Cargo transportation insurance pricing and insurance method and system based on logistics whole-link data

Abstract

The application discloses a kind of based on logistics whole link data's freight transport insurance pricing and method and system, the present application is based on logistics order data, to build out insurance feature vector;Then, historical logistics order data is clustered, obtains several order class clusters;Then, the preceding insurance feature vector is searched feature, to match out with the most similar order class cluster of logistics order data, and in the most similar order class cluster, filter out similar order set corresponding to logistics order data;Then, the transport insurance information of each similar order in the similar order set can be based on, to generate the best transport insurance scheme containing freight transport insurance pricing information and insurance scheme information;Finally, it is sent to user end for user reference, and after user confirms insurance, then the automatic insurance of target goods can be completed;Therefore, compared with traditional technology, the present application improves efficiency, reduces labor cost, and improves user experience.

Description

Technical Field

[0001] This invention belongs to the field of insurance recommendation technology, specifically relating to a pricing and underwriting method and system for cargo transportation insurance based on full-chain logistics data. Background Technology

[0002] During transportation, goods inevitably encounter various risks (such as natural disaster risks, accident risks, loading and unloading accident risks, etc.), which are often difficult to completely avoid or control. Therefore, during the transportation of goods, transportation insurance is usually purchased to effectively transfer and disperse various risks during transportation, thereby reducing losses for enterprises and individuals.

[0003] Currently, most cargo transportation insurance schemes are recommended manually to achieve pricing and insurance coverage. This involves communication between insurance agents and policyholders or insured parties, followed by recommendations based on the communication results and relevant risk considerations. However, the aforementioned existing technologies have the following shortcomings: (1) Manual recommendations require analyzing orders and querying historical cases one by one, which is time-consuming and labor-intensive, and cannot cope with massive orders and complex and ever-changing logistics scenarios; (2) Reliance on professional insurance consultants or sales teams results in high labor costs, making it difficult to support large-scale, high-concurrency business needs; (3) Customers often have to wait a long time to obtain a recommended scheme and apply for insurance, which leads to a poor user experience. Therefore, based on the aforementioned shortcomings, how to provide a highly efficient, low-cost, and user-friendly cargo transportation insurance pricing and insurance method based on full-chain logistics data has become an urgent problem to be solved. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for pricing and underwriting cargo transportation insurance based on full-chain logistics data, in order to solve the problems of low efficiency, high cost and poor user experience in the existing technology.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a method for pricing and underwriting cargo transportation insurance based on end-to-end logistics data is provided, including: Obtain logistics order data for the target goods and a historical cargo transportation insurance database. The logistics order data includes cargo information, transportation information, and weather information, and the historical cargo transportation insurance database contains several historical logistics order data and corresponding transportation insurance information for each historical logistics order data. Based on cargo information, transportation information, and weather information in logistics order data, an insurance feature vector for the target cargo is constructed. Clustering is performed on historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters; Based on the insurance feature vector, the order cluster most similar to the logistics order data is selected from several order clusters and used as the target cluster; Using the insurance feature vector, a set of similar orders corresponding to the logistics order data is determined from the target cluster; Based on the transportation insurance information of each similar order in the similar order set, the optimal transportation insurance plan for the target goods is generated and sent to the user terminal. The optimal transportation insurance plan includes cargo transportation insurance pricing information and insurance plan information. In response to the user's interactive operation for insurance application, the optimal transportation insurance plan is sent to the insurance client to complete the transportation insurance application for the target goods.

[0006] Based on the aforementioned disclosure, this invention first acquires logistics order data for target goods, including cargo information, transportation information, and weather information. Then, based on the cargo information, transportation information, and weather information in the logistics order data, an insurance feature vector for the target goods is constructed. Subsequently, based on the insurance feature vector and combined with historical logistics order data and corresponding transportation insurance information in the historical cargo transportation insurance database, a transportation insurance plan can be recommended, thereby achieving transportation insurance pricing and automatic insurance coverage. Specifically, this invention first performs clustering processing on the historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters. Then, based on the preceding... The insurance feature vector is used to filter out the order clusters most similar to the logistics order data from several order clusters, thus serving as the target clusters. Then, the insurance feature vector is used to filter out the set of similar orders corresponding to the aforementioned logistics order data from the target clusters. Subsequently, based on the transportation insurance information of each similar order in the set of similar orders, the optimal transportation insurance plan, which includes cargo transportation insurance pricing information and insurance plan information, can be generated. Finally, this plan can be sent to the user's terminal for reference, and in response to the user's human-computer interaction operation for insurance, the optimal transportation insurance plan can be sent to the insurance client, thereby automatically completing the transportation insurance application for the target goods.

[0007] Through the above design, compared with traditional technologies, the present invention improves efficiency, reduces labor costs, reduces user waiting time, and improves user experience; therefore, it is very suitable for large-scale application and promotion.

[0008] In one possible design, based on cargo information, transportation information, and weather information from the logistics order data, an insurance feature vector for the target cargo is constructed, including: Based on the cargo information, the cargo type, unit price, and attributes are determined, and cargo characteristics are generated using the cargo type, unit price, and attributes. Based on the transportation information, the transportation mode, transportation distance, transportation time and carrier information are determined; Using the aforementioned transportation mode, transportation distance, transportation duration, and carrier information, transportation characteristics are generated; Based on the weather information, generate the weather characteristics of the target cargo during transportation; The insurance feature vector is generated using the cargo features, the transportation features, and the weather features.

