Financial product matching method and device, computer equipment, storage medium and computer program product

Abstract

The invention relates to a financial product matching method and device, computer equipment, a storage medium and a computer program product. The method comprises the steps of performing feature extraction processing on preprocessed object information of a to-be-analyzed object through a pre-trained object type recognition model to obtain a first information feature and a second information feature corresponding to the to-be-analyzed object, and performing fusion processing for multiple times to obtain a fused information feature corresponding to the to-be-analyzed object; according to the first information feature, determining financial product acquisition probability information of the to-be-analyzed object, and according to the fusion information feature, determining financial product resource transfer information of the to-be-analyzed object; and determining a target object type of the to-be-analyzed object according to the financial product acquisition probability information and the financial product resource transfer information, and determining a target object matched with the to-be-analyzed financial product from the to-be-analyzed object according to the target object type. By adopting the method, the accuracy of financial product matching can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a financial product matching method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] At present, in order to ensure the stable operation of the insurance industry, it is crucial to accurately match financial products with corresponding customers.

[0003] In traditional technology, in the process of matching financial products, manual matching is generally adopted; however, this manual matching method has subjective factors and is prone to errors, resulting in a low accuracy rate in financial product matching. Summary of the invention

[0004] Based on this, it is necessary to provide a financial product matching method, apparatus, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of financial product matching in order to address the above technical problems.

[0005] In a first aspect, the present application provides a financial product matching method, comprising:

[0006] Obtaining object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed;

[0007] Preprocessing the object information to obtain preprocessed object information;

[0008] Through the pre-trained object type recognition model, feature extraction processing is performed on the pre-processed object information to obtain first information features and second information features corresponding to the object to be analyzed, and the first information features and the second information features are fused multiple times to obtain fused information features corresponding to the object to be analyzed;

[0009] Determine the financial product acquisition probability information of the object to be analyzed based on the first information feature, and determine the financial product resource transfer information of the object to be analyzed based on the fusion information feature;

[0010] The target object type of the object to be analyzed is determined according to the financial product acquisition probability information and the financial product resource transfer information, and the target object matching the financial product to be analyzed is determined from the objects to be analyzed according to the target object type.

[0011] In one embodiment, the performing multiple fusion processes on the first information feature and the second information feature to obtain the fused information feature corresponding to the object to be analyzed includes:

[0012] fusing the first information feature and the second information feature to obtain a third information feature corresponding to the object to be analyzed;

[0013] Performing feature extraction processing on the first information feature to obtain a fourth information feature corresponding to the object to be analyzed, and performing feature extraction processing on the third information feature to obtain a fifth information feature corresponding to the object to be analyzed;

[0014] The fourth information feature is used as a new first information feature, the fifth information feature is used as a new second information feature, and the process proceeds to the step of fusing the first information feature and the second information feature to obtain a third information feature corresponding to the object to be analyzed, until the obtained third information feature meets a preset condition;

[0015] When the obtained third information feature meets the preset condition, the obtained third information feature is used as the fused information feature.

[0016] In one embodiment, the pre-trained object type recognition model is trained in the following manner:

[0017] Acquire sample object information of the sample object; the sample object information includes at least basic attribute information and financial product information of the sample object;

[0018] Preprocessing the sample object information to obtain preprocessed sample object information;

[0019] By using the object type recognition model to be trained, feature extraction processing is performed on the preprocessed sample object information respectively to obtain a first sample information feature and a second sample information feature corresponding to the sample object, and the first sample information feature and the second sample information feature are fused multiple times to obtain a sample fusion information feature corresponding to the sample object;

[0020] Determine the predicted financial product acquisition probability information of the sample object according to the first sample information feature, and determine the predicted financial product resource transfer information of the sample object according to the sample fusion information feature;

[0021] The actual financial product acquisition probability information and the actual financial product resource transfer information of the sample object are obtained, and according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, the object type recognition model to be trained is iteratively trained to obtain the pre-trained object type recognition model.

[0022] In one embodiment, the iterative training of the object type recognition model to be trained based on the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and based on the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, to obtain the pre-trained object type recognition model includes:

[0023] Obtaining a first loss value according to a difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and obtaining a second loss value according to a difference between the predicted financial product resource transfer information and the actual financial product resource transfer information;

[0024] fusing the first loss value and the second loss value to obtain a target loss value;

[0025] According to the target loss value, the object type recognition model to be trained is iteratively trained to obtain the pre-trained object type recognition model.

[0026] In one embodiment, the preprocessing the object information to obtain the preprocessed object information includes:

[0027] Obtaining application scenario information corresponding to the object to be analyzed;

[0028] Determining the correlation between each object information and the application scenario information;

[0029] From each of the object information, identifying object information whose correlation with the application scenario information is less than a preset correlation, as the object information to be processed;

[0030] The object information to be processed in each of the object information is deleted to obtain the pre-processed object information.

[0031] In one embodiment, determining the target object type of the object to be analyzed according to the financial product acquisition probability information and the financial product resource transfer information includes:

[0032] Obtaining multiple prediction object types of the object to be analyzed and the prediction probability of each prediction object type according to the financial product acquisition probability information and the financial product resource transfer information;

[0033] From each of the prediction object types, the prediction object type with the largest prediction probability is screened out as the target object type.

[0034] In a second aspect, the present application also provides a financial product matching device, including:

[0035] An information acquisition module, used to acquire object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed;

[0036] An information processing module, used for preprocessing the object information to obtain preprocessed object information;

[0037] A feature extraction module is used to perform feature extraction processing on the preprocessed object information respectively through a pre-trained object type recognition model to obtain a first information feature and a second information feature corresponding to the object to be analyzed, and perform multiple fusion processing on the first information feature and the second information feature to obtain a fused information feature corresponding to the object to be analyzed;

[0038] an information determination module, configured to determine the financial product acquisition probability information of the object to be analyzed according to the first information feature, and to determine the financial product resource transfer information of the object to be analyzed according to the fusion information feature;

[0039] The product matching module is used to determine the target object type of the object to be analyzed according to the financial product acquisition probability information and the financial product resource transfer information, and determine the target object matching the financial product to be analyzed from the objects to be analyzed according to the target object type.

