Photovoltaic (PV) power forecasting is often developed under an implicit assumption of stationarity, yet the weather–power relationship evolves over time and induces distribution shifts commonly known as concept drift. Existing PV power forecasting models either ignore this issue entirely or address it in a limited and insufficient manner. To better understand this non-stationarity, we distinguish between internal and external drift, which motivate the key components of our design. Therefore, we present ADrift , a drift-aware forecasting framework that integrates patch-based temporal modeling, weather-guided representation learning, and lightweight online adaptation. To model internal drift, the backbone employs prototype-guided weather experts that capture diverse meteorological patterns within each input window. To cope with external drift, a learnable adapter updates its parameters through a temporal gap attention mechanism that enables targeted adjustments to the model. In addition, a proactive update strategy further mitigates supervision delays under rapidly changing conditions. Experiments on three real-world PV datasets show that ADrift consistently improves forecasting accuracy over static and online-learning baselines, demonstrating its potential for practical deployment under evolving weather conditions.

• To the best of our knowledge, this work is the first to quantify weather-type imbalance in PV power datasets based on the intrinsic volatility of PV power, rather than relying on external parameters. • We observe that samples with high PV power volatility account for most of the training loss, which significantly reduces forecasting performance. • We design a novel volatility-aware re-weighting framework (ReMAV) that adjusts the importance of training samples based on their volatility levels, thereby improving model accuracy under imbalanced PV power datasets. • We validate the proposed framework on three benchmark datasets and demonstrate that our proposed ReMAV framework effectively handles weather-type imbalance in PV power datasets and consistently outperforms existing baseline models in forecasting accuracy. Recent short-term photovoltaic (PV) power forecasting methods have primarily focused on improving model architectures to enhance forecasting accuracy, they often overlook the issue of weather-type imbalance in PV power datasets. To this end, we first introduce a new metric, Mean Accumulated Volatility (MAV) , which quantifies the volatility of each sample. By translating unquantified weather-type imbalance into a measurable form of volatility imbalance, we observe that high-MAV samples account for most of the training loss, thereby harming the model’s forecasting accuracy. Then, we further propose ReMAV , a volatility-aware Re -weighting framework that down-weights the losses of high-MAV samples and up-weights those of low-MAV samples based on the MAV -based density. Extensive experiments on eleven baseline forecasting models across three real-world PV power datasets demonstrate that our proposed ReMAV framework effectively handles PV power with weather-type imbalance and consistently outperforms existing baseline models in forecasting accuracy. For example, on the Alice Springs dataset, ReMAV reduces average MAE by 8.53% over baselines, while on the PVOD dataset, MAE drops by 5.46% on average.

In this work, we propose a new Dual Min-Max Games (DMMG) based self-supervised skeleton action recognition method by augmenting unlabeled data in a contrastive learning framework. Our DMMG consists of a viewpoint variation min-max game and an edge perturbation min-max game. These two min-max games adopt an adversarial paradigm to perform data augmentation on the skeleton sequences and graph-structured body joints, respectively. Our viewpoint variation min-max game focuses on constructing various hard contrastive pairs by generating skeleton sequences from various viewpoints. These hard contrastive pairs help our model learn representative action features, thus facilitating model transfer to downstream tasks. Moreover, our edge perturbation min-max game specializes in building diverse hard contrastive samples through perturbing connectivity strength among graph-based body joints. The connectivity-strength varying contrastive pairs enable the model to capture minimal sufficient information of different actions, such as representative gestures for an action while preventing the model from overfitting. By fully exploiting the proposed DMMG, we can generate sufficient challenging contrastive pairs and thus achieve discriminative action feature representations from unlabeled skeleton data in a self-supervised manner. Extensive experiments demonstrate that our method achieves superior results under various evaluation protocols on widely-used NTU-RGB+D, NTU120-RGB+D and PKU-MMD datasets.

