Abstract: The present application provides a three-mode storage method for program blocks based on check, comprising: first dividing the program to be injected into N blocks and checking each block to obtain the first checksum; then injecting the program to be injected and the first checksum twice in the program storage area, and the two injected programs are noted as the main program and the backup program, respectively; dividing the main program into N blocks, and checking each program block to obtain the second checksum during program run; dividing the backup program into N blocks and checking each block to obtain the third checksum; and comparing the checksum, second checksum and third checksum: if the three are consistent, performing no operation; updating the program or checksum and checking again, if the three are inconsistent.

Abstract: A navigation satellite time system and its autonomous recovery method are provided, including a load time system, the load time system is configured to generate and maintain the load time, and the load time system comprises an space borne atomic clock, a time-frequency processing unit and a plurality of load time backups module, the time information of the load time is obtained from the ground station time; the pulse-per-second signals of the load time are generated and maintained by the space borne atomic clock and the time-frequency processing unit; when the time-frequency processing unit fails, the first-level recovery state is triggered: the time-frequency processing unit compares the time information and the pulse-per-second signals reversely output by the multiple load time backup modules to perform load time recovery.

Abstract: A carrier aircraft, comprising: a body for carrying at least one object or passenger; and multiple rotatable fan devices disposed on two sides of said body; wherein an angle of rotation of one of said rotatable fan devices can be changed, such that said carrier aircraft flies along a direction at an angle, said angle being between 0 degrees to 30 degrees.

Abstract: Methods, computer program products, and systems are presented for service cost prediction using machine learning. A claims risk model is formed for predicting a per member per month cost based on variables trained in early prediction models and variables of claim record data from data sources. A training dataset for the claims risk model is modified based on a preconfigured cap value and two distinctive datasets are generated, which trains the claims risk model into a capped submodel and an outlier submodel, respectively. Each submodel makes prediction of the per member per month cost and a slice model interpolates a cost predicted by the capped submodel and another predicted by the outlier submodel. The splice model redistributes the outcomes of the capped submodel and the outlier submodel.

Abstract: Methods, computer program products, and systems are presented for service cost prediction using machine learning. A claims risk model is formed for predicting a per member per month cost based on variables trained in early prediction models and variables of claim record data from data sources. A training dataset for the claims risk model is modified based on a preconfigured cap value and two distinctive datasets are generated, which trains the claims risk model into a capped submodel and an outlier submodel, respectively. Each submodel makes prediction of the per member per month cost and a slice model interpolates a cost predicted by the capped submodel and another predicted by the outlier submodel. The splice model redistributes the outcomes of the capped submodel and the outlier submodel.