[0040] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0041] Obtaining object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed;

[0042] Preprocessing the object information to obtain preprocessed object information;

[0043] Through the pre-trained object type recognition model, feature extraction processing is performed on the pre-processed object information to obtain first information features and second information features corresponding to the object to be analyzed, and the first information features and the second information features are fused multiple times to obtain fused information features corresponding to the object to be analyzed;

[0044] Determine the financial product acquisition probability information of the object to be analyzed based on the first information feature, and determine the financial product resource transfer information of the object to be analyzed based on the fusion information feature;

[0045] The target object type of the object to be analyzed is determined according to the financial product acquisition probability information and the financial product resource transfer information, and the target object matching the financial product to be analyzed is determined from the objects to be analyzed according to the target object type.

[0046] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0047] Obtaining object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed;

[0048] Preprocessing the object information to obtain preprocessed object information;

[0049] Through the pre-trained object type recognition model, feature extraction processing is performed on the pre-processed object information to obtain first information features and second information features corresponding to the object to be analyzed, and the first information features and the second information features are fused multiple times to obtain fused information features corresponding to the object to be analyzed;

[0050] Determine the financial product acquisition probability information of the object to be analyzed based on the first information feature, and determine the financial product resource transfer information of the object to be analyzed based on the fusion information feature;

[0051] The target object type of the object to be analyzed is determined according to the financial product acquisition probability information and the financial product resource transfer information, and the target object matching the financial product to be analyzed is determined from the objects to be analyzed according to the target object type.

[0052] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0053] Obtaining object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed;

[0054] Preprocessing the object information to obtain preprocessed object information;

[0055] Through the pre-trained object type recognition model, feature extraction processing is performed on the pre-processed object information to obtain first information features and second information features corresponding to the object to be analyzed, and the first information features and the second information features are fused multiple times to obtain fused information features corresponding to the object to be analyzed;

[0056] Determine the financial product acquisition probability information of the object to be analyzed based on the first information feature, and determine the financial product resource transfer information of the object to be analyzed based on the fusion information feature;

[0057] The target object type of the object to be analyzed is determined according to the financial product acquisition probability information and the financial product resource transfer information, and the target object matching the financial product to be analyzed is determined from the objects to be analyzed according to the target object type.

[0058] The above-mentioned financial product matching method, apparatus, computer equipment, storage medium and computer program product first obtain basic attribute information and financial product information of the object to be analyzed as object information of the object to be analyzed, and then pre-process the object information to obtain pre-processed object information. Then, through the pre-trained object type recognition model, feature extraction processing is performed on the pre-processed object information respectively to obtain the first information feature and the second information feature corresponding to the object to be analyzed, and the first information feature and the second information feature are fused for multiple times to obtain the fused information feature corresponding to the object to be analyzed. Then, based on the first information feature, the financial product acquisition probability information of the object to be analyzed is determined, and based on the fused information feature, the financial product resource transfer information of the object to be analyzed is determined. Finally, based on the financial product acquisition probability information and the financial product resource transfer information, the target object type of the object to be analyzed is determined, and based on the target object type, the target object matching the financial product to be analyzed is determined from the object to be analyzed. In this way, in the process of matching financial products, by obtaining the object information of the object to be analyzed from multiple different sources, misjudgment caused by the one-sidedness of a single information source is avoided, so that the subsequent preprocessing and feature extraction processing of the object to be analyzed is established on a more comprehensive basis, and then based on the accurate target object type of the object to be analyzed, and based on the target object type, the target object that matches the financial product to be analyzed can be accurately determined from the object to be analyzed, which is conducive to improving the accuracy of financial product matching; moreover, the entire process does not require human intervention, avoiding the subjective factors in the manual matching method, which is prone to errors and leads to the defect of low accuracy of financial product matching, and further improves the accuracy of financial product matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 A flowchart of a financial product matching method in one embodiment;

[0061] Figure 2 A schematic diagram of the structure of an object type identification model in one embodiment;

[0062] Figure 3 A flowchart of a financial product matching method in another embodiment;

[0063] Figure 4 A flowchart of a method for digital management of insurance customers based on artificial intelligence and big data in one embodiment;

[0064] Figure 5 is a structural block diagram of a financial product matching device in one embodiment;

[0065] Figure 6 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0067] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0068] In an exemplary embodiment, Figure 1 As shown, a financial product matching method is provided. This embodiment uses the method applied to a server as an example for illustration. It is understandable that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smart phones, and tablet computers; the server can be implemented as an independent server or a server cluster consisting of multiple servers. In this embodiment, the method includes the following steps:

[0069] Step S101, obtaining object information of an object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed.

[0070] The object to be analyzed refers to an object whose object type needs to be identified. In an actual scenario, the object to be analyzed refers to a customer whose object type needs to be identified.

[0071] The object information includes at least basic attribute information and financial product information of the object to be analyzed.

[0072] Among them, basic attribute information refers to the basic user information of the object to be analyzed, including the name, gender, date of birth, mobile phone number, marital status, customer star rating, whether the policyholder is an insured person, and other basic information of the object to be analyzed.

[0073] Among them, financial product information refers to information related to the financial products to be analyzed, including salesperson information, historical policy information, user claims information and other information; salesperson information includes whether the salesperson is employed, historical policy records, customer visit records, salesperson praise scores and other information; historical policy information includes policy number, policy type, policy name, number of policy physical examinations, policy effective status, policy value and other information; user claims information includes claim time, claim application amount, claim amount, claim policy number and other information.

