumption by capturing the spatial and temporal features of multivariate time series. In the model, the autoencoder is used to achieve the nonlinear dimensionality reduction and transfer the data to a low-dimensional space. Furthermore, a Convolutional Neural Network is used to extract more effective features from the decoded data, and long short-term memory is employed to identify the temporal dependencies between extracted features and total energy consumption. Shapley additive explanation is introduced to interpret the outputs of the black-box model. The superior performance of the proposed method with high accuracy and good adaptability is verified by the comparisons with conventional forecasting models. This method provides an insight into the regional energy consumption analyzing contributions of weather variables to energy consumption, which helps administers in understanding regional energy performance for enhancing energy efficiency.

The Impact of the Three-Dimensional Structure of a Subduction Zone on Time-dependent Crustal Deformation Measured by HR-GNSS

Accurately modeling time-dependent coseismic crustal deformation as observed on high-rate Global Navigation Satellite System (HR-GNSS) lends insight into earthquake source processes and improves local earthquake and tsunami early warning algorithms. Currently, time-dependent crustal deformation modeling relies most frequently on simplified 1D radially symmetric Earth models. However, for shallow subduction zone earthquakes, even low-frequency shaking is likely affected by the many strongly heterogeneous structures such as the subducting slab, mantle wedge, and the overlying crustal structure. We demonstrate that including 3D structure improves the estimation of key features of coseismic HR-GNSS time series, such as the peak ground displacement (PGD), the time to PGD (tPGD), static displacements (SD), and waveform cross-correlation values. We computed synthetic 1D and 3D, 0.25 Hz and 0.5 Hz waveforms at HR-GNSS stations for four M7.3+ earthquakes in Japan using MudPy and SW4, respectively. From these synthetics, we computed intensity-measure residuals between the synthetic and observed GNSS waveforms. Comparing 1D and 3D residuals, we observed that the 3D simulations show better fits to the PGD and SD in the observed waveforms than the 1D simulations for both 0.25 Hz and 0.5 Hz simulations. We find that the reduction in PGD residuals in the 3D simulations is a combined effect of both shallow and deep 3D structures; hence incorporating only the upper 30 km of 3D structure will still improve the fit to the observed PGD values. Our results demonstrate that 3D simulations significantly improve models of GNSS waveform characteristics and will not only help understand the underlying processes, but also improve local tsunami warning.

CIVET: Exploring Compact Index for Variable-Length Subsequence Matching on Time Series

Nowadays the demands for managing and analyzing substantially increasing collections of time series are becoming more challenging. Subsequence matching, as a core subroutine in time series analysis, has drawn significant research attention. Most of the previous works only focus on matching the subsequences with equal length to the query. However, many scenarios require support for efficient variable-length subsequence matching. In this paper, we propose a new representation, Uniform Piecewise Aggregate Approximation (UPAA) with the capability of aligning features for variable-length time series while remaining the lower bounding property. Based on UPAA, we present a compact index structure by grouping adjacent subsequences and similar subsequences respectively. Moreover, we propose an index pruning algorithm and a data filtering strategy to efficiently support variable-length subsequence matching without false dismissals. The experiments conducted on both real and synthetic datasets demonstrate that our approach achieves considerably better efficiency, scalability, and effectiveness than existing approaches.

Turbulence Intermittency Effects on the Initiation Threshold of Sediment Motion in Natural Waters

AbstractThe initiation threshold of sediment motion, a key component in quantifying sediment transport, has potential link to intermittent turbulence bursts. Here, we elaborated in situ experiments on coastal sea bottom covered with cohesive sediments, to obtain intermittency parameters. The variable “waiting time” between turbulence bursts was utilized to capture the occurrence of sediment initiation events in natural waters. A relationship between waiting time and shear stress reveals the different intermittency features of sediment flux time series before and after reaching the threshold, which can be used to determine the initiation threshold of sediment motion. Multi‐site results demonstrate the limitations of traditional empirical formulae for fine‐grained sediments, where cohesiveness becomes more pronounced as grain size decreases and the deviation can reach 600%. The empirical formula was modified using grain size, and the modified calculations were in good agreement with observed values, which will greatly assist in sediment transport and geomorphology model predictions.