[0009] In one possible design, historical logistics order data in the historical cargo transportation insurance database is clustered to obtain several order clusters, including: For any historical logistics order data in the historical cargo transportation insurance database, obtain the k nearest neighbor historical logistics order data of any historical logistics order data; Calculate the similarity between any historical logistics order data and each of its nearest neighboring historical logistics order data, and calculate the cluster density of any historical logistics order data based on each similarity. Based on the k nearest neighbor historical logistics order data, the cluster distance of any historical logistics order data is calculated; Based on the cluster density and the cluster distance, the cluster decision value of any historical logistics order data is determined, and after all historical logistics order data has been queried, the cluster decision value of each historical logistics order data is obtained. Based on the clustering decision values ​​of various historical logistics order data, several cluster centers are determined; Using several cluster centers, all historical logistics order data are clustered to obtain several order clusters.

[0010] In one possible design, the similarity between any given historical logistics order data and each of its nearest neighboring historical logistics order data is calculated, including: For any neighboring historical logistics order data, similar historical order data corresponding to the neighboring historical logistics order data are determined from the historical cargo transportation insurance database, and a first set is formed using the neighboring historical logistics order data and the similar historical order data; Find the intersection between the first set and the second set to obtain the order intersection, wherein the second set contains any historical logistics order data and the k nearest neighbor historical logistics order data corresponding to any historical logistics order data; Based on the intersection of the orders, the similarity between any historical logistics order data and any nearest neighbor historical logistics order data is calculated.

[0011] In one possible design, based on the order intersection, the similarity between any historical logistics order data and any nearest neighbor historical logistics order data is calculated, including: Determine whether any of the historical logistics order data and any of the nearest neighbor historical logistics order data exist in the order intersection; If not, then determine whether any of the historical logistics order data or any of the nearest historical logistics order data exists in the intersection of the orders; If so, then obtain the first historical insurance feature vector corresponding to any historical logistics order data, and the second historical insurance feature vector corresponding to any nearest neighbor historical logistics order data; Based on the first historical insurance feature vector and the second historical insurance feature vector, the vector distance between any historical logistics order data and any nearest historical logistics order data is calculated; Based on the total number of data points in the order intersection and the vector distance, the similarity between any historical logistics order data and any nearest neighbor historical logistics order data is calculated.

[0012] In one possible design, the clustering distance of any historical logistics order data is calculated based on k nearest neighbor historical logistics order data, including: Obtain similar historical order data from each neighboring historical logistics order database; Calculate the first distance between any historical logistics order data and each of its nearest historical logistics order data, and sum the obtained first distances to obtain the total first distance; For the a-th nearest historical logistics order data, calculate the second distance between the a-th nearest historical logistics order data and each similar historical order data corresponding to the a-th nearest historical logistics order data; Sum the individual second distances to obtain the total second distance; Summing the first and second total distances yields the third total distance. The initial clustering distance is obtained by multiplying the first distance between any historical logistics order data and the a-th nearest neighbor historical logistics order data by the sum of the third distances. Increment 'a' by 1 and recalculate the second distance between the a-th nearest historical logistics order data and each similar historical order data corresponding to the a-th nearest historical logistics order data, until a equals k, to obtain multiple initial clustering distances, where the initial value of a is 1; From multiple initial cluster distances, the smallest initial cluster distance is selected as the cluster distance for any historical logistics order data.

[0013] In one possible design, based on the clustering decision values ​​of various historical logistics order data, several cluster centers are determined, including: The historical logistics order data are sorted in descending order of clustering decision values ​​to obtain a sorted sequence; Based on the total number of historical logistics order data, a data truncation threshold is determined. Based on the data truncation threshold, several historical logistics order data are extracted from the sorted sequence to form a specified set; The mutation threshold is calculated based on the clustering decision value of each historical logistics order data in the specified set; For the q-th historical logistics order data in the specified set, calculate the difference between the clustering decision value of the q-th historical logistics order data and the clustering decision value of the (q-1)-th historical logistics order data, and use it as the first mutation intermediate value; Determine whether the absolute value of the first mutation intermediate value is greater than or equal to the mutation threshold; If so, record q in a one-dimensional array, increment q by 1, and recalculate the difference between the clustering decision value of the q-th historical logistics order data and the clustering decision value of the (q-1)-th historical logistics order data until q equals n, thus obtaining the final one-dimensional array. The initial value of q is 2, and n is the total number of data in the specified set. From the final one-dimensional array, select the element with the largest value, and use the historical logistics order data corresponding to the element with the largest value as the cluster mutation point; Historical logistics order data that occur before the cluster mutation point in the sorted sequence are used as cluster centers.

[0014] In one possible design, the cluster center of any order cluster corresponds to a historical insurance feature vector. Based on this insurance feature vector, the order clusters most similar to the logistics order data are selected from several order clusters as target clusters, including: Calculate the distance between the insurance feature vector and the historical insurance feature vector of the distance center of each order cluster, and sort the order clusters in ascending order of distance to obtain the cluster sequence; The order clusters that rank u first in the sorted cluster sequence are taken as the target clusters, where u is a positive integer.

[0015] Secondly, a cargo transportation insurance pricing and underwriting system based on end-to-end logistics data is provided, including: The acquisition unit is used to acquire logistics order data of the target goods and historical cargo transportation insurance database. The logistics order data includes cargo information, transportation information and weather information, and the historical cargo transportation insurance database contains several historical logistics order data and transportation insurance information corresponding to each historical logistics order data. The feature construction unit is used to construct the insurance feature vector of the target goods based on the cargo information, transportation information and weather information in the logistics order data; Clustering units are used to cluster historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters. The search unit is used to select the order cluster that is most similar to the logistics order data from several order clusters based on the insurance feature vector, and use it as the target cluster. The search unit is also used to use the insurance feature vector to determine the set of similar orders corresponding to the logistics order data from the target cluster; The insurance planning unit is used to generate the best transportation insurance plan for the target goods based on the transportation insurance information of each similar order in the similar order set, and send it to the user terminal. The best transportation insurance plan includes cargo transportation insurance pricing information and insurance plan information. The insurance unit is used to respond to the user's human-computer interaction operation for insurance application, and to send the optimal transportation insurance plan to the insurance client to complete the transportation insurance application for the target goods.

