The user operates the smart home devices year in year out, have produced mass operation data, but these data do not be utilized well in past. Nowadays, these data can be used to predict user's behavior custom with the development of big data and machine learning technologies, and then the prediction results can be employed to enhance the intelligence of a smart home system. In view of this, this paper proposes a novel unsupervised user behavior prediction (UUBP) algorithm, which employs an artificial neural network and proposes a forgetting factor to overcome the shortcomings of the previous prediction algorithm. This algorithm has a high-level of autonomous and self-organizing learning ability while does not require too much human intervention. Furthermore, the algorithm can better avoid the influence of user's infrequent and out-of-date operation records, because of the forgetting factor. Finally, the use of real end user's operation records to demonstrate that UUBP algorithm has a better level of performance than other algorithms from effectiveness.

This paper proposes a novel intelligent disease diagnosis method based on fuzzy concept lattice. Symptoms and the corresponding extents (e.g., frequency, severity, and duration) of each disease can be extracted to form a fuzzy concept lattice. The fuzzy concept lattice of the symptoms and their extents to be diagnosed needs to be constructed to match the fuzzy concept lattice of possible diseases. The similarity between the above two types of concept lattices can be calculated and used to aid for effective diagnosis. Naturally, the disease with the largest similarity is the finding of intelligent diagnosis. In the future, more efficient fuzzy concept lattice construction method and update algorithm will be explored, which are presumed to be very complicated.

The vehicle-road cooperative perception needs high accuracy and real-time automatic background filtering to separate background objects from foreground objects in complex traffic scenes. Reducing the influence of foreground objects to improve accuracy, and introducing a new framework to improve real-time performance are two main challenges in automatic background filtering. This paper proposes an Automatic Background Filtering method with innovative Frame Selection and Background Matrix Extraction modules (ABF-FSBME) to address these challenges. Firstly, a new space division method with equal hitting probability is proposed to divide the 3D point cloud formed by roadside Light Detection and Ranging (LiDAR), which can reduce the influence of slight LiDAR vibrations. Secondly, the terminal-edge-cloud framework is introduced to balance delay-constrained tasks and computation-intensive tasks in automatic background filtering. Thirdly, a variance-based frame selection strategy with a sliding window mechanism is proposed to select candidate frames with fewer foreground objects. This strategy can reduce the influence of foreground objects in a coarse-grained way. Meanwhile, a new background matrix extraction method is proposed to construct the background matrix. This method can further reduce the influence of foreground objects in a fine-grained way. Finally, based on the extracted background matrix from a cloud server, the edge server can filter the raw frame in real-time. The experimental results show that the proposed ABF-FSBME method has better accuracy than other methods in error rate and integrity rate. Besides, the proposed ABF-FSBME can complete fame filtering within 10ms, and has almost no network delay, so it can satisfy the real-time requirement.