[0074] Exemplarily, in response to a financial product matching instruction, the server obtains basic attribute information and financial product information of the object to be analyzed from a financial database; then, the server combines the basic attribute information and the financial product information in a preset combination method to obtain the object information of the object to be analyzed.

[0075] Step S102: preprocess the object information to obtain preprocessed object information.

[0076] The preprocessed object information refers to the object information after preprocessing.

[0077] Exemplarily, the server performs preliminary cleaning and processing of the object information (including removing noise data, filling missing values, discrete feature encoding, data interval scaling, etc.) to obtain pre-processed object information.

[0078] Step S103, using a pre-trained object type recognition model, feature extraction processing is performed on the pre-processed object information to obtain first information features and second information features corresponding to the object to be analyzed, and the first information features and the second information features are fused multiple times to obtain fused information features corresponding to the object to be analyzed.

[0079] Among them, the object type recognition model refers to a network model that can use the pre-processed object information of the object to be analyzed to obtain the financial product acquisition probability information and financial product resource transfer information corresponding to the object to be analyzed, such as a neural network model.

[0080] The first information feature refers to the information feature corresponding to the probability information of obtaining financial products and the financial product resource transfer information of the object to be analyzed.

[0081] The second information feature refers to the information feature corresponding to the financial product resource transfer information of the object to be analyzed that is used to calculate the information.

[0082] The fused information feature refers to an information feature obtained by fusing the first information feature and the second information feature multiple times.

[0083] Exemplarily, the server inputs the preprocessed object information into a pre-trained object type recognition model, and performs feature extraction processing on the pre-processed object information through the pre-trained object type recognition model to obtain first information features and second information features corresponding to the object to be analyzed; then, the server performs multiple fusion processing on the first information features and the second information features to obtain fused information features corresponding to the object to be analyzed.

[0084] For example, see Figure 2 The server inputs the preprocessed object information into the pre-trained object type recognition model, obtains the first information feature corresponding to the object to be analyzed through network unit 1, network unit 2, ..., network unit n, network unit n+1 in the object type recognition model, and obtains the second information feature corresponding to the object to be analyzed through network unit 1, network unit 2, ..., network unit n, network unit n+4 in the object type recognition model; then, the server performs multiple fusion processing on the first information feature and the second information feature to obtain the fused information feature corresponding to the object to be analyzed.

[0085] Step S104: determining the financial product acquisition probability information of the object to be analyzed based on the first information feature, and determining the financial product resource transfer information of the object to be analyzed based on the fusion information feature.

[0086] The financial product acquisition probability information is used to represent the transaction probability of the financial product to be analyzed.

[0087] The financial product resource transfer information is used to represent the insurance policy value corresponding to the transaction of the financial product to be analyzed. It should be noted that the financial product resource transfer information is also called customer value.

[0088] Exemplarily, the server determines, based on the first information feature, the financial product acquisition probability information corresponding to the first information feature, as the financial product acquisition probability information of the object to be analyzed; then, the server determines, based on the fused information feature, the financial product resource transfer information corresponding to the fused information feature, as the financial product resource transfer information of the object to be analyzed.

[0089] For example, see Figure 2 , through network unit n+2, network unit n+3 and Softmax (normalized exponential function) layer, the transaction probability corresponding to the first information feature is obtained as the transaction probability of the object to be analyzed; then, through network unit n+6, the customer value corresponding to the fused information feature is obtained as the customer value of the object to be analyzed.

[0090] Step S105, determining the target object type of the object to be analyzed according to the financial product acquisition probability information and the financial product resource transfer information, and determining the target object matching the financial product to be analyzed from the objects to be analyzed according to the target object type.

[0091] The target object type refers to the object type of the object to be analyzed.

[0092] The financial products to be analyzed refer to the financial products that need to be matched.

[0093] The target object refers to the object to be analyzed that matches the financial product to be analyzed.

[0094] Exemplarily, the server queries the correspondence between the financial product acquisition probability information, the financial product resource transfer information and the object type based on the financial product acquisition probability information and the financial product resource transfer information, and obtains the object type corresponding to the financial product acquisition probability information and the financial product resource transfer information as the target object type of the object to be analyzed; for example, the server takes the object whose financial product acquisition probability information is greater than the preset probability information and whose financial product resource transfer information is greater than the preset resource transfer information as the target object (such as a high-potential customer who is easy to place an order); then, the server determines the object level information of the object to be analyzed based on the target object type of the object to be analyzed; for example, the server takes the high-potential customer who is easy to place an order as a high-level customer, takes the low-potential customer who is easy to place an order and the high-potential customer who is difficult to place an order as a mid-level customer, and takes the low-potential customer who is difficult to place an order as a low-level customer; then, the server determines the target object that matches the financial product to be analyzed from the objects to be analyzed based on the object level information of the objects to be analyzed and the product information of the financial product to be analyzed.

[0095] In the above financial product matching method, the basic attribute information and financial product information of the object to be analyzed are first obtained as the object information of the object to be analyzed, and then the object information is preprocessed to obtain the preprocessed object information. Then, the pre-trained object type recognition model is used to perform feature extraction processing on the preprocessed object information to obtain the first information feature and the second information feature corresponding to the object to be analyzed, and the first information feature and the second information feature are fused for multiple times to obtain the fused information feature corresponding to the object to be analyzed. Then, the financial product acquisition probability information of the object to be analyzed is determined based on the first information feature, and the financial product resource transfer information of the object to be analyzed is determined based on the fused information feature. Finally, the target object type of the object to be analyzed is determined based on the financial product acquisition probability information and the financial product resource transfer information, and the target object matching the financial product to be analyzed is determined from the object to be analyzed based on the target object type. In this way, in the process of matching financial products, by obtaining the object information of the object to be analyzed from multiple different sources, misjudgment caused by the one-sidedness of a single information source is avoided, so that the subsequent preprocessing and feature extraction processing of the object to be analyzed is established on a more comprehensive basis, and then based on the accurate target object type of the object to be analyzed, and based on the target object type, the target object that matches the financial product to be analyzed can be accurately determined from the object to be analyzed, which is conducive to improving the accuracy of financial product matching; moreover, the entire process does not require human intervention, avoiding the subjective factors in the manual matching method, which is prone to errors and leads to the defect of low accuracy of financial product matching, and further improves the accuracy of financial product matching.