[0016] Thirdly, a cargo transportation insurance pricing and underwriting device based on logistics end-to-end data is provided. Taking the device as an electronic device as an example, it includes a memory, a processor, and a transceiver that are connected in sequence. The memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer programs and execute the cargo transportation insurance pricing and underwriting method based on logistics end-to-end data as described in the first aspect or any possible design in the first aspect.

[0017] Fourthly, a storage medium is provided, on which instructions are stored, which, when executed on a computer, perform the cargo transportation insurance pricing and underwriting method based on the entire logistics chain data as described in the first aspect or any possible design in the first aspect.

[0018] Fifthly, a computer program product containing instructions is provided, which, when executed on a computer, causes the computer to perform the cargo transportation insurance pricing and underwriting method based on end-to-end logistics data as described in the first aspect or any possible design of the first aspect.

[0019] Beneficial effects: (1) Based on logistics order data, this invention constructs an insurance feature vector containing cargo information, transportation information, and weather information. Then, it performs clustering processing on historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters. Next, using the aforementioned insurance feature vector as search features, it matches the order clusters most similar to the logistics order data, and filters out the similar order set corresponding to the logistics order data from the most similar order clusters. Then, based on the transportation insurance information of each similar order in the similar order set, it generates the best transportation insurance plan containing cargo transportation insurance pricing information and insurance plan information. Finally, it sends the plan to the user for reference, and after the user confirms the insurance, the automatic insurance of the target cargo can be completed. Thus, compared with traditional technology, this invention improves efficiency, reduces labor costs, reduces user waiting time, and improves user experience. Therefore, it is very suitable for large-scale application and promotion. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the steps of the cargo transportation insurance pricing and underwriting method based on end-to-end logistics data provided in this embodiment of the invention. Figure 2 This is a structural diagram of a cargo transportation insurance pricing and underwriting system based on end-to-end logistics data, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0022] It should be understood that although the terms first, second, etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit, without departing from the scope of the exemplary embodiments of the invention.

[0023] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, and A and B exist simultaneously. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone, and A and B exist alone. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0024] Example: See Figure 1 As shown, the cargo transportation insurance pricing and underwriting method based on end-to-end logistics data provided in this embodiment constructs an insurance feature vector based on logistics order data. Then, it performs clustering processing on historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters. Next, using the aforementioned insurance feature vector as a search feature, it matches the order cluster most similar to the logistics order data, and filters out a set of similar orders corresponding to the logistics order data from the most similar order clusters. Then, based on the transportation insurance information of each similar order in the similar order set, it generates the optimal transportation insurance plan containing cargo transportation insurance pricing information and insurance plan information. Finally, it sends the plan to the user for reference, and after the user confirms the insurance, the automatic insurance of the target cargo is completed. Therefore, compared with traditional technologies, this method improves efficiency, reduces labor costs, reduces user waiting time, and improves user experience. Thus, it is very suitable for large-scale application and promotion. For example, this method can run on, but is not limited to, the insurance server side. It is understood that the aforementioned execution entity does not constitute a limitation on the embodiments of this application. Accordingly, the operation steps of this method can be, but are not limited to, as shown in steps S1 to S7 below.

[0025] S1. Obtain logistics order data for the target goods and a historical cargo transportation insurance database, wherein the logistics order data includes cargo information, transportation information, and weather information, and the historical cargo transportation insurance database contains several historical logistics order data and corresponding transportation insurance information for each historical logistics order data.

[0026] In this embodiment, the cargo information may include, but is not limited to, the category, total value, weight, and storage conditions of the target cargo; the transportation information may include, but is not limited to, the origin and destination, route, distance, estimated duration, and carrier information (company, vehicle, driver); and the weather information may include, but is not limited to, weather forecast information during the transportation period. Similarly, the content contained in the historical logistics order data is also the same, and will not be repeated here. Furthermore, the transportation insurance information may include, but is not limited to, the type of insurance, the insured amount, the premium, and the deductible.

[0027] Thus, after obtaining the aforementioned information, feature construction can be performed so that the constructed features can be used to recommend transportation insurance schemes; the feature construction process is as shown in step S2 below.

[0028] S2. Based on the cargo information, transportation information, and weather information in the logistics order data, construct the insurance feature vector of the target cargo; in specific applications, for example, but not limited to, the following steps S21 to S25 can be used to construct the insurance feature vector.

[0029] S21. Based on the cargo information, determine the cargo type, unit price, and attributes, and generate cargo characteristics using the cargo type, unit price, and attributes. In specific implementation, the cargo type is the category of goods, such as electronic products, fresh produce, or general building materials; the unit price is calculated by dividing the total value of the goods by the weight of the goods; and the cargo attributes are obtained based on storage conditions, such as whether temperature control is required in this embodiment. Thus, after determining the aforementioned information, cargo characteristics can be constructed based on it.

[0030] In specific implementation, different numbers are used to represent different types of goods, such as electronic products using the number 11, fresh produce using the number 12, and ordinary building materials using the number 13, etc. At the same time, if temperature control is required, the goods attribute is 1, and otherwise it is 0. In this way, after constructing the goods characteristics, the transportation characteristics can be constructed, and the process is shown in step S22 below.

