71results about How to "Expand the range of values" patented technology

The invention discloses an Internet financial credit evaluation method based on a PSO-BP neural network, and the method comprises the steps: obtaining a result true value of information, carrying outthe normalization processing and principal component analysis dimensionality reduction of obtained data, dividing a test set and a training set, initializing the number of input nodes, the number of output nodes and the number of hidden layer nodes of the BP neural network; using a traditional gradient descent method and back propagation for continuously adjusting the weight and the threshold of anetwork to construct a BP neural network model, using a particle swarm algorithm for optimizing the connection weight and the threshold to obtain a PSOBP neural network model, and using a verification set for verification and optimization; and deploying the model to an application system to perform feature parameter extraction and prediction classification on data of a real-time application client. According to the invention, the convergence rate of the BP neural network is greatly improved, the obtained credit evaluation model of the PSO-BP neural network can accurately and quickly realize credit evaluation of an Internet financial applicant, the service timeliness of application approval is effectively improved, and the risk control cost and the application fraud risk are reduced.

A ring network protection method for the optical transmission network relates to the digital information transmission and includes using the overhead byte APS/PCC in the ODUk to transmit the protection information of the APS. This method adds fields in the APS/PCC and further includes: a. the network element in which the protection switching takes place in the OTN sends the protection switching information to other network elements through the fields set in the APS/PCC of the ODUk; b. the network element which received the ODUk performs corresponding processing according to the fields carried by the APS/PCC in the ODUk and implements the protection switching for the OTN ring network. In the technical solution of the present invention, the overhead byte APS/PCC is extended and is used to transmit the APS information. The network elements in the OTN ring network perform the corresponding processing according the fields in the overhead byte APS/PCC so that while the protection switching is implemented for the OTN ring network, the problems occurring in the SDH APS will not appear in the OTN ring network.

A method in which the wavelength of the light of a tunable light source is varied periodically over an absorption line of interest for the gas component to measure the concentration of a gas component in a measurement gas based on one of two measurement methods of direct absorption spectroscopy and wavelength modulation spectroscopy and, where in the case of wavelength modulation spectroscopy, the wavelength of the light is additionally sinusoidally modulated at a high frequency and with a small amplitude, where the intensity of the light is detected after transradiation of the measurement gas and processed to yield a measurement result, and where in order to increase the accuracy and reliability of the measurement, the two measurement methods are applied simultaneously during each period, or alternately in consecutive periods, and their results are combined by averaging to form the measurement result.

The invention relates to a slip-limiting differential mechanism with a noncircular bevel gear. The slip-limiting differential mechanism with the noncircular bevel gear consists of a differential mechanism shell, a noncircular planet gear and a noncircular half axle gear. The planet gear is connected with the shell by a shaft handle or a linear shaft and rotates with respect to the shell. The gear ratio of the noncircular planet gear to the noncircular half axle gear is 1: 2. The planet gear and the half axle gear adopt the noncircular bevel gear structures, the transmission ratio in an engagement process is not a fixed value, the change law of the transmission ratio is not limited by the number of the teeth, and the change range of the numerical value can be greatly improved. Therefore, the locking coefficient of the differential mechanism is improved, the pulsation and the impact are reduced simultaneously, and the cross-country requirements of cars are met. The invention is particularly suitable for various wheeled engineering machines of cross-country cars, dump trucks, wheel loaders, bulldozers and the like.

The invention provides a control method of Risley prism system applied to airborne infrared aided navigation. The method comprises the following steps: using sight lines to point to vector commands as input, using the reverse exact analysis algorithm based on the non-paraxial vector ray tracing algorithm, calculating the rotation angles of the prism which can accurately realize the pointing requirement of the sight lines, and using them as control model outputs. This method can effectively solve the problem in practical use that the rotation angles of the prism are on constant large expansion due to the continuous sight line scanning and the mutually independent calculation processes for the rotation angles of the prism corresponding to the vectors pointed to by scattered sight lines, therefore improving the stability and efficiency of prism rotation. The simulation and the implementation results show that the control method can achieve the aim of the invention, has the advantages of precise control of the pointing of sight lines and good control effect for prism rotation. At the same time, as a software control algorithm, the method is cost-effective and can be promoted for wider applications.