[0096] In an exemplary embodiment, the above-mentioned step S103, in which the first information feature and the second information feature are fused multiple times to obtain a fused information feature corresponding to the object to be analyzed, specifically includes the following contents: fusing the first information feature and the second information feature to obtain a third information feature corresponding to the object to be analyzed; performing feature extraction processing on the first information feature to obtain a fourth information feature corresponding to the object to be analyzed, and performing feature extraction processing on the third information feature to obtain a fifth information feature corresponding to the object to be analyzed; using the fourth information feature as a new first information feature and the fifth information feature as a new second information feature, and jumping to the step of fusing the first information feature and the second information feature to obtain the third information feature corresponding to the object to be analyzed, until the obtained third information feature meets the preset condition; when the obtained third information feature meets the preset condition, the obtained third information feature is used as the fused information feature.

[0097] The third information feature refers to the information feature obtained by fusing the first information feature and the second information feature.

[0098] The fourth information feature refers to the information feature obtained by performing feature extraction processing on the first information feature.

[0099] The fifth information feature refers to the information feature obtained by performing feature extraction processing on the third information feature.

[0100] The preset condition refers to a pre-set condition, such as a preset number, which is used to judge the obtained third information feature. If the obtained third information feature meets the preset condition, that is, the obtained third information feature meets the preset number, such as the obtained third information feature is the second third information feature, the obtained third information feature is used as the fusion information feature; if the obtained third information feature does not meet the preset condition, that is, the obtained third information feature does not meet the preset number, such as the obtained third information feature is not the second third information feature, the first information feature is subjected to feature extraction processing to obtain the fourth information feature corresponding to the object to be analyzed, and the third information feature is subjected to feature extraction processing to obtain the fifth information feature corresponding to the object to be analyzed; the fourth information feature is used as the new first information feature, the fifth information feature is used as the new second information feature, and the process jumps to the step of fusing the first information feature and the second information feature to obtain the third information feature corresponding to the object to be analyzed.

[0101] Exemplarily, the server fuses the first information feature and the second information feature to obtain a fused feature corresponding to the first information feature and the second information feature as the third information feature corresponding to the object to be analyzed; then, the server judges the obtained third information feature; if the obtained third information feature does not meet the preset conditions, the server performs feature extraction processing on the first information feature to obtain a fourth information feature corresponding to the object to be analyzed, and performs feature extraction processing on the third information feature to obtain a fifth information feature corresponding to the object to be analyzed; then, the server uses the fourth information feature as the new first information feature and the fifth information feature as the new second information feature, and jumps to the step of fusing the first information feature and the second information feature to obtain the third information feature corresponding to the object to be analyzed, until the obtained third information feature meets the preset conditions; finally, when the obtained third information feature meets the preset conditions, the server uses the obtained third information feature as the fused information feature.

[0102] For example, see Figure 2 , through network unit n+2, feature extraction processing is performed on the first information feature to obtain a fourth information feature corresponding to the object to be analyzed; through network unit n+5, feature extraction processing is performed on the third information feature to obtain a fifth information feature corresponding to the object to be analyzed.

[0103] In this embodiment, the third information feature is obtained by fusing the first information feature and the second information feature, and then new features are continuously extracted on this basis and the extracted features are used as new inputs for re-fusion in a loop. This enables deep fusion and interaction between features at different levels, avoiding the limitations of single features or simple fusion features in describing objects.

[0104] In an exemplary embodiment, the financial product matching method provided by the present application also includes a training step of a pre-trained object type recognition model, which specifically includes the following contents: obtaining sample object information of the sample object; the sample object information includes at least basic attribute information and financial product information of the sample object; pre-processing the sample object information to obtain pre-processed sample object information; performing feature extraction processing on the pre-processed sample object information through the object type recognition model to be trained, respectively, to obtain a first sample information feature and a second sample information feature corresponding to the sample object, and performing multiple fusion processing on the first sample information feature and the second sample information feature to obtain a sample fusion information feature corresponding to the sample object; determining the predicted financial product acquisition probability information of the sample object according to the first sample information feature, and determining the predicted financial product resource transfer information of the sample object according to the sample fusion information feature; obtaining the actual financial product acquisition probability information and the actual financial product resource transfer information of the sample object, and iteratively training the object type recognition model to be trained according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, to obtain a pre-trained object type recognition model.

[0105] The sample object refers to an object used to train the object type recognition model.

[0106] The sample object information refers to the object information of the sample object.

[0107] The preprocessed sample object information refers to the sample object information after preprocessing.

[0108] The first sample information feature refers to the information feature corresponding to the predicted financial product acquisition probability information and the predicted financial product resource transfer information used to calculate the sample object.

[0109] The second sample information feature refers to the information feature corresponding to the predicted financial product resource transfer information used to calculate the sample object.

[0110] The sample fusion information feature refers to an information feature obtained by fusing the first sample information feature and the second sample information feature multiple times.

[0111] The predicted financial product acquisition probability information refers to the predicted value corresponding to the financial product acquisition probability information of the sample object.

[0112] The predicted financial product resource transfer information refers to the predicted value corresponding to the financial product resource transfer information of the sample object.

[0113] The actual financial product acquisition probability information refers to the actual value corresponding to the financial product acquisition probability information of the sample object.

