Spectrometry is practical for rapid analysis of fruit maturity. RBF LSSVR model is effective for Vis-NIR calibration. However, nonlinear effect of the high-dimensional spectrum data influences the model prediction accuracy. A fuzzy optimization strategy was proposed to improve performance of RBF LSSVR models in Vis-NIR determination of pomelo maturity. A linguistic iterative mode was introduced. The input variables were features extracted by fuzzy transform PCA. The output was a polynomial formula. In application, the maturity of pomelo was recorded as indices of L*, a* and b*. The fuzzy transformed data was input to the fuzzy optimized LSSVR model for calibration, validation and test. Parameter uncertainty was also investigated to examine the advantage of the proposed strategy. Results showed that the fuzzy optimized model gained appreciate predictions on the indices. This strategy is effective for implementation of Vis-NIR LSSVR models and expected for rapid prediction of fruit maturity.

Federated learning (FL) has recently been adopted to train shared models across industrial Internet of Things (IoT) devices without revealing their private raw data. Conventional FL usually relies on a central server for coordination. However, in reality, the central server is not fully trusted, which means that it may be collecting data, raising concerns about data leakage and misuse. Here, we propose a decentralized FL algorithm based on deep neural networks to address the problem of untrusted central servers. The original problem is decomposed into several subproblems with consensus constraints, which can be solved by local computation and communication. The proposed algorithm combines the decentralized average consensus and alternating direction method of multipliers. Several decentralized algorithms are employed for comparison, and the issue of heterogeneous data is discussed. Experimental evaluation shows the effectiveness of our proposed algorithm.

With the widespread application of smart devices, the security of internet of things (IoT) systems faces entirely new challenges. The IoT data stream operates in a non-stationary, dynamic environment, making it prone to concept drift. This paper focused on addressing the issue of concept drift in data streams, with a key emphasis on introducing an innovative drift detection method-ensemble multiple non-parametric concept localization detectors, abbreviated as EMNCD. EMNCD employs an ensemble of non-parametric statistical methods, including the Kolmogorov-Smirnov, Wilcoxon rank sum and Mann-Kendall tests. By comparing sample distributions within a sliding window, EMNCD accurately detects concept drift, achieving precise localization of drift points, and enhancing overall detection reliability. Experimental results demonstrated the superior performance of EMNCD compared to classical methods on artificial datasets. Simultaneously, to enhance the robustness of data stream processing, we presented an online anomaly detection method based on the isolation forest (iForest). Additionally, we proposedwhale optimization algorithm (WOA)-extreme gradient boosting (XGBoost), a drift adaptation model employing XGBoost as a base classifier. This model dynamically updates using drift points detected by EMNCD and fine-tunes parameters through the WOA. Real-world applications on the edge-industrial IoTset (IIoTset) intrusion dataset explore the impact of concept drift on intrusion detection, where IIoT is a subclass of IoT. In summary, this paper focused on EMNCD, introducing innovative approaches for drift detection, anomaly detection, and drift adaptation. The research provided practical and viable solutions to address concept drift in data streams, enhancing security in IoT systems.

Urban sensing has become prevalent for monitoring dynamic urban status due to the development of intelligent computing. Fast sensing and accurate analyzing of soil quality promote to make data-driven urban planning and management. Soil contamination hiders the sustainable development of the city environmental ecology. Decentralized sensing based on an Internet of Things (IoT) architecture is a state-of-art technology. Portable spectral detection serves for immediate quantitative analysis of target components. In this case, portable sensing combined with federated learning simply shares the informative features extracted from the IoT-based spectral sensing data. However, the modeling technique for the analysis of spectroscopic sensing data faces several challenges due to the emerge of the big data problem in IoT-based sceneries. The consideration of possible federal learning of multi-node collective data should ask for deep studies of modeling methodologies. In this paper, we built up an IoT-based portable spectral sensing system for simultaneous sensing and detecting the soil data at distributed sensing places. The collaborate training model was established for dealing with the dynamic sensing data, by the fusion design of a broad learning network (BLN) and a fuzzy partial least square (fPLS) model. The BLN output feature variables are produced from the scalable pseudo input layer with training linking weights adaptively, and the number of fuzzy rules is scalable. In experiment, the decentralized sensing system and the fusion modeling framework carry out the model optimization for the portable near-infrared spectroscopic sensing data of heavy metal in soil samples. The quantification of two elements demonstrated that the proposed framework yielded superior prediction results to the conventional PLS model for the training and evaluating on the dynamic sensing data. In interaction with the IoT architecture, the framework is prospectively expected as an advanced intelligent technical support for fast analysis of the portable spectroscopic sensing data.