[0031] S22. Based on the transportation information, determine the transportation mode, transportation distance, transportation duration, and carrier information; in this embodiment, carrier information may include, but is not limited to, the carrier company's credit rating and the driver's historical risk score (which is obtained based on the driver's historical driving behavior and pre-stored in the insurance server); therefore, after obtaining the aforementioned transportation-related information, transportation characteristics can be constructed based on this, as shown in step S23 below.

[0032] S23. Using the aforementioned mode of transport, transport distance, transport duration, and carrier information, transport characteristics are generated; in specific implementation, the aforementioned mode of transport, transport distance, transport duration, carrier credit rating, and driver historical risk score are used to form transport characteristics; then, weather characteristics can be constructed, the process of which is shown in step S24 below.

[0033] S24. Based on the weather information, generate the weather characteristics of the target cargo during transportation; in this embodiment, the weather forecast risk level during transportation can be determined based on the weather forecast information in the weather information, and then used as the weather characteristic; of course, different weather forecasts correspond to different levels, which can be obtained by looking up the weather forecast table.

[0034] Thus, after obtaining the weather characteristics, the aforementioned cargo characteristics and transportation characteristics can be combined to generate an insurance feature vector, as shown in step S25 below.

[0035] S25. Using the cargo features, transportation features, and weather features, generate the insurance feature vector; in specific applications, for example, but not limited to, first perform feature vectorization on the cargo features, transportation features, and weather features, that is, convert all the above features into a numerical vector for mathematical calculation; then, concatenate the feature vectorized cargo features, transportation features, and weather features to obtain the insurance feature vector; wherein, for example, but not limited to, one-hot encoding can be used to achieve the vectorization of the aforementioned features.

[0036] After constructing the insurance feature vector of the target goods through the aforementioned steps S21 to S25, the historical logistics order data and corresponding transportation insurance information in the historical cargo transportation insurance database can be combined to recommend a transportation insurance plan for the target goods, that is, to realize the pricing and insurance of the target goods' transportation insurance.

[0037] In this embodiment, historical logistics order data is first clustered to obtain several order clusters. Then, the insurance feature vector of the target goods is used as a search condition to perform cluster matching in the order clusters, thereby obtaining the order clusters most similar to the logistics order data of the target goods. Then, similar order data of the logistics order data of the target goods are determined in the most similar order clusters. Finally, transportation insurance pricing and insurance are performed based on the similar order data. Optionally, the clustering process of historical logistics order data can be, but is not limited to, the following step S3.

[0038] S3. Cluster the historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters. In specific applications, this embodiment determines the cluster center by calculating the cluster decision value of each historical logistics order data. Then, based on the cluster center, the historical logistics order data is clustered. The process can be, but is not limited to, the steps S31 to S36 below.

[0039] S31. For any historical logistics order data in the historical cargo transportation insurance database, obtain the k nearest neighbor historical logistics order data of the any historical logistics order data. In this embodiment, the historical insurance feature vector of each historical logistics order data is first constructed. Then, based on the historical insurance feature vector of any historical logistics order data, its corresponding k nearest neighbor historical logistics order data are determined. Specifically, the k nearest neighbor historical logistics order data is determined based on the historical insurance feature vector of any historical logistics order data and using the k nearest neighbor algorithm. Further, the k nearest neighbor algorithm refers to the process of finding the K points closest to the current point in a known category dataset. It is a commonly used data search technique, and its principle will not be elaborated here.

[0040] After obtaining the k nearest neighbor historical logistics order data of any historical logistics order data, the similarity can be calculated so that the cluster density of any historical logistics order data can be determined based on the calculated similarity. The calculation process of similarity and cluster density is as shown in step S32 below.

[0041] S32. Calculate the similarity between any historical logistics order data and each of the nearest historical logistics order data, and calculate the cluster density of any historical logistics order data based on each similarity. In this embodiment, a similarity calculation method based on k-nearest neighbor intersection is proposed, and the process can be, but is not limited to, the steps S32a to S32c below.

[0042] S32a. For any nearest neighbor historical logistics order data, determine the similar historical order data corresponding to the nearest neighbor historical logistics order data from the historical cargo transportation insurance database, and use the nearest neighbor historical logistics order data and the similar historical order data to form a first set; in this embodiment, the aforementioned similar historical order data is: the k nearest neighbor order data of the nearest neighbor historical logistics order data; thus, it is equivalent to using any nearest neighbor historical logistics order data and its corresponding k nearest neighbor order data to form the first set.

[0043] After obtaining the first set, the intersection of the k nearest neighbors can be determined, as shown in step S32b below.

[0044] S32b. Find the intersection between the first set and the second set to obtain the order intersection, wherein the second set contains any historical logistics order data and the k nearest neighbor historical logistics order data corresponding to the historical logistics order data; in this embodiment, after obtaining the k nearest neighbor intersection (i.e., the order intersection) between the first set and the second set, the similarity between the historical logistics order data and the corresponding nearest neighbor historical logistics order data can be calculated based on the k nearest neighbor intersection, and the calculation process is shown in step S32c below.

[0045] S32c. Based on the order intersection, calculate the similarity between any historical logistics order data and any neighboring historical logistics order data. In this embodiment, different similarity calculation methods are used depending on whether the order intersection includes any historical logistics order data and any neighboring historical logistics order data. The process can be, but is not limited to, the steps S32c1 to S32c5 below.