[0114] The actual financial product resource transfer information refers to the actual value corresponding to the financial product resource transfer information of the sample object.

[0115] Exemplarily, in response to a model training instruction for an object type recognition model to be trained, the server obtains basic attribute information and financial product information of a sample object from a financial database as sample object information of the sample object; then, the server preprocesses the sample object information to obtain preprocessed sample object information; then, the server inputs the preprocessed sample object information into the object type recognition model to be trained, and performs feature extraction processing on the preprocessed sample object information through the object type recognition model to be trained, respectively, to obtain a first sample information feature and a second sample information feature corresponding to the sample object, and performs multiple fusion processing on the first sample information feature and the second sample information feature to obtain a sample fusion information feature corresponding to the sample object; then, the server performs a fusion processing on the first sample information feature and the second sample information feature according to the first sample information feature. Based on a sample information feature, the predicted financial product acquisition probability information corresponding to the first sample information feature is determined as the predicted financial product acquisition probability information of the sample object, and based on the sample fusion information feature, the predicted financial product resource transfer information corresponding to the sample fusion information feature is determined as the predicted financial product resource transfer information of the sample object; then, the server obtains the actual financial product acquisition probability information and the actual financial product resource transfer information of the sample object, and according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, iteratively trains the object type recognition model to be trained to obtain a pre-trained object type recognition model.

[0116] In this embodiment, by pre-training the object type recognition model, it is convenient to predict the target object type of the object to be analyzed after obtaining the object information of the object to be analyzed in actual applications; moreover, the object type recognition model receives new data in each round of iteration, and performs internal improvements and optimizations on the model, so as to be able to make predictions more effectively, which is beneficial to improving the prediction accuracy of the object type recognition model.

[0117] In an exemplary embodiment, according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, the object type recognition model to be trained is iteratively trained to obtain a pre-trained object type recognition model, which specifically includes the following contents: according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, a first loss value is obtained, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, a second loss value is obtained; the first loss value and the second loss value are fused to obtain a target loss value; according to the target loss value, the object type recognition model to be trained is iteratively trained to obtain a pre-trained object type recognition model.

[0118] The first loss value refers to the loss value obtained based on the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information.

[0119] The second loss value refers to the loss value obtained according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information.

[0120] The target loss value refers to the loss value obtained by fusing the first loss value and the second loss value.

[0121] Exemplarily, the server obtains a first loss value based on the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information; then, the server obtains a second loss value based on the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information; then, the server obtains a first weight corresponding to the first loss value, and a second weight corresponding to the second loss value; then, the server fuses the first loss value and the second loss value according to the first weight and the second weight to obtain a target loss value; then, the server adjusts the model parameters of the object type recognition model to be trained according to the target loss value; then, the server re-trains the object type recognition model after the model parameters are adjusted until the target loss value obtained by the trained object type recognition model is less than the loss value threshold, then stops training, and uses the trained object type recognition model as the pre-trained object type recognition model.

[0122] For example, the server can obtain the target loss value through the following formula:

[0123] , formula (1)

[0124] Among them, i represents the i-th sample object, n represents the total number of sample objects, and p iRefers to the probability information of actual financial products, p i ' refers to the probability information of predicting financial products, δ is the loss adjustment coefficient, v i Refers to the actual financial product resource transfer information, v i ' refers to predicting financial product resource transfer information.

[0125] In this embodiment, the loss value is calculated by taking into account the differences in two different dimensions, namely, the financial product acquisition probability information and the financial product resource transfer information. This avoids the one-sidedness problem that may be caused by evaluating the model from only a single perspective, thereby making the evaluation of the model more accurate, thereby improving the training accuracy of the object type recognition model.

[0126] In an exemplary embodiment, the above step S102 preprocesses the object information to obtain preprocessed object information, which specifically includes the following contents: obtaining application scenario information corresponding to the object to be analyzed; determining the correlation between each object information and the application scenario information; identifying, from each object information, object information whose correlation with the application scenario information is less than a preset correlation, as the object information to be processed; deleting the object information to be processed in each object information to obtain the preprocessed object information.

[0127] The application scenario information is used to represent the scenario information of the object to be analyzed. In the actual scenario, the application scenario information refers to the insurance business.

[0128] The correlation degree is used to indicate the degree of correlation between each object information and the application scenario information. For example, in an actual scenario, the correlation degree between the user name of the object to be analyzed and the insurance business is 20%, and the correlation degree between the date of birth of the object to be analyzed and the insurance business is 90%.

[0129] The preset correlation refers to a preset correlation threshold, such as 80%. It should be noted that the preset correlation depends on the situation.

[0130] Exemplarily, the server obtains application scenario information corresponding to the object to be analyzed; then, the server inputs each object information and application scenario information into a pre-trained correlation prediction model to obtain the correlation between each object information and the application scenario information; then, the server identifies object information whose correlation with the application scenario information is less than a preset correlation from each object information, and uses this object information as object information to be processed; then, the server deletes the object information to be processed in each object information to obtain pre-processed object information.

[0131] In this embodiment, by determining the correlation between each object information and the application scenario information, and eliminating the object information with a correlation less than a preset correlation, redundant data that has little correlation with the core application scenario of the insurance business can be removed, thereby avoiding interference from irrelevant information and further improving the pertinence and effectiveness of the analysis.

[0132] In an exemplary embodiment, the above step S105 determines the target object type of the object to be analyzed based on the financial product acquisition probability information and the financial product resource transfer information, and specifically includes the following contents: according to the financial product acquisition probability information and the financial product resource transfer information, multiple prediction object types of the object to be analyzed and the prediction probability of each prediction object type are obtained; from each prediction object type, the prediction object type with the largest prediction probability is screened out as the target object type.

[0133] The predicted object type refers to the predicted value corresponding to the object type of the object to be analyzed.