Despite its efficiency, the classical Hopfield network was a highly impractical data-searching solution due to its limited storage capacity. While the recently released modern Hopfield variant has increased its storage capacity, its searching ability is heavily affected by local minima and saddle points, which prevented it from becoming a worthy successor of the classical Hopfield network. We propose a novel unsupervised clustering approach to bypass local minima and saddle points to enhance the overall robustness of the Hopfield network. Our experimental results on benchmark MNIST indicate that our algorithm can increase the retrieval accuracy by over (20%) in general against the Hopfield Update Rule, proving that it is a far superior modelling solution.

In a data stream, concept drift refers to unpredictable distribution changes over time, which violates the identical-distribution assumption required by conventional machine learning methods. Current concept drift adaptation techniques mostly focus on a data stream with changing distributions. However, since each variable of a data stream is a time series, these variables normally have temporal dependency problems in the real world. How to solve concept drift and temporal dependency problems at the same time is rarely discussed in the concept-drift literature. To solve this situation, this article proves and validates that the testing error decreases faster if a predictor is trained on a temporally reconstructed space when drift occurs. Based on this theory, a novel drift adaptation regression (DAR) framework is designed to predict the label variable for data streams with concept drift and temporal dependency. A new statistic called local drift degree (LDD+) is proposed and used as a drift adaptation technique in the DAR framework to discard outdated instances in a timely way, thereby guaranteeing that the most relevant instances will be selected during the training process. The performance of DAR is demonstrated by a set of experimental evaluations on both synthetic data and real-world data streams.

A critical challenge in human-autonomy teaming is for human players to comprehend their nonhuman teammates (agents). Transparency in agents' behaviors is the key for such comprehension, which may be obtained by embedding a self-explanation ability into the agent to explain its own behaviors. Previous studies have relied on searching for the executed functions and logics to generate explanations for behaviors of goal-following logic-based agents. With the increasing number of functions and logics, current methods, such as component and process-based methods, have become impractical. This article proposes a new method exploiting the agent's artificial situation awareness states for generating explanations that involves several techniques: A Bayesian network, fuzzy theory, and Hamming distance. Our new method is evaluated in a collaborative driving context, in which a significant number of accidents recently occurred around the globe due to the lack of understanding of the autopilot agents. Using an autonomous driving simulator called Carla, two typical scenarios in collaborative driving, namely, traffic light and overtaking situations, are used. The findings show that the new method potentially reduces the search space in generating explanations and exhibits better computational performance and a lower cognitive workload. This work is important to calibrate human trust and to enhance comprehension of the agent.

Distributed energy sources including renewable energy (RE) sources and electric vehicle (EV) discharging offer opportunities to improve grid performance. Although uncoordinated EV charging can perfectly meet users’ driving needs, it brings great challenges in maintaining the quality of low voltage (LV) distribution grids. Most existing EV control strategies do not take into account the heterogeneous nature of EV charging, including the various users’ travel needs and the intermittent nature of RE sources. This article proposes a hierarchical control method that simultaneously coordinates EV charging by considering a dynamic energy tariff, RE sources, EV hardware characteristics, and particularly the EV’s driving needs. The proposed local controller sends the EV’s charging priority and required energy to the central controller based on the EV’s hardware characteristics and users’ driving needs. The central controller then uses this information along with the energy tariff from the retailer and the present grid performance to control EV’s charging or discharging power. The efficacy of this hierarchical control method is evaluated using an Australian LV grid. The obtained results show that this method can reduce the neutral current and voltage imbalance with the maximized usage of renewable energy resources while the EV users’ driving needs are met.

In concept drift adaptation, we aim to design a blind or an informed strategy to update our best predictor for future data at each time point. However, existing informed drift adaptation methods need to wait for an entire batch of data to detect drift and then update the predictor (if drift is detected), which causes adaptation delay. To overcome the adaptation delay, we propose a sequentially updated statistic, called drift-gradient to quantify the increase of distributional discrepancy when every new instance arrives. Based on drift-gradient, a segment-based drift adaptation (SEGA) method is developed to online update our best predictor. Drift-gradient is defined on a segment in the training set. It can precisely quantify the increase of distributional discrepancy between the old segment and the newest segment when only one new instance is available at each time point. A lower value of drift-gradient on the old segment represents that the distribution of the new instance is closer to the distribution of the old segment. Based on the drift-gradient, SEGA retrains our best predictors with the segments that have the minimum drift-gradient when every new instance arrives. SEGA has been validated by extensive experiments on both synthetic and real-world, classification and regression data streams. The experimental results show that SEGA outperforms competitive blind and informed drift adaptation methods.