A novel framework for predicting non-stationary production time series of shale gas based on BiLSTM-RF-MPA deep fusion model

Shale gas, as an environmentally friendly fossil energy resource, has gained significant commercial development and shows immense potential. However, accurately predicting shale gas production faces substantial challenges due to the complex law of decline, nonlinear and non-stationary features in production data, which greatly repair the robustness of current models in predicting shale gas production time series. To address these challenges and improve accuracy in production forecasting, this paper introduces a novel and innovative approach: a hybrid proxy model that combines the bi-directional long short-term memory (BiLSTM) neural network and random forest (RF) through deep learning. The BiLSTM neural network is adept at capturing long-term dependencies, making it suitable for understanding the intricate relationships between input and output variables in shale gas production. On the other hand, RF serves a dual purpose: reducing model variance and addressing the concept drift problem that arises in non-stationary time series predictions made by BiLSTM. By integrating these two models, the hybrid approach effectively captures the inherent dependencies present in long and nonstationary production time series, thereby reducing model uncertainty. Furthermore, the combination of BiLSTM and RF is optimized using the recently-proposed marine predators algorithm (MPA) to fine-tune hyperparameters and enhance the overall performance of the proxy model. The results demonstrate that the proposed BiLSTM-RF-MPA model achieves higher prediction accuracy and demonstrates stronger generalization capabilities by effectively handling the complex nonlinear and non-stationary characteristics of shale gas production time series. Compared to other models such as LSTM, BiLSTM, and RF, the proposed model exhibits superior fitting and prediction performance, with an average improvement in performance indicators exceeding 20%. This innovative framework provides valuable insights for forecasting the complex production performance of unconventional oil and gas reservoirs, which sheds light on the development of data-driven proxy models in the field of subsurface energy utilization.

Prediction of the axial strain development curve of soft clay under dynamic triaxial conditions based on LSTM

The dynamic triaxial test results of soft clay show that a long-term cyclic traffic load lead to significant deformation of soft clay. Considering that the dynamic triaxial test data of soft clay under cyclic loading have apparent time series characteristics, the long short-term memory (LSTM) model can be used to predict the deformation law of soft clay. Through a dynamic triaxial test of Shanghai soft clay, time series data, such as axial stress, pore water pressure, and axial strain were obtained to establish the sample data of the LSTM model. The LSTM model is then used to predict the axial strain development curve. The results showed that the LSTM model had a good fitting precision for the sample data under different dynamic stresses and frequencies, and its accuracy was above 97% in all conditions. The prediction results can reproduce the fluctuation law of the axial strain under cyclic loads and can be used to study elastic and plastic strains. The research results can provide technical support for disaster prevention and control in soft clay.

Recognition of EEG-based movement intention combined with channel selection adopting deep learning methods