[0134] The prediction probability is used to indicate the possibility that the object type recognition model determines that the predicted object type is correct.

[0135] Exemplarily, the server uses a pre-trained object type recognition model to obtain probability information of financial products and resource transfer information of financial products to obtain multiple predicted object types of the object to be analyzed and the predicted probability of each predicted object type; then, the server filters out the predicted object type with the largest predicted probability from each predicted object type, and uses the predicted object type as the target object type.

[0136] In this embodiment, the object with the highest prediction probability is selected from each predicted object type as the target object type, which is equivalent to determining the object type that best matches the object to be analyzed from multiple predicted object types, which is beneficial to improving the accuracy of determining the target object type of the object to be analyzed and facilitates subsequent data analysis and processing.

[0137] In an exemplary embodiment, Figure 3 As shown, another financial product matching method is provided, which is described by taking the method applied to a server as an example, and includes the following steps:

[0138] Step S301, obtaining object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed.

[0139] Step S302, obtain application scenario information corresponding to the object to be analyzed; determine the correlation between each object information and the application scenario information; identify from each object information the object information whose correlation with the application scenario information is less than a preset correlation, as the object information to be processed; delete the object information to be processed in each object information to obtain the pre-processed object information.

[0140] Step S303: Perform feature extraction processing on the preprocessed object information through a pre-trained object type recognition model to obtain first information features and second information features corresponding to the object to be analyzed.

[0141] Step S304: fusing the first information feature and the second information feature to obtain a third information feature corresponding to the object to be analyzed.

[0142] Step S305: perform feature extraction processing on the first information feature to obtain a fourth information feature corresponding to the object to be analyzed, and perform feature extraction processing on the third information feature to obtain a fifth information feature corresponding to the object to be analyzed.

[0143] Step S306: Use the fourth information feature as a new first information feature, use the fifth information feature as a new second information feature, and jump to step S304 until the obtained third information feature meets the preset condition.

[0144] Step S307: When the obtained third information feature meets a preset condition, the obtained third information feature is used as a fused information feature.

[0145] Step S308: determining the financial product acquisition probability information of the object to be analyzed based on the first information feature, and determining the financial product resource transfer information of the object to be analyzed based on the fusion information feature.

[0146] Step S309 , obtaining multiple prediction object types of the object to be analyzed and the prediction probability of each prediction object type according to the financial product acquisition probability information and the financial product resource transfer information.

[0147] Step S310 , selecting the prediction object type with the highest prediction probability from each prediction object type as the target object type.

[0148] Step S311: determining target objects that match the financial product to be analyzed from the objects to be analyzed according to the target object type.

[0149] In the above-mentioned financial product matching method, in the process of matching financial products, by obtaining the object information of the object to be analyzed from multiple different sources, misjudgment caused by the one-sidedness of a single information source is avoided, so that the subsequent preprocessing and feature extraction processing of the object to be analyzed are established on a more comprehensive basis, and then based on the accurate target object type of the object to be analyzed, and based on the target object type, the target object that matches the financial product to be analyzed can be accurately determined from the object to be analyzed, which is conducive to improving the accuracy of financial product matching; moreover, the entire process does not require human intervention, avoiding the subjective factors in the manual matching method, which is prone to errors and leads to the defect of low accuracy of financial product matching, and further improving the accuracy of financial product matching.

[0150] In an exemplary embodiment, in order to more clearly illustrate the financial product matching method provided in the embodiment of the present application, the financial product matching method is specifically described below with a specific embodiment. Figure 4 As shown, this application also provides a digital management method for insurance customers based on artificial intelligence and big data. In the process of matching financial products, the user's basic information, salesperson information, historical policy information, user claim information and other data are first obtained, and then the data is subjected to feature preprocessing, feature depth processing and feature integration processing, and input into the model. The user's transaction probability is predicted through the main trunk of the model, and the intermediate extraction branch is further calculated to obtain the policy value of the user's transaction, so that the user type can be determined, and based on the user type, the target user matching the financial product to be analyzed is determined. Specifically, it includes the following contents:

[0151] 1. Data extraction module:

[0152] As the starting point of the system data flow, the data extraction module carries the core responsibility of obtaining key information from a complex and ever-changing data environment. This module has the ability to extract key customer feature data from a multi-source heterogeneous data environment.

[0153] The data extraction module includes multiple modules such as user basic information extraction module, salesperson information extraction module, historical policy information extraction module, and user claims information extraction module.

[0154] The extraction module can be used to obtain the customer's basic user information, corresponding salesperson information, historical insurance policy information and other characteristic data. For example, the basic user information includes the user's name, gender, date of birth, mobile phone number, marital status, customer star rating, whether the policyholder is insured, whether the insured is insured and other basic information; the salesperson information includes whether the salesperson is employed, historical order records, customer visit records, salesperson praise scores, etc.; historical insurance policy information includes policy number, policy type, policy name, number of insurance policy physical examinations, policy effective status, policy value, etc.; user claim information includes claim time, claim application amount, claim amount, claim policy number, etc.

[0155] 2. Feature processing module:

[0156] (1) Feature preprocessing module:

[0157] After data extraction, the feature preprocessing module performs preliminary cleaning and processing of the data, including removing noise data, filling missing values, discrete feature encoding, data interval scaling, etc., to generate high-quality feature data.

[0158] In this step, multiple related basic features are combined to form new features, and the feature data is amplified so that the subsequent model can have better results.

[0159] (2) Large model feature deep processing module:

[0160] In the model training phase, after feature preprocessing is completed, all feature names, feature meanings, and feature values ​​are input into the large model. The large model further processes the features according to the feature meanings and data feature values, removes features with low relevance to the prediction scenario, performs data processing on the features, and combines them to form new features.

[0161] During the model inference phase, the large model processes the data in the same way as in the training phase.