[0046] S32c1. Determine whether any of the historical logistics order data and any of the nearest historical logistics order data exist in the order intersection; in this embodiment, if the aforementioned conditions are met, it is necessary to calculate the similarity based on the distance between each order in the order intersection and any of the historical logistics order data and any of the nearest historical logistics order data. The process is as follows: Step 1: Calculate the fourth distance between any historical logistics order data and each order in the order intersection, and the fifth distance between any nearest historical logistics order data and each order in the order intersection; in this embodiment, the fourth distance is calculated as the cosine distance between the historical insurance feature vector of any historical logistics order data and the historical insurance feature vector of each order in the order disclosure; similarly, the fifth distance is calculated in the same way, and will not be repeated here.

[0047] Step 2: Summate the fourth distance and the fifth distance to obtain the total distance.

[0048] Step 3: Based on the total number of data in the intersection of the orders and the sum of the distances, calculate the similarity between any historical logistics order data and any nearest neighbor historical logistics order data.

[0049] In practical applications, the following formula can be used, for example but not limited to, to calculate the similarity; ; In the formula, This indicates the similarity between any historical logistics order data and any nearest neighboring historical logistics order data. Indicates the total number of data. This indicates that the first historical logistics order data in the intersection with the order data represents the... The fourth distance between orders This represents the first nearest neighbor historical logistics order data and the order intersection. The fifth distance between orders This represents the intersection of the aforementioned orders; thus, the aforementioned and Summing these values ​​gives the distance sum.

[0050] Based on the aforementioned formula, the similarity between two historical logistics order data can be calculated when the order intersection contains both the historical logistics order data of the aforementioned historical logistics order data of the aforementioned historical logistics order data.

[0051] If the aforementioned conditions are not met, further judgment is required, namely, the following step S32c2.

[0052] S32c2. If not, determine whether any of the historical logistics order data or any of the nearest historical logistics order data exists in the intersection of the orders; in this embodiment, if the condition in step S32c2 is met, it is necessary to use the historical insurance feature vectors of the two to calculate the vector distance between them, and the process is shown in steps S32c3 and S32c4 below.

[0053] S32c3. If so, obtain the first historical insurance feature vector corresponding to any historical logistics order data and the second historical insurance feature vector corresponding to any neighboring historical logistics order data; in this embodiment, the construction process of the historical insurance feature vector can be referred to the aforementioned step S2, and will not be repeated here.

[0054] After obtaining the historical insurance feature vectors of each group of any historical logistics order data and any neighboring historical logistics order data, the distance between the two can be calculated based on these vectors, as shown in step S32c4 below.

[0055] S32c4. Based on the first historical insurance feature vector and the second historical insurance feature vector, calculate the vector distance between any historical logistics order data and any neighboring historical logistics order data; in this embodiment, the cosine distance between the first and second historical insurance feature vectors is used as the vector distance; then, the similarity between the two can be calculated by combining the total number of order intersections, as shown in step S32c5 below.

[0056] S32c5. Based on the total number of data in the intersection of the orders and the vector distance, calculate the similarity between any historical logistics order data and any nearest neighbor historical logistics order data.

[0057] In practical applications, examples such as, but not limited to, using the following formula can be used to calculate the similarity between the two.

[0058] ; In the formula, This indicates the similarity between any historical logistics order data and any nearest neighboring historical logistics order data. This represents the maximum nearest neighbor number (a preset value, which is essentially k). Indicates the total number of data. Represents vector distance.

[0059] In addition, in this embodiment, when the order intersection does not contain either of the historical logistics order data or either of the nearest neighbor historical logistics order data, the similarity between the two is 0.

[0060] Thus, through the aforementioned steps S32c1 to S32c5, the similarity between any historical logistics order data and any nearest neighbor historical logistics order data can be calculated; then, in the same manner as described above, the similarity between any historical logistics order data and each nearest neighbor historical logistics order data can be calculated; then, based on the calculated similarities, the cluster density of any historical logistics order data can be determined.

[0061] Specifically, in this embodiment, the sum of each similarity is first calculated to obtain the total similarity; then, the total number of neighboring historical logistics order data corresponding to any historical logistics order data is multiplied by the total similarity to obtain the cluster density.

[0062] After calculating the cluster density of any historical logistics order data based on the aforementioned step S32, the cluster distance can be calculated, as shown in step S33 below.

[0063] S33. Calculate the clustering distance of any historical logistics order data based on the k nearest neighbor historical logistics order data; in this embodiment, for example, but not limited to, the following steps S33a to S33h can be used to calculate the clustering distance.

[0064] S33a. Obtain similar historical order data for each neighboring historical logistics order data from the historical cargo transportation insurance database; in this embodiment, the similar order data for any neighboring historical logistics order data is the k-nearest neighbor order of that neighboring historical logistics order data, which has been described above and will not be repeated here.

[0065] After obtaining the similar historical order data of each neighboring historical logistics order data, the distance between orders can be calculated, as shown in step S33b below.

[0066] S33b. Calculate the first distance between any historical logistics order data and each of the nearest historical logistics order data, and sum the obtained first distances to obtain the first distance sum; in this embodiment, the cosine distance between the historical insurance feature vector of any historical logistics order data and the historical insurance feature vector of each of the nearest historical logistics order data is still used as each first distance; in this way, after summing each first distance to obtain the first distance sum, the distance between each of the nearest historical logistics order data and their respective similar order data can be calculated, and the process is as shown in step S33c below.

[0067] S33c. For the a-th nearest neighbor historical logistics order data, calculate the second distance between the a-th nearest neighbor historical logistics order data and each similar historical order data corresponding to the a-th nearest neighbor historical logistics order data; in this embodiment, the calculation process of the second distance is the same as that of the first distance, and will not be repeated here.