Accurately crude oil price prediction remains challenging so far. Despite the abundant research achievements of crude oil price prediction, most of them emphasize the linear and deterministic modeling, which cannot adequately capture the complex nonlinear characteristics and uncertainties involved, thus impeding further developments in the field. In this article, a novel learning system with the aim of obtaining the deterministic and probabilistic predictions is presented to model the nonlinear dynamics in crude oil price, composed by the modules of recurrence analysis, outlier detection, data preprocessing, feature selection, predictive modeling based on deep learning, and system evaluation. In particular, the temporal convolution is developed to perform feature selection, thus improving the generalization of the system. Additionally, the extensions, including the predictive performance test evaluation, convergence investigation, and sensitivity analysis, are carried out. The experimental simulations show that the proposed system can yield the deterministic and probabilistic predictions with higher accuracy and feasibility compared with the benchmarks considered, further indicating its effectiveness.

How to handle concept drift problem is a big challenge for algorithms designed for the data streams. Currently, techniques related to the concept drift problem focus on single data stream. However, it normally needs to handle multiple relevant data streams in the real-world application. Current concept drift methods can not be directly used in the multistream setting. They can only be limitedly applied on each stream separately, which omits the drift correlation between streams. In the multi-stream scenario, when drift occurs in a stream, other streams may face or have faced a similar drift problem as well. This pattern of simultaneous or delayed occurrence of drift is critical to analyze and predict multiple streams as a whole dynamic system. To fill the gap in the multi-stream scenario, this paper proposes a fuzzy drift variance (FDV) to measure the correlated drift patterns among streams. FDA is able to present how the pattern of drift occurrence for any two streams correlates and how delayed this correlation is. Seven synthetic streams are designed to validate FDA. The experimental results show a good presentation ability of FDA for drift-correlated multiple streams.

It is highly desirable that web search engines know users well and provide just what the user needs. Although great effort has been devoted to achieve this dream, the commonly used web search engines still provide a “one-fit-all” results. One of the barriers is lack of an accurate representation of user search context that supports personalised web search. This article presents a method to represent user search context and incorporate this representation to produce personalised web search results based on Google search results. The key contributions are twofold: a method to build contextual user profiles using their browsing behaviour and the semantic knowledge represented in a domain ontology; and an algorithm to re-rank the original search results using these contextual user profiles. The effectiveness of proposed new techniques were evaluated through comparisons of cases with and without these techniques respectively and a promising result of 35% precision improvement is achieved.

Current models and algorithms have been increasingly required to learn in a nonstationary environment because the phenomenon of concept drift (or pattern shift) may occur, that is, the assumption that data are identically distributed may be invalid in data streams. Once the data pattern changes, a well-trained model built on the previous, now obsolete data cannot provide an accurate prediction for future data. To obtain reliable prediction, it is important to understand the existing patterns in the data stream and to know which pattern the current examples belong to during the modeling process. However, it is ambiguous to classify an example to a certain pattern in many real-world cases. In this paper, we propose a novel adaptive regression approach, called FUZZ-CARE, to dynamically recognize, train, and store patterns, and assign the membership degree of the upcoming examples belonging to these patterns. Membership degrees are presented by the membership matrix obtained from a kernel fuzzy c-means clustering, which is synchronously trained and adapted with regression parameters. Rather than designing a complicated procedure to continuously chase the newest pattern, which is a common approach in the literature, FUZZ-CARE abstracts useful past information to help predict newly arrived examples. It thus effectively avoids the risk of insufficient training due to the lack of new data and improves prediction accuracy. Experiments on six synthetic datasets and 21 real-world datasets validate the high accuracy and robustness of our approach.