[0162] (3) Feature integration module:

[0163] Finally, the data processed in the above steps are integrated according to {user id (identity), feature 1, feature 2…feature n} to form a sample set.

[0164] During the training phase, all samples need to be divided into training set, test set, and validation set.

[0165] 3. Reasoning module:

[0166] Considering that the user's transaction probability and customer value are strongly correlated, for example, when the customer's transaction probability is 0, according to the definition, the customer value is the value of the insurance policy that the customer has transacted, and at this time the customer value must also be 0.

[0167] Therefore, we designed a neural network model, the main trunk predicts the user's transaction probability, and the middle extraction branch further performs model reasoning to obtain the policy value. This not only meets the characteristics of the prediction results being correlated, but also greatly reduces the amount of calculation compared to the general situation where two models are required to predict the probability and value respectively.

[0168] The model is as follows: the model trunk combines multiple network units (the network unit consists of a fully connected layer and an activation function) and outputs the user transaction probability through the Softmax layer. The model branch network is used to predict the customer value, extract the intermediate data of a certain level of the trunk model, calculate it through the network unit, and then accumulate the data of the next level of the trunk, calculate it through the network unit, and obtain the customer value after multiple calculations, such as Figure 2 shown.

[0169] The loss function of the model is as follows:

[0170] , formula (1)

[0171] Among them, i represents the i-th sample, n represents the total number of training samples, and p i represents the true value of the sample, p i ' is the probability value predicted by the model, δ is the loss adjustment coefficient, v i represents the real customer value of the sample, v i ' is the customer value predicted by the model. The loss function is used to determine when the model reaches convergence during training.

[0172] In the above-mentioned embodiment, in the process of matching financial products, by obtaining the object information of the object to be analyzed from multiple different sources, misjudgment caused by the one-sidedness of a single information source is avoided, so that the subsequent preprocessing and feature extraction processing of the object to be analyzed are established on a more comprehensive basis, and then based on the accurate target object type of the object to be analyzed, and based on the target object type, the target object that matches the financial product to be analyzed can be accurately determined from the object to be analyzed, which is conducive to improving the accuracy of financial product matching; moreover, the entire process does not require manual intervention, avoiding the defect of subjective factors and prone to errors in manual matching, resulting in a low accuracy rate of financial product matching, and further improving the accuracy of financial product matching. At the same time, this solution uses big data technology to extract massive user data from the ODS (Operational Data Store) data layer to make model predictions more accurate; this solution uses large model technology to help reduce the cost of feature selection and model training; this solution calculates the feature contribution of each sample in the AI ​​(Artificial Intelligence) model and gives the ranking results, thereby enhancing the interpretability of the black box model; this solution improves the accuracy of identifying high-potential customers who are easy to place orders, filters low-potential customers in advance, improves the experience and effectiveness of insurance customers, and reduces operating costs.

[0173] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0174] Based on the same inventive concept, the embodiment of the present application also provides a financial product matching device for implementing the financial product matching method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more financial product matching device embodiments provided below can refer to the limitations of the financial product matching method above, and will not be repeated here.

[0175] In an exemplary embodiment, Figure 5As shown, a financial product matching device is provided, including: an information acquisition module 501, an information processing module 502, a feature extraction module 503, an information determination module 504 and a product matching module 505, wherein:

[0176] The information acquisition module 501 is used to acquire object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed.

[0177] The information processing module 502 is used to preprocess the object information to obtain preprocessed object information.

[0178] The feature extraction module 503 is used to perform feature extraction processing on the preprocessed object information through a pre-trained object type recognition model to obtain the first information feature and the second information feature corresponding to the object to be analyzed, and perform multiple fusion processing on the first information feature and the second information feature to obtain the fused information feature corresponding to the object to be analyzed.

[0179] The information determination module 504 is used to determine the financial product acquisition probability information of the object to be analyzed according to the first information feature, and to determine the financial product resource transfer information of the object to be analyzed according to the fusion information feature.

[0180] The product matching module 505 is used to determine the target object type of the object to be analyzed according to the financial product acquisition probability information and the financial product resource transfer information, and determine the target object matching the financial product to be analyzed from the objects to be analyzed according to the target object type.

[0181] In an exemplary embodiment, the feature extraction module 503 is also used to fuse the first information feature and the second information feature to obtain a third information feature corresponding to the object to be analyzed; perform feature extraction on the first information feature to obtain a fourth information feature corresponding to the object to be analyzed, and perform feature extraction on the third information feature to obtain a fifth information feature corresponding to the object to be analyzed; use the fourth information feature as a new first information feature, use the fifth information feature as a new second information feature, and jump to the step of fusing the first information feature and the second information feature to obtain the third information feature corresponding to the object to be analyzed, until the obtained third information feature meets the preset condition; when the obtained third information feature meets the preset condition, use the obtained third information feature as the fused information feature.

[0182] In an exemplary embodiment, the financial product matching device also includes a model training module, which is used to obtain sample object information of the sample object; the sample object information includes at least basic attribute information and financial product information of the sample object; the sample object information is preprocessed to obtain preprocessed sample object information; through the object type recognition model to be trained, feature extraction processing is performed on the preprocessed sample object information respectively to obtain a first sample information feature and a second sample information feature corresponding to the sample object, and the first sample information feature and the second sample information feature are fused multiple times to obtain a sample fusion information feature corresponding to the sample object; according to the first sample information feature, the predicted financial product acquisition probability information of the sample object is determined, and according to the sample fusion information feature, the predicted financial product resource transfer information of the sample object is determined; the actual financial product acquisition probability information and the actual financial product resource transfer information of the sample object are obtained, and according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, the object type recognition model to be trained is iteratively trained to obtain a pre-trained object type recognition model.