[0068] After obtaining the second distance between the a-th nearest historical logistics order data and its corresponding similar order data, the same summation is performed to obtain the second distance summation, as shown in step S33d below.

[0069] S33d. Summate the obtained second distances to obtain the total second distance.

[0070] After obtaining the second total distance, it can be added to the second total distance to obtain the first total distance, as shown in step S33e below.

[0071] S33e. Summing the first and second distance sums yields the third distance sum.

[0072] After obtaining the third sum of distances, the initial clustering distance can be calculated by combining the first distance between any historical logistics order data and the a-th nearest historical logistics order data, as shown in step S33f below.

[0073] S33f. Using the first distance between any historical logistics order data and the a-th nearest historical logistics order data, multiply by the sum of the third distances to obtain the initial clustering distance.

[0074] After the calculation of the a-th nearest neighbor historical logistics order data is completed, the calculation of the next nearest neighbor historical logistics order data can be performed in the same way as described above, until all k nearest neighbor historical logistics order data of any historical logistics order data have been polled. At this time, k initial clustering distances can be obtained; wherein, the polling process is as shown in step S33g below.

[0075] S33g. Increment 'a' by 1 and recalculate the second distance between the a-th nearest historical logistics order data and each similar historical order data corresponding to the a-th nearest historical logistics order data, until a equals k, to obtain multiple initial clustering distances, where the initial value of a is 1.

[0076] After obtaining k initial cluster distances, the smallest initial cluster distance can be selected as the cluster distance for any historical logistics order data. The process is shown in step S33h below.

[0077] S33h. Select the smallest initial clustering distance from multiple initial clustering distances to use as the clustering distance for any of the historical logistics order data.

[0078] Thus, through the aforementioned steps S33a to S33h, the cluster distance of any historical logistics order data can be calculated; then, the cluster decision value can be calculated by combining its corresponding cluster density, so that the cluster center can be selected based on the cluster decision value; wherein, the calculation process of the cluster decision value is shown in step S34 below.

[0079] S34. Based on the cluster density and the cluster distance, determine the cluster decision value of any historical logistics order data, and obtain the cluster decision value of each historical logistics order data after polling all historical logistics order data; in this embodiment, the cluster decision value of any historical logistics order data can be obtained by multiplying the cluster density by the cluster distance; thus, after polling all historical logistics order data in the database in the aforementioned manner, the cluster decision value of each historical logistics order data can be obtained.

[0080] Then, cluster centers can be selected based on the clustering decision value, as shown in step S35 below.

[0081] S35. Based on the clustering decision values ​​of each historical logistics order data, several cluster centers are determined. In specific implementation, this embodiment determines the cluster mutation points, and then selects the cluster centers based on the cluster mutation points. The process can be, but is not limited to, the steps S35a to S35i below.

[0082] S35a. Sort the historical logistics order data in descending order of clustering decision values ​​to obtain a sorted sequence. In this embodiment, after sorting the historical logistics order data, it is necessary to determine the data truncation threshold so that the data can be extracted from the sorted sequence based on the data truncation threshold.

[0083] The calculation process for the data truncation threshold is shown in step S35b below.

[0084] S35b. Based on the total number of historical logistics order data, determine the data truncation threshold; in this embodiment, the calculation process of the data truncation threshold is as follows: first, perform a square root operation on the total number of historical logistics order data, and take the positive result of the square root operation as the initial data truncation threshold; then, round up the initial data truncation threshold to obtain the data truncation threshold.

[0085] After obtaining the data truncation threshold, data extraction can be performed, as shown in step S35c below.

[0086] S35c. Based on the data truncation threshold, extract several historical logistics order data from the sorted sequence to form a specified set; in this embodiment, assuming the data truncation threshold is n, then the first n historical logistics order data from the sorted sequence are taken to form the specified set; then, the mutation threshold can be calculated based on the specified set, as shown in step S35d below.

[0087] S35d. Based on the clustering decision values ​​of each historical logistics order data in the specified set, a mutation threshold is calculated. In this embodiment, the difference between the clustering decision values ​​of each adjacent pair of historical logistics order data in the specified set is first calculated, and the sum of the calculated differences is obtained to get the second mutation intermediate value. Then, the mutation threshold is calculated based on the second mutation intermediate value. Specifically, the difference between the clustering decision value of the q-th historical logistics order data in the specified set and the clustering decision value of the (q-1)-th historical logistics order data is calculated. Then, q is incremented by 1 until q equals n, resulting in multiple differences. Finally, the sum of the absolute values ​​of each difference is obtained to get the second mutation intermediate value, where the initial value of q is 2.

[0088] Specifically, the mutation threshold is calculated by dividing the second mutation intermediate value by the target value, where the target value is n-2.

[0089] Thus, after calculating the mutation threshold, the mutation point can be selected based on it, as shown in steps S35e to S35i below.

[0090] S35e. For the q-th historical logistics order data in the specified set, calculate the difference between the clustering decision value of the q-th historical logistics order data and the clustering decision value of the (q-1)-th historical logistics order data, and use it as the first mutation intermediate value; after obtaining the first mutation intermediate value, it can be compared with the mutation threshold, and the process is as shown in step S35f below.

[0091] S35f. Determine whether the absolute value of the first mutation intermediate value is greater than or equal to the mutation threshold. In this embodiment, if the aforementioned condition is met, then q is recorded; otherwise, the calculation of the next historical logistics order data in the specified set is performed, as shown in step S35g below.

