In real-world conditions, it is challenging to acquire high-quality gear signals from measured vibration data, as the signals are often masked by noise, transmission path effects, and interference from other components. This makes it difficult to identify health states. To address this issue, this paper proposes a novel Physics-informed machine learning (PIML) framework, named Modulated Model Network (MMN), for high-precision reconstruction of signals under different health conditions. MMN integrates a modulation signal model—including harmonic, impulsive, amplitude modulation (AM), and frequency modulation (FM) components—into a neural network. The loss function is designed to quantify the spectral discrepancy between the original and reconstructed signals, ensuring that the reconstruction process retains and further enhances the vibration characteristics. This fully customized architecture establishes a one-to-one correspondence between network parameters and signal model parameters, thereby endowing MMN with complete physical interpretability and controllability. Two experimental settings were considered: different fault types and different crack severities in gears. For comparison, the variational mode decomposition (VMD) method was employed. Results demonstrate that the reconstructed signals perform best in both time and frequency domains, successfully preserving and enhancing gear vibration features. Finally, a one-dimensional convolutional neural network (1-D CNN) was used to classify the original signals, VMD-processed signals, and MMN-reconstructed signals under various health conditions. The classification accuracies of the three types of signals in the first experiment are 93.00 %, 94.33 %, and 99.50 %, respectively, whereas in the second experiment, the accuracies were 88.67 %, 92.89 %, and 97.11 %, respectively.

As an important part of rotating machinery, gear is easy to appear some unexpected fault states, and its fault diagnosis is very important. Fourier decomposition method (FDM) is a common method for gear fault diagnosis, but the noise robustness, period recognition, and extraction capabilities of FDM are unsatisfactory. Based on this, in this article, Ramanujan Fourier mode decomposition (RFMD) method is proposed. The RFMD not only has a complete mathematical theory foundation but also has an excellent ability to identify and extract periodic components. Emulational and experimental results of planetary gearbox show that the RFMD method has good noise robustness and can accurately extract gear fault characteristic information. Thus, it is an effective gear fault diagnosis method.

The Web has become an excellent source for gathering consumer opinions. There are now numerous Web sites containing such opinions, e.g., customer reviews of products, forums, discussion groups, and blogs. This paper focuses on online customer reviews of products. It makes two contributions. First, it proposes a novel framework for analyzing and comparing consumer opinions of competing products. A prototype system called Opinion Observer is also implemented. The system is such that with a single glance of its visualization, the user is able to clearly see the strengths and weaknesses of each product in the minds of consumers in terms of various product features. This comparison is useful to both potential customers and product manufacturers. For a potential customer, he/she can see a visual side-by-side and feature-by-feature comparison of consumer opinions on these products, which helps him/her to decide which product to buy. For a product manufacturer, the comparison enables it to easily gather marketing intelligence and product benchmarking information. Second, a new technique based on language pattern mining is proposed to extract product features from Pros and Cons in a particular type of reviews. Such features form the basis for the above comparison. Experimental results show that the technique is highly effective and outperform existing methods significantly.