[0183] In an exemplary embodiment, the model training module is also used to obtain a first loss value based on the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and to obtain a second loss value based on the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information; to fuse the first loss value and the second loss value to obtain a target loss value; and to iteratively train the object type recognition model to be trained based on the target loss value to obtain a pre-trained object type recognition model.

[0184] In an exemplary embodiment, the information processing module 502 is also used to obtain application scenario information corresponding to the object to be analyzed; determine the correlation between each object information and the application scenario information; identify, from each object information, object information whose correlation with the application scenario information is less than a preset correlation, as object information to be processed; delete the object information to be processed in each object information to obtain pre-processed object information.

[0185] In an exemplary embodiment, the product matching module 505 is also used to obtain probability information and financial product resource transfer information based on financial products, to obtain multiple prediction object types of the object to be analyzed, and the prediction probability of each prediction object type; from each prediction object type, the prediction object type with the largest prediction probability is screened out as the target object type.

[0186] Each module in the above-mentioned financial product matching device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0187] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as basic attribute information and financial product information. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a financial product matching method is implemented.

[0188] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0189] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0190] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0191] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0192] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0193] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0194] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A financial product matching method, characterized in that: The method comprises: Obtaining object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed; Preprocessing the object information to obtain preprocessed object information; Through the pre-trained object type recognition model, feature extraction processing is performed on the pre-processed object information to obtain first information features and second information features corresponding to the object to be analyzed, and the first information features and the second information features are fused multiple times to obtain fused information features corresponding to the object to be analyzed; Determine the financial product acquisition probability information of the object to be analyzed based on the first information feature, and determine the financial product resource transfer information of the object to be analyzed based on the fusion information feature; The target object type of the object to be analyzed is determined according to the financial product acquisition probability information and the financial product resource transfer information, and the target object matching the financial product to be analyzed is determined from the objects to be analyzed according to the target object type.

2. The method according to claim 1, characterized in that: The performing multiple fusion processing on the first information feature and the second information feature to obtain the fused information feature corresponding to the object to be analyzed includes: The first information feature and the second information feature are fused to obtain a third information feature corresponding to the object to be analyzed; Performing feature extraction processing on the first information feature to obtain a fourth information feature corresponding to the object to be analyzed, and performing feature extraction processing on the third information feature to obtain a fifth information feature corresponding to the object to be analyzed; The fourth information feature is used as a new first information feature, the fifth information feature is used as a new second information feature, and the process proceeds to the step of fusing the first information feature and the second information feature to obtain a third information feature corresponding to the object to be analyzed, until the obtained third information feature meets a preset condition; When the obtained third information feature meets the preset condition, the obtained third information feature is used as the fused information feature.

3. The method according to claim 1, characterized in that The pre-trained object type recognition model is trained in the following way: Acquire sample object information of the sample object; the sample object information includes at least basic attribute information and financial product information of the sample object; Preprocessing the sample object information to obtain preprocessed sample object information; By using the object type recognition model to be trained, feature extraction processing is performed on the preprocessed sample object information respectively to obtain a first sample information feature and a second sample information feature corresponding to the sample object, and the first sample information feature and the second sample information feature are fused multiple times to obtain a sample fusion information feature corresponding to the sample object; Determine the predicted financial product acquisition probability information of the sample object according to the first sample information feature, and determine the predicted financial product resource transfer information of the sample object according to the sample fusion information feature; The actual financial product acquisition probability information and the actual financial product resource transfer information of the sample object are obtained, and according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, the object type recognition model to be trained is iteratively trained to obtain the pre-trained object type recognition model.

4. The method according to claim 3, characterized in that The iterative training of the object type recognition model to be trained according to the difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and according to the difference between the predicted financial product resource transfer information and the actual financial product resource transfer information, to obtain the pre-trained object type recognition model, comprises: Obtaining a first loss value according to a difference between the predicted financial product acquisition probability information and the actual financial product acquisition probability information, and obtaining a second loss value according to a difference between the predicted financial product resource transfer information and the actual financial product resource transfer information; fusing the first loss value and the second loss value to obtain a target loss value; According to the target loss value, the object type recognition model to be trained is iteratively trained to obtain the pre-trained object type recognition model.

5. The method according to claim 1, characterized in that: The preprocessing of the object information to obtain preprocessed object information includes: Obtaining application scenario information corresponding to the object to be analyzed; Determining the correlation between each of the object information and the application scenario information; From each of the object information, identifying object information whose correlation with the application scenario information is less than a preset correlation as the object information to be processed; The object information to be processed in each of the object information is deleted to obtain the pre-processed object information.

6. The method according to any one of claims 1 to 5, characterized in that The step of determining the target object type of the object to be analyzed according to the financial product acquisition probability information and the financial product resource transfer information includes: Obtaining multiple prediction object types of the object to be analyzed and the prediction probability of each prediction object type according to the financial product acquisition probability information and the financial product resource transfer information; From each of the predicted object types, the predicted object type with the largest predicted probability is screened out as the target object type.

7. A financial product matching device, characterized in that: The device comprises: An information acquisition module, used to acquire object information of the object to be analyzed; the object information at least includes basic attribute information and financial product information of the object to be analyzed; An information processing module, used for preprocessing the object information to obtain preprocessed object information; A feature extraction module is used to perform feature extraction processing on the preprocessed object information respectively through a pre-trained object type recognition model to obtain a first information feature and a second information feature corresponding to the object to be analyzed, and perform multiple fusion processing on the first information feature and the second information feature to obtain a fused information feature corresponding to the object to be analyzed; an information determination module, configured to determine the financial product acquisition probability information of the object to be analyzed according to the first information feature, and to determine the financial product resource transfer information of the object to be analyzed according to the fusion information feature; The product matching module is used to determine the target object type of the object to be analyzed according to the financial product acquisition probability information and the financial product resource transfer information, and determine the target object matching the financial product to be analyzed from the objects to be analyzed according to the target object type.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images