Concept drift referring to the changes of data distributions has been one critical challenge typically associated with mining data streams. Current drift detection and adaptation methods focus on how to immediately detect the distribution changes once the concept drift occurs and swiftly update the model to be applicable to the newly arrived data instances. Most of those methods assume the data does not have noise or the noise is too weak to affect the modeling procedure. However, realworld data are normally contaminated, and denoise techniques are highly preferred as a necessary preprocess. This issue is more complex for a data stream with concept drift because the noise is very likely to be confused with drift. Motivated by that, this paper proposes a Noise-tolerant Fuzzy c-means based drift Adaptation method (NFA) which can adapt to the changing distributions and is suitable for noisy data streams. The concept drift problem is solved by using a fuzzy c-means based regression model to continuously include the most relevant data instances to the latest pattern in the training set. In addition, a denoise technique is designed in NFA to remove noise, and the ability of incremental updating enables it to be embedded in the incremental drift adaptation process, and therefore NFA can solve concept drift and noise problems at the same time. Experimental evaluation results also show good performance of our method on handling data streams with concept drift and noise.

Prediction in non-stationary environments, where data streams are ever-changing at very high speeds, has become more and more important in real-world applications. The uncertainty in data streams caused by changes in data distribution is described as concept drift. The appearance of concept drift in a data stream results in inconsistencies between the existing data and incoming data. Such inconsistencies pose a great challenge to conventional machine learning methods, given they are built on the assumption of independent and identically distributed data and cannot adapt to unpredictable changes in knowledge patterns. To solve such data stream uncertainty problem, this paper presents a window-based self-adaptive fuzzy network called adaptive fuzzy network (AFN), which can continuously modify the network through identifying new knowledge from the previous data samples. Three components are embedded in ANF: a drift detection module to identify whether the current window of data samples presents different pattern from the previous; a drift adaption module to retain useful knowledge in previous samples; and a fuzzy inference system, which integrates the detection and adaption modules for prediction. ANF has been evaluated through a set of experiments on non-stationary data streams. The experimental results show a good effectiveness of our method.

Concept drift, given the huge volume of high-speed data streams, requires traditional machine learning models to be self-adaptive. Techniques to handle drift are especially needed in regression cases for a wide range of applications in the real world. There is, however, a shortage of research on drift adaptation for regression cases in the literature. One of the main obstacles to further research is the resulting model complexity when regression methods and drift handling techniques are combined. This paper proposes a self-adaptive algorithm, based on a fuzzy kernel c-means clustering approach and a lazy learning algorithm, called FKLL, to handle drift in regression learning. Using FKLL, drift adaptation first updates the learning set using lazy learning, then fuzzy kernel c-means clustering is used to determine the most relevant learning set. Experiments show that the FKLL algorithm is better able to respond to drift as soon as the learning sets are updated, and is also suitable for dealing with reoccurring drift, when compared to the original lazy learning algorithm and other state-of-the-art regression methods.

Many web-based online learning systems focus more on textual and/or image based content delivery without including experiment systems, or if included they are usually operated under pre-defined conditions, such as fixed scenarios and pre-determined delivery orders. These limitations hinder personalized learning and collaboration between students and discourage student engagement. To circumvent these limitations, an Intelligent Web-based Experiment Management System (IWEMS) using multi-agent concept is presented. In the system, three kinds of software agents are used: (i) Student-Agent, responsible for assessing the knowledge levels of students. A fuzzy set based algorithm is used and the results are plotted through a dynamic polar chart; (ii) Teacher-Agent, responsible for tracking experiment progress of each student and recommending personalized the next-to-do experiment to him or her; and (iii) Co-Agent, responsible for group formation based on similar knowledge levels to facilitate collaborative learning between students. A prototype of this system is developed using a Java Agent Development Framework(JADE), where a client/server architecture and a MySQL database are used. It demonstrates the validity of the design and effectiveness of this system's functionality, achieves the personalization recommendation of next-to-do experiment and collaborative learning environment.

