Various attempts have been conducted to improve the performance of text-based sentiment analysis. These significant attempts have focused on text representation and model classifiers. This paper introduced a hybrid model based on the text representation and the classifier models, to address sentiment classification with various topics. The combination of BERT and a distilled version of BERT (DistilBERT) was selected in the representative vectors of the input sentences, while the combination of long short-term memory and temporal convolutional networks was taken to enhance the proposed model in understanding the semantics and context of each word. The experiment results showed that the proposed model outperformed various counterpart schemes in considered metrics. The reliability of the proposed model was confirmed in a mixed dataset containing nine topics.

This study aimed to design chaos-based synchronized dynamic keys and develop an improved chaos-based advanced encryption standard (AES) algorithm with the proposed synchronized random keys. First, based on sliding mode control (SMC) technology, a rippling control scheme was introduced to guarantee the synchronization between master–slave discrete chaotic systems. Under the synchronization, the same dynamic random chaos signals could be simultaneously obtained at the transmitter and receiver in communication systems. Then, a novel modified AES cryptosystem with dynamic random keys based on chaos synchronization was presented. In a traditional AES cryptosystem, a static key is used, and it must be exchanged in advance and confirmed to be safely kept. However, in the proposed design, by introducing the synchronization technology of chaotic systems, the static key becomes dynamic and random, and it does not need to be kept or transmitted in open channels. Consequently, the disadvantage of key storage could be eliminated and the security of encryption could be improved. Finally, the developed chaos-based AES (CAES) algorithm has been applied to construct a novel image encryption algorithm. The statistical analysis, histogram, information entropy, and correlation indexes have been calculated and analyzed through simulation experiments in order to demonstrate the capability and improvement of this presented CAES cryptosystem.

Power consumption forecasting is a crucial need for power management to achieve sustainable energy. The power demand is increasing over time, while the forecasting of power consumption possesses challenges with nonlinearity patterns and various noise in the datasets. To this end, this paper proposes the RobustSTL and temporal convolutional network (TCN) model to forecast hourly power consumption. Through the RobustSTL, instead of standard STL, this decomposition method can extract time series data despite containing dynamic patterns, various noise, and burstiness. The trend, seasonality, and remainder components obtained from the decomposition operation can enhance prediction accuracy by providing significant information from the dataset. These components are then used as input for the TCN model applying deep learning for forecasting. TCN employing dilated causal convolutions and residual blocks to extract long-term data patterns outperforms recurrent networks in time series forecasting studies. To assess the proposed model, this paper conducts a comparison experiment between the proposed model and counterpart models. The result shows that the proposed model can grasp the rules of historical time series data related to hourly power consumption. Our proposed model overcomes the counterpart schemes in MAPE, MAE, and RMSE metrics. Additionally, the proposed model obtains the best results in precision, recall, and F1-score values. The result also indicates that the predicted data can fit the pattern of the actual data.

This study is concerned with robust synchronization for master–slave chaotic systems with matched/mismatched disturbances and uncertainty in the control input. A robust sliding mode control (SMC) is presented to achieve chaos synchronization even under the influence of matched/mismatched disturbances and uncertainty of inputs. A proportional-integral (PI) switching surface is introduced to make the controlled error dynamics in the sliding manifold easy to analyze. Furthermore, by using the proposed SMC scheme even subjected to input uncertainty, we can force the trajectories of the error dynamics to enter the sliding manifold and fully synchronize the master–slave systems in spite of matched uncertainties and input nonlinearity. As for the mismatched disturbances, the bounds of synchronization errors can be well estimated by introducing the limit of the Riemann sum, which is not well addressed in previous works. Simulation experiments including matched and mismatched cases are presented to illustrate the robustness and synchronization performance with the proposed SMC synchronization controller.

In an industrial mass production pattern, quality prediction is one of the important processes when guarding quality. The products are extracted periodically or quantitatively for inspection in order to observe the relationship between process variation and engineering specification. When these irregularities are not instantly detected by lot sampling inspection, lot defectives are produced, and the defective cost increases. Failure to identify defects during sampling inspection leads to product returns or harm to business reputation. Press casting is a common mass production method in the metal industry. After the metal is molten at a high temperature, high pressure is injected into the mold, and then it is solidified and formed in the mold. Thus, pressure stability inside the mold is one of the key factors that influences quality. The melting point of aluminum alloy is normally around 650 °C, but there was no sensor that could withstand this high temperature. To combat this, we developed a high temperature resistant sensor and installed it into pressure casting mold grounded on the principles of fluid mechanics and experts’ suggestions in order to realize the impact of pressure change on the mold. To our limited knowledge, it was a seminal study on predicting mold’s casting quality via in-mold pressure data. We propose a press casting quality prediction method based on machine learning. By collecting the in-mold pressure data in real time. Savitzky-Golay Filter is used for data smoothing, and first-order difference is taken to extract the time interval of an actual injection of molten metal in-to the mold. We extract the key data that influence the quality and employ XGBoost to establish a classifier. In the experiment we prove that the method achieves good accuracy of quality prediction and recall of defectives for in-mold pressure. Via this model, we not only can save large amount of time and costs, but also can carry out maintenance warnings in advance, notify professionals to stop produce defective products, reduce the shipping risk and maintain reputation so as to strengthen its competitive edge.