[0092] S35g. If so, record q in a one-dimensional array, increment q by 1, and recalculate the difference between the clustering decision value of the q-th historical logistics order data and the clustering decision value of the (q-1)-th historical logistics order data until q equals n, thus obtaining the final one-dimensional array, where the initial value of q is 2 and n is the total number of data in the specified set.

[0093] In this embodiment, the one-dimensional array is initially empty. When the difference between the clustering decision value of the second historical logistics order data and the clustering decision value of the first historical logistics order data is greater than or equal to the mutation threshold, 2 is added to the one-dimensional array. Then, the difference between the clustering decision value of the third historical logistics order data and the clustering decision value of the second historical logistics order data is calculated, and a new judgment is made. This process is repeated until all historical logistics order data in the specified set has been polled, resulting in a one-dimensional array containing multiple elements. Then, based on this one-dimensional array, the cluster mutation point can be determined, as shown in step S35h below.

[0094] S35h. From the final one-dimensional array, select the element with the largest value, and take the historical logistics order data corresponding to the element with the largest value as the cluster mutation point; in this embodiment, assuming that the one-dimensional array is [2,4,6,8], then the 8th historical logistics order data in the specified set will be taken as the cluster mutation point; of course, the above example is only an example, and the cluster mutation point will also change when the elements contained in the one-dimensional array are different, which will not be listed one by one here.

[0095] After obtaining the cluster mutation point, the cluster center can be determined based on it, as shown in step S35i below.

[0096] S35i. Use historical logistics order data that are before the cluster mutation point in the sorted sequence as the cluster center.

[0097] Therefore, after determining the cluster center through the aforementioned steps S35a to S35i, the historical logistics order data can be clustered based on this, as shown in step S36 below.

[0098] S36. Using several cluster centers, cluster all historical logistics order data to obtain several order clusters; in specific implementation, for example, but not limited to, using the k-means clustering algorithm to measure the distance of historical logistics orders, thereby obtaining several order clusters.

[0099] Furthermore, this embodiment also provides an improved clustering algorithm to improve the accuracy of clustering, the process of which is shown in steps one to eleven below.

[0100] Step 1: Sort the cluster centers in descending order of their cluster decision values ​​to obtain a cluster sequence.

[0101] Step 2: For the j-th cluster center in the clustering sequence, obtain the historical logistics order data of the k nearest neighbors of the j-th cluster center from the historical cargo transportation insurance database to form the target order set.

[0102] After forming the target order set by using the historical logistics order data of the k nearest neighbors of the jth cluster center, the nearest neighbor orders of each target order in the target order set can be clustered and divided, as shown in steps three to seven below.

[0103] Step 7: For the p-th target order in the target order set, obtain the k nearest neighbor historical logistics order data of the p-th target order as the designated order; in this embodiment, it is equivalent to performing a k nearest neighbor search on the p-th target order to obtain the corresponding designated order, and then performing clustering on the designated order, as shown in steps 4 to 7 below.

[0104] Step 4. For any specified order, calculate the third distance between the p-th target order and any specified order; in this embodiment, the cosine distance between the historical insurance feature vector of the specified order and the historical insurance feature vector of the p-th target order is still used as the third distance; then, the mean distance can be calculated, as shown in Step 5 below.

[0105] Step 5: Calculate the average distance between the p-th target order and each specified order to obtain the average distance; in this embodiment, after calculating the average distance, the distance judgment can be performed, and the process is shown in Step 6 below.

[0106] Step Six: Determine whether the third distance is less than or equal to the mean distance. In practical applications, if the third distance is greater than the mean distance, then no division is made for any specified order, and the division of the next specified order can be carried out directly. Otherwise, the specified order needs to be divided into the cluster to which the j-th cluster center belongs, as shown in Step Seven below.

[0107] Step 7: If yes, then assign any specified order to the cluster corresponding to the cluster center of the p-th target order, until all specified orders have been circumvented. The cluster center of the p-th target order is the j-th cluster center. In specific implementation, after the cluster division of any specified order is completed, the same principle can be used to divide the remaining specified orders until all specified orders have been circumvented. At this point, the clustering of the nearest neighbor orders of the p-th target order can be completed. Then, the clustering of the nearest neighbor orders of the first target order can be performed until all target orders have been circumvented, and the initial cluster of the j-th cluster center can be obtained. The circumventing clustering process is as shown in Step 8 below.

[0108] Step 8: Increment p by 1, and reacquire the historical logistics order data of the k nearest neighbors of the p-th target order until p equals k, thus obtaining the initial cluster corresponding to the j-th cluster center.

[0109] After obtaining the initial cluster corresponding to the j-th cluster center, the historical logistics order data of the k nearest neighbors of the next cluster center can be obtained. Then, in the same way as described above, the initial cluster of the next cluster center is obtained by clustering until all cluster centers have been queried. Then, several initial clusters can be obtained. The process is shown in step nine below.

[0110] Step 9: Increment j by 1, and reacquire the historical logistics order data of the k nearest neighbors of the j-th cluster center until j equals J, thus obtaining several initial clusters, where the initial value of j is 1, and j is the total number of cluster centers in the clustering sequence.

[0111] After obtaining several initial clusters, the remaining orders can be divided, as shown in steps 10 and 11 below.

[0112] Step 10: Filter out the unclustered historical logistics order data from the historical cargo transportation insurance database to serve as the remaining orders; in this embodiment, the unclustered historical logistics order data refers to the order data in the database excluding the historical logistics order data in several initial clusters.

[0113] After obtaining the remaining orders, the remaining orders can be divided, as shown in step eleven below.

[0114] Step 11: Based on the neighboring historical logistics order data of the remaining orders, divide the remaining orders into several initial clusters, so that after the division is completed, several order clusters are obtained.