Air quality has received continuous attention from both environmental managers and citizens. Accordingly, early-warning systems for air pollution are very useful tools to avoid negative health effects and develop effective prevention programs. However, developing robust early-warning systems is very challenging, as well as necessary. This paper develops a reliable and effective early-warning system that consists of air quality prediction and assessment modules. In the prediction module, a hybrid forecasting method is developed for predicting pollutant concentrations that effectively estimates future air quality conditions. In developing this proposed model, we suggest the use of a back propagation neural network algorithm, combined with a probabilistic parameter model and data preprocessing techniques, to address the uncertainties involved in future air quality prediction. Meanwhile, a pre-analysis is implemented, primarily by using optimized distribution functions to examine and analyze statistical characteristics and emission behaviors of air pollutants. The second method, which is developed as part of the second module, is based on fuzzy set theory and the Analytic Hierarchy Process, and it performs air quality assessments to provide a clear and intelligible description of air quality conditions. Using data from the Ministry of Environmental Protection of China and six stages of air quality classification levels, specifically good, moderate, lightly polluted, moderately polluted, heavily polluted and severely polluted , two cities in China, Chengdu and Hangzhou, are used as illustrative examples to verify the effectiveness of the developed early-warning system. The results demonstrate that the proposed methods are effective and reliable for use by environmental supervisors in air pollution monitoring and management.

An improved particle swarm optimization (PSO) algorithm based on self-adaptive excellence coefficients, Cauchy operator and 3-opt, called SCLPSO, is proposed in this paper in order to deal with the issues such as premature convergence and low accuracy of the basic discrete PSO when applied to traveling salesman problem (TSP). To improve the optimization ability and convergence speed of the algorithm, each edge is assigned a self-adaptive excellence coefficient based on the principle of roulette selection, which can be adjusted dynamically according to the process of searching for the solutions. To gain better global search ability of the basic discrete PSO, the Cauchy distribution density function is used to regulate the inertia weight so as to improve the diversity of the population. Furthermore, the 3-opt local search technique is utilized to increase the accuracy and convergence speed of the algorithm. Through simulation experiments with MATLAB, the performance of the proposed algorithm is evaluated on several classical examples taken from the TSPLIB. The experimental results indicate that the proposed SCLPSO algorithm performs better in terms of accuracy and convergence speed compared with several other algorithms, and thus is a potential intelligence algorithm for solving TSP.

Web-page recommendation plays an important role in intelligent Web systems. Useful knowledge discovery from Web usage data and satisfactory knowledge representation for effective Web-page recommendations are crucial and challenging. This paper proposes a novel method to efficiently provide better Web-page recommendation through semantic-enhancement by integrating the domain and Web usage knowledge of a website. Two new models are proposed to represent the domain knowledge. The first model uses an ontology to represent the domain knowledge. The second model uses one automatically generated semantic network to represent domain terms, Web-pages, and the relations between them. Another new model, the conceptual prediction model, is proposed to automatically generate a semantic network of the semantic Web usage knowledge, which is the integration of domain knowledge and Web usage knowledge. A number of effective queries have been developed to query about these knowledge bases. Based on these queries, a set of recommendation strategies have been proposed to generate Web-page candidates. The recommendation results have been compared with the results obtained from an advanced existing Web Usage Mining (WUM) method. The experimental results demonstrate that the proposed method produces significantly higher performance than the WUM method.

Current semantic recommender systems aim to exploit the website ontologies to produce valuable web recommendations. However, Web usage knowledge for recommendation is presented separately and differently from the domain ontology, this leads to the complexity of using inconsistent knowledge resources. This paper aims to solve this problem by proposing a novel ontology-style model of Web usage to represent the non-taxonomic visiting relationship among the visited pages. The output of this model is an ontology-style document which enables the discovered web usage knowledge to be sharable and machine-understandable in semantic Web applications, such as recommender systems. A case study is presented to show how this model is used in conjunction of the web usage mining and web recommendation. Two real-world datasets are used in the case study.

E-government is becoming more attentive towards providing intelligent personalized online services to citizens so that citizens can receive better services with less time and effort. This paper proposes a new conceptual framework for delivering personalized e-government services to citizens from a citizen-centric approach, called Pe-Gov service framework. This framework outlines the main components and their interconnections. Detailed explanations about these components are given and the special features of this framework are highlighted. The Pe-Gov framework has the potential to outperform the existing e-Gov service systems as illustrated by two real life examples. © 2010 ACM.