The invention discloses a high-entropy alloy selective laser melting process parameter optimization method. The method comprises the following steps: 1, printing a high-entropy alloy sample by adopting a selective laser melting method which is a cross-over experiment of randomly selecting or specifying three process parameters of power, exposure time and line-dot pitch; 2, detecting the relative density of the high-entropy alloy sample obtained in the step 1; 3, performing polynomial fitting on the relative density obtained in the step 2 and the corresponding three process parameters of the power, the exposure time and the line-dot pitch in the step 1, wherein a polynomial is composed of a cubic term, a quadratic term, a first term and an interactive term about the three parameters respectively; and 4, performing extreme value solving of the relative density on the fitted polynomial, wherein the power, the exposure time and the line-dot pitch corresponding to the extreme value of the relative density are the optimal process parameters. By using the method, an optimization result which is more accurate than that of a traditional optimization process can be obtained.

The invention provides a transaction information verification method and a transaction information verification device. The transaction information verification method comprises the following steps: receiving transaction information submitted by a user, and performing digital signature and encryption processing; transmitting encryption transaction information with a digital signature to an electronic bank server, verifying the digital signature and judging the legality of the transaction information through the electronic bank server, and after judging that the digital signature is qualified and the transaction information is legal, generating a transaction verification code according to the transaction information; receiving a transaction verification code and a server electronic signature fed back from the electronic bank server, and verifying the server electronic signature; after the verification is approved, prompting the user to input the verification code in a mode that button lamps are lightened or tuned off or the colors of the button lamps are changed through corresponding buttons according to the transaction verification code; judging whether the verification code input by the user through operation buttons is correct or not; after the verification code is correctly input, generating verification approval information, transmitting to the electronic bank server, and processing the transaction information through the electronic bank server according to the verification approval information, and completing the transaction.

A bearing suspension device includes first and second elastically flexible members extending in an axial direction; and first and second thrust points that are movable in the axial direction and that are in radial contact respectively with the first and second elastically flexible members. The bearing suspension device is configured to transmit a radial force by the first elastically flexible member bending between the first thrust point and a first end of the first elastically flexible member, and also to transmit the radial force by the second elastically flexible member bending between the second thrust point and a first end of the second elastically flexible member.

The invention discloses a variable coefficient underclocking model prediction direct power control method and device, and belongs to the electric power electron technical field; the method comprises the following steps: obtaining an output phase voltage and an output phase current of a three-phase grid connected inverter; converting the sampled phase voltage and phase current into a two-phase static coordinate system; calculating a variable coefficient underclocking model prediction direct power control object function J value corresponding to each output vector; determining the minimum value Jmin of the object function J values; obtaining the switch state combination of the vector corresponding to the minimum value Jmin, thus realizing the variable coefficient underclocking model prediction direct power control of the three-phase grid connected inverter. The method can effectively reduce the switch frequency while stably tracking the three phase grid connected inverter active power and reactive power, can reduce the switch loss, and has very well dynamic and static state performances.

The invention relates to a flow measuring device, in particular to a low-pressure-drop super acceleration gauge. The low-pressure-drop super acceleration gauge comprises a Venturi tube, and is characterized in that the Venturi tube is an inserted type external Venturi tube, a rectangular tubular generator is matched in the Venturi tube, and two pressure tapping tubes perpendicular to the axis direction of the Venturi tube and the rectangular tubular generator are matched with the Venturi tube and the rectangular tubular generator. The low-pressure-drop super acceleration gauge is unique in structure, solves the problems that accuracy is low and measuring range is narrow during low-Reynolds-number fluid measurement, solves the problems that under low pressure drop, measurement accuracy is low and measuring range is narrow, and achieves the purpose of installing and replacing a measuring instrument online.

