The embodiment of the invention discloses a method and a device for pushing information. A specific embodiment of the method comprises the steps of predicting a consumption amount of a client in a future set time period based on historical behavior data of the client; determining a product between the consumption amount and each first numerical value in a preset first numerical value set, taking the product as a target consumption amount corresponding to the first numerical value, and enabling the first numerical value to correspond to a second numerical value in advance; Generating first feedback information based on the target consumption amount and the second value corresponding to the first value in the first value set; and pushing the first feedback information to the client of the client. According to the embodiment of the invention, targeted information pushing is realized.

Method and device for pushing information

The invention discloses a user identification method and device, a storage medium and a processor. The method comprises: acquiring multi-dimensional feature information of a user; converting the multi-dimensional feature information into a multi-dimensional array, the numerical value in the multi-dimensional array being within a predetermined value interval; and analyzing the multi-dimensional array according to a predetermined machine learning model, identifying a user type corresponding to the multi-dimensional array, the predetermined machine learning model being obtained by training at least one group of training data through machine learning, and each group of training data comprising the multi-dimensional array and the user type corresponding to the multi-dimensional array. The technical problems that when a traditional machine learning algorithm carries out user classification, feature extraction is difficult and inaccurate, the learning efficiency is low, and overfitting is likely to happen are solved.

User identification method and device, storage medium and processor

The disclosed embodiment discloses a method, an apparatus, an electronic device and a readable storage medium for determining object characteristic parameters. The method comprises: acquiring a firstobject portrait of a first object to generate a training set; Training the training set to establish a second object acquisition success rate model of the second object acquisition success rate underdifferent characteristic parameters of the first object; Calculating a characteristic parameter sensitivity of the first object related to acquiring a second object according to the second object acquisition success rate model; Calculating a characteristic parameter of the first object related to the acquisition of the second object based on a sensitivity of the characteristic parameter of the first object related to the acquisition of the second object, The first object can be directly used to model the acquisition success rate data of the second object without relying on artificial experience and manual annotation, and then the sensitivity of the first object's characteristic parameters can be deduced to differentiate the characteristic parameters.

Object characteristic parameter determination method, device, electronic device, and readable storage medium

A method and device for predicting a foreign exchange transaction volume. The method comprises: obtaining time series data consisting of foreign exchange transaction business data series (102), wherein each foreign exchange transaction business data series comprises foreign exchange transaction business data of multiple dimensions within the same business day; and then, inputting the time series data to a pre-trained multilayer convolutional neural network to predict a foreign exchange transaction volume in a future time period (104). The mutual influence of the foreign exchange transaction business data between adjacent time nodes is considered during prediction, and the multi-dimensional characteristic attribute of a single time node is utilized, thereby improving the prediction accuracy.

Method and device for predicting foreign exchange transaction volume

The invention relates to a consumer consumption behavior prediction method based on a multi-factor recurrent neural network. The method comprises: 1) performing data preprocessing to obtain training data, standardizing four sequences S, H, C, T to obtain training data X= [SHCT]; 2) generating a recurrent neural network mode including an input layer X, a hidden layer Z and an output layer Y; 3) using the BPTT algorithm and the training data to train the model, and using the BPTT algorithm to iteratively update three weight values (U, V, W) and two deviation values (bz, by) according to the gradient descent method; and 4) using the trained neural network to calculate the output sequence Y and restore the output sequence to the final prediction result sT+1. The invention has the beneficial effects that the method considers some natural factors that affect the consumption behavior, and can accurately and efficiently predict offline consumption behaviors of different users; in addition to the three natural factors proposed by the invention, the method can be expanded to continue to introduce other factors that may affect the offline consumer behaviors of the users.

Consumer consumption behavior prediction method based on multi-factor recurrent neural network

The embodiment of the invention provides a consumption ability prediction method and apparatus, an electronic device, and a readable storage medium, and relates to the technical field of computers. The consumption ability prediction method includes the steps: acquiring the statistical characteristic data and the time sequence characteristic data of the target object from the historical data of thetarget user, based on the statistical characteristic data and the time sequence characteristic data, and utilizing the preset hybrid neural network prediction model to determine the consumption ability value of the target user for the target object. The consumption ability prediction method can solve the problem that the prior art utilizes the price of the commodity which is purchased by the userat the last time, the price of the commodity which is purchased randomly, or the price mean value of the commodities which are purchased in the history to determine the consumption ability value of the user, thus being lower in the accuracy. The consumption ability prediction method and apparatus combines with the time sequence characteristic data on the basis of the statistical characteristic data so as to be able to realize sequential dimension characteristic extraction of the historical data to enable the consumption ability value which is predicted by the hybrid neural network predictionmodel to be more accurate.

Consumption ability prediction method and apparatus, electronic device, and readable storage medium

The invention provides a stimulating behavior sensitivity prediction method and device, equipment and a readable storage medium. The method comprises the steps that a stimulating behavior characteristic combination and a non-stimulating behavior characteristic combination are acquired according to the subject data of a target subject; the stimulating behavior characteristic combination and the non-stimulating behavior characteristic combination are input into a target parameter prediction model for a target object, so as to acquire a stimulating behavior target parameter and a non-stimulatingbehavior target parameter; and according to the difference between the stimulating behavior target parameter and the non-stimulating behavior target parameter, the stimulating behavior sensitivity ofthe target subject to the target object is determined. The problems that the sensitivity accuracy is low and a preset effect cannot be effectively realized when a statistical rule is used to predict the stimulating behavior of the subject in the prior art are solved.

Stimulating behavior sensitivity prediction method, device and equipment and readable storage medium

An embodiment of the present invention relates to the technical field of computers, and provided thereby are a method and apparatus for predicting spending power, as well as an electronic device and a readable storage medium. According to an embodiment of the described method, by means of acquiring from among historical data of a target user one or more pieces of statistical feature data and one or more pieces of timing sequence feature data for a target object, a preset hybrid neural network prediction model may, on the basis of the one or more pieces of statistical feature data and the one or more pieces of timing sequence feature data, be used to determine the spending power of the target user with respect to the target object.

Wang Ziwei

Papers

Consumption ability prediction method and apparatus, electronic device, and readable storage medium

Stimulating behavior sensitivity prediction method, device and equipment and readable storage medium

Prediction of spending power