Brain-computer interface (BCI) is an emerging technology which provides a road to control communication and external devices. Electroencephalogram (EEG)-based motor imagery (MI) tasks recognition has important research significance for stroke, disability and others in BCI fields. However, enhancing the classification performance for decoding MI-related EEG signals presents a significant challenge, primarily due to the variability across different subjects and the presence of irrelevant channels. To address this issue, a novel hybrid structure is developed in this study to classify the MI tasks via deep separable convolution network (DSCNN) and bidirectional long short-term memory (BLSTM). First, the collected time-series EEG signals are initially processed into a matrix grid. Subsequently, data segments formed using a sliding window strategy are inputted into proposed DSCNN model for feature extraction (FE) across various dimensions. And, the spatial-temporal features extracted are then fed into the BLSTM network, which further refines vital time-series features to identify five distinct types of MI-related tasks. Ultimately, the evaluation results of our method demonstrate that the developed model achieves a 98.09% accuracy rate on the EEGMMIDB physiological datasets over a 4-second period for MI tasks by adopting full channels, outperforming other existing studies. Besides, the results of the five evaluation indexes of Recall, Precision, Test-auc, and F1-score also achieve 97.76%, 97.98%, 98.63% and 97.86%, respectively. Moreover, a Gradient-class Activation Mapping (GRAD-CAM) visualization technique is adopted to select the vital EEG channels and reduce the irrelevant information. As a result, we also obtained a satisfying outcome of 94.52% accuracy with 36 channels selected using the Grad-CAM approach. Our study not only provides an optimal trade-off between recognition rate and number of channels with half the number of channels reduced, but also it can also advances practical application research in the field of BCI rehabilitation medicine, effectively.

Hydroclimatic scenario generation using two-stage stochastic simulation framework

Climate change poses significant challenges for decision-making processes across a range of sectors. From the water resources planning and management perspective, the interest is often in evaluating the performance of a water supply system in a future state considering the potential changes in rainfall and streamflow characteristics. With observed climate change signals, scenario-based projections of rainfall and streamflow simulations are crucial for evaluating the potential impacts of climate change on water resource systems. Given the complexity of the existing approaches, their applications for generating scenario-based projections of streamflow and rainfall are limited. We developed a non-parametric bootstrapping approach, NPScnGen, for future scenario generation of any hydroclimatic variable. The developed approach is flexible and can be used with any physical hydrological or data-driven stochastic model that provides simulations of hydroclimatic variables of interest for the historical climate condition. In NPScnGen, samples of any set of time-series characteristics, such as mean and standard deviation, are generated from a multivariate Gaussian process for the considered scenario, and then bootstrapping is performed to select the closest sample from the historical simulation of that hydroclimatic variable. We have also proposed a modified wavelet-based model, Wavelet-HMM, and used that model to synthetically generate historical climate time-series as a baseline. We present the application of the developed framework consisting of historical climate simulation and future climate projection approaches on rainfall and streamflow datasets for the Tampa Bay region in Florida.Plain Language Summary: Water resources managers require a wide range of hypothetical but potential changes in hydroclimatic variables such as streamflow and rainfall to evaluate the sustenance of water supply systems in future. Existing scenario generation approaches are limited by either the complexity of statistical models or dependency on climate models which have their own limitations. In such a scenario, the developed non-parametric scenario generation framework in this study, NPScnGen, can be very useful. The developed framework can be applied with any sophisticated time-series generation model that can generate synthetic hydroclimatic traces for baseline climate condition, and it is also flexible in generating a wide range of potential climate change scenarios. We show the application of the framework on both streamflow and rainfall datasets.

A Novel Spatiotemporal Periodic Polynomial Model for Predicting Road Traffic Speed

Accurate and fast traffic prediction is the data-based foundation for achieving traffic control and management, and the accuracy of prediction results will directly affect the effectiveness of traffic control and management. This paper proposes a new spatiotemporal periodic polynomial model for road traffic, which integrates the temporal, spatial, and periodic features of speed time series and can effectively handle the nonlinear mapping relationship from input to output. In terms of the model, we establish a road traffic speed prediction model based on polynomial regression. In terms of spatial feature extraction methods, we introduce a maximum mutual information coefficient spatial feature extraction method. In terms of periodic feature extraction methods, we introduce a periodic trend modeling method into the prediction of speed time series, and effective fusion is carried out. Four strategies are evaluated based on the Guangzhou road speed dataset: a univariate polynomial model, a spatiotemporal polynomial model, a periodic polynomial model, and a spatiotemporal periodic polynomial model. The test results show that the three methods proposed in this article can effectively improve prediction accuracy. Comparing the spatiotemporal periodic polynomial model with multiple machine learning models and deep learning models, the prediction accuracy is improved by 5.94% compared to the best feedforward neural network. The research in this article can effectively deal with the temporal, spatial, periodic, and nonlinear characteristics of speed prediction, and to a certain extent, improve the accuracy of speed prediction.