The present invention provides a flow branch data rationality validation method. Flow branch data comprises a line parameter, a transformer parameter and a branch load rate. The method comprises the validation of the rationality of the line parameter, the transformer parameter and the branch load rate. According to the method, the branch load rate validation does not depend on flow convergence, the DC flow without convergence is used to estimate branch power, the similar result of branch power is obtained to carry out inspection, and the flow misconvergence caused by irrational branch load rate and the unable obtainment of a branch flow result are avoided. According to the method, a branch parameter with a large range and difficult direction inspection is translated into a visual line length and a transformer short-circuit voltage percentage with a typical range, the method thinking is clear, and the method is intuitive and practical.

The invention relates to a reset circuit. The reset circuit comprises a charging and discharging circuit, a follower circuit and a reset signal circuit. Wherein the charging and discharging circuit comprises a resistor R1, a capacitor C1 and a power supply port J1, and the reset signal circuit comprises a reset chip. When the circuit is started and powered on, the voltage of the capacitor C1 rises, the input voltage of the control end of the follower circuit rises, and the output voltage of the follower circuit rises along with the input voltage of the control end, namely, the voltage of the input end of the reset chip follows the voltage of the capacitor C1. The reset time can be adjusted by adjusting the time constant of the charging and discharging circuit. Due to the existence of the follower circuit, the resistor R1 is not influenced by the input impedance of the pin of the reset chip, the value range of the resistor R1 is increased, and the adjustment range of the reset time canbe further improved.

The invention discloses a pulse sequence delay correction method for magnetic resonance imaging. The method comprises the steps of conducting the delay correction before a pulse sequence is compiled into hardware codes; respectively merging continuous and identical output states in each channel on a pulse sequence generator; updating the output state based on merged data; and ensuring the holding time of the output state of each channel to meet the minimum pulse width requirement of the pulse sequence generator after the time delay correction. According to the technical scheme of the invention, the time delay correction is independent of a specific pulse sequence type, thus being strong in universality. The time delay correction is conducted before the pulse sequence is compiled into hardware codes, so that the application amount of hardware logic resources is reduced. On the premise that the hardware complexity is not increased and the time delay correction precision is guaranteed, the value range of the time delay correction is expanded.

The invention embodiment discloses a power control method and a mobile station. The method disclosed by the invention embodiment comprises the following steps: determining a closed loop power control value according to the corresponding relation of a code point built in advance and each closed loop power control value, wherein the code point consists of a TPC domain bit and a combination control signaling domain bit; and adjusting transmitting power by using the determined the closed loop power control value. The power can be adjusted more effectively through the invention embodiment.

A method for processing a digital image including image elements having a first luminance component. The first luminance component has a first value in a first predetermined interval. The display can render second values of second luminance components included in a second predetermined interval. The method includes: determining information representative of image brightness perceived by an observer, based on the first values of the first luminance component; calculating an expansion exponent as a function of the determined brightness information; transforming the first luminance components into the second luminance components, including calculating an intermediate luminance value by applying the calculated expansion exponent to the first luminance component value, multiplying the intermediate value by a length of the second interval, and obtaining the second value of the second luminance component based on the multiplied intermediate luminance value. The expansion exponent is based on a logarithm, of the determined brightness information.

The invention relates to a self-adaptive relation modeling method for structured data, and belongs to the technical field of artificial intelligence prediction. K*o exponential neurons are used for modeling a feature interaction relation between structured data attributes, K is the number of attention heads, and o is the number of exponential neurons of each attention head; all exponential neurons of each attention head share a bilinear attention function weight matrix W[att]; and the i exponential neuron yi of each attention head is expressed as formulas defined in the specification, a dynamic feature interaction weight w[i] is given by a formula w[i]=z[i]*v[i], z[i] is used as a gate, is dynamically generated through bilinear attention alignment scores and is expressed as formulas defined in the specification, alpha-entmax(.) is sparse softmax, and the sparsity is increased along with the increase of alpha. Compared with the prior art, the method has the advantages that input is not needed and must be positive, and the effectiveness of cross feature modeling is improved; the white box design is followed, and the modeling process is more transparent; the cross feature of any order are dynamically and selectively modeled according to the input data, so that the accuracy and the efficiency are improved; the global/local interpretability of the model is improved through the attention weight and the gating mechanism.