The Impact of Operating Cash Flow on Corporate Profitability: Amman Stock Exchange case study

The problem of the study is the challenges faced by the growth of different sectors, which requires the identification of an appropriate operating cash flow policy to increase corporate Profitability. Unbalanced policies followed in operating cash flows can negatively affect companies' Profitability. The study's objective is to investigate the impact of operating cash flow on the Profitability of companies in different sectors listed on the Oman Stock Exchange. For five years (2017–2021) for 71 companies. In these sectors, Profitability and metrics have been used carefully. To show how these elements affect the success of each of these sectors? It is because investors can use this information to make sound investment decisions using different operating cash flow methods to evaluate company profitability. The researchers created a checklist to test the research hypothesis, in which each operating cash flow (OCF) was used as an independent variable. (Return on Assets (ROA) and Return on Equity (ROE)) was determined as the type of firm Profitability, which is the dependent variable. The results are shown through two models. Data using descriptive statistical analysis. Regression Approaches Using EViews 12, Given that the data included both time series and cross-sectional characteristics, fixed least squares (PLS) regression effects models were used to estimate the study models. The results reveal that operating cash flow has an impact of 57.9% on ROA, a measure of corporate Profitability, and other variables have an impact of 42.1% that is not mentioned. Meanwhile, the independent variable affects ROE by 57.7% and other variables by 42.3%. Paper type: Research paper.

Meta-learning-based approach for tool condition monitoring in multi-condition small sample scenarios

Tool Condition Monitoring (TCM) technology in machining is crucial for maintaining safety and optimizing costs. However, its practical application faces two significant challenges: difficulties in data collection and a decline in generalization performance across different monitoring tasks. To this end, a hybrid feature boundary-enhanced meta-learning network with adaptive gradients (HFBEAML) is proposed. This method combines cutting process parameters with time-series signal features, employing a multimodal one-dimensional Convolutional Neural Network (1D-CNN) based on the DenseNet architecture and a multi-head self-attention mechanism (MHSA) to mine multi-dimensional features sensitive to tool wear. To further enhance the model’s feature discrimination capability, a multi-loss joint optimization strategy is introduced, combining task-level loss with an improved triplet loss. This approach incorporates strategies for adaptive boundaries and handling the hardest positive and negative sample sets, thereby enhancing the robustness of feature metrics. Additionally, this research innovatively proposes an adaptive meta-level gradient update mechanism, dynamically adjusting gradient weights according to the characteristics of multiple tasks, aiming to improve the model’s generalization ability in multi-task learning environments. The effectiveness of the proposed method is demonstrated through experiments in two different scenarios, comparing its results with five other models showcasing its significant advantages in multi-task, small-sample data monitoring environments.

Using conditional Invertible Neural Networks to perform mid‐term peak load forecasting

AbstractMeasures for balancing the electrical grid, such as peak shaving, require accurate peak forecasts for lower aggregation levels of electrical loads. Thus, the Big Data Energy Analytics Laboratory (BigDEAL) challenge—organised by the BigDEAL—focused on forecasting three different daily peak characteristics in low aggregated load time series. In particular, participants of the challenge were asked to provide long‐term forecasts with horizons of up to 1 year in the qualification. The authors present the approach of the KIT‐IAI team from the Institute for Automation and Applied Informatics at the Karlsruhe Institute of Technology. The approach to the challenge is based on a hybrid generative model. In particular, the authors use a conditional Invertible Neural Network (cINN). The cINN gets the forecast of a sliding mean as representative of the trend, different weather features, and calendar information as conditioning input. By this, the proposed hybrid method achieved second place overall and won two out of three tracks of the BigDEAL challenge.