[0115] In this embodiment, for any remaining order, the k nearest neighbor historical logistics order data of the remaining order are obtained as the similar order data corresponding to the remaining order. Then, the number of similar order data corresponding to the remaining order contained in each initial cluster is counted. Next, the initial cluster containing the most similar order data corresponding to the remaining order is taken as the optimal cluster of the remaining order (e.g., if initial cluster A contains 2 similar orders and initial cluster B contains 3 similar orders, then initial cluster B is taken as the optimal cluster). Finally, the remaining order is assigned to the optimal cluster, thus completing the division of the remaining order. In this way, the remaining orders are divided in the aforementioned manner to obtain several order clusters.

[0116] After the historical logistics order data is clustered through the aforementioned steps S31 to S36, the clusters can be matched based on the insurance feature vector of the target goods, as shown in step S4 below.

[0117] S4. Based on the insurance feature vector, select the order cluster most similar to the logistics order data from several order clusters as the target cluster. In this embodiment, as explained above, the cluster center of each order cluster is essentially a historical logistics order data. Therefore, the cluster center of any order cluster corresponds to a historical insurance feature vector. Based on this, this embodiment first calculates the distance between the insurance feature vector and the historical insurance feature vector of the distance center of each order cluster, and sorts each order cluster in ascending order of distance to obtain a cluster sequence. Then, the order clusters ranked in the top u positions of the cluster sequence are selected as the target clusters.

[0118] In this embodiment, u is a positive integer, and for example, it is 1 or 2; of course, it can be set according to actual use, and no specific limitation is made here.

[0119] After obtaining the target cluster, the historical logistics order data that is most similar to the logistics order data can be found in the target cluster, as shown in step S5 below.

[0120] S5. Using the insurance feature vector, determine the set of similar orders corresponding to the logistics order data from the target cluster; in this embodiment, calculate the cosine distance between the insurance feature vector and the historical insurance feature vector of each historical logistics order data in the target cluster; then, sort the historical logistics order data in the target cluster according to the cosine distance; finally, take the top 3 or 5 historical logistics order data to form a set of similar orders.

[0121] Thus, after obtaining the set of similar orders, the optimal transportation insurance plan for the target goods can be generated based on the transportation insurance information of the similar orders in the set of similar orders, as shown in step S6 below.

[0122] S6. Based on the transportation insurance information of each similar order in the similar order set, generate the optimal transportation insurance plan for the target goods and send it to the user terminal. The optimal transportation insurance plan includes cargo transportation insurance pricing information and insurance plan information. In specific implementation, for example, but not limited to, statistically analyzing the frequency of different types of insurance (such as basic insurance, comprehensive insurance, and supplementary insurance) in the transportation insurance information of each similar order, and using the insurance type with the highest frequency as the recommended insurance type.

[0123] Simultaneously, the average insured amount in the transportation insurance information of various similar orders is calculated as the recommended insured amount; next, the average premium in the transportation insurance information of various similar orders is calculated as the cargo transportation insurance pricing information; furthermore, the deductible settings in the transportation insurance information of various similar orders can be statistically analyzed to obtain the recommended deductible; thus, the recommended insurance type, recommended insured amount, and recommended deductible can be used to form insurance plan information; finally, the cargo transportation insurance pricing information and insurance plan information can be used to generate the aforementioned optimal transportation insurance plan.

[0124] Thus, after obtaining the best transportation insurance plan, it can be sent to the user's terminal for reference; at the same time, a one-click insurance button can be set on the user's terminal; when the user clicks the one-click insurance button, the user's terminal will receive the insurance human-computer interaction operation, and the best transportation insurance plan will be used as the optimal plan for automatic insurance, as shown in step S7 below.

[0125] S7. In response to the user's interactive operation for insurance application, the optimal transportation insurance plan is sent to the insurance client to complete the transportation insurance application for the target goods. In this embodiment, the insurance client can use the insurance plan information in the optimal transportation insurance plan to apply for transportation insurance for the target goods. At the same time, it sends payment information to the user, so that the transportation insurance application for the target goods is completed after the user pays.

[0126] Therefore, through the cargo transportation insurance pricing and underwriting method based on logistics end-to-end data described in detail in steps S1 to S7 above, this invention constructs an insurance feature vector containing cargo information, transportation information, and weather information based on logistics order data. Then, it performs clustering processing on historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters. Next, using the aforementioned insurance feature vector as search features, it matches the order clusters most similar to the logistics order data, and filters out similar order sets corresponding to the logistics order data from the most similar order clusters. Then, based on the transportation insurance information of each similar order in the similar order set, it generates the optimal transportation insurance plan containing cargo transportation insurance pricing information and insurance plan information. Finally, it sends this plan to the user for reference, and after the user confirms the insurance, the automatic insurance of the target cargo is completed. Therefore, compared with traditional technologies, this invention improves efficiency, reduces labor costs, reduces user waiting time, and improves user experience; thus, it is very suitable for large-scale application and promotion.

[0127] like Figure 2 As shown, the second aspect of this embodiment provides a hardware system for implementing the cargo transportation insurance pricing and underwriting method based on end-to-end logistics data as described in the first aspect of the embodiment, comprising: The acquisition unit is used to acquire logistics order data of the target goods and a historical cargo transportation insurance database. The logistics order data includes cargo information, transportation information and weather information, and the historical cargo transportation insurance database contains several historical logistics order data and transportation insurance information corresponding to each historical logistics order data.

[0128] The feature construction unit is used to construct the insurance feature vector of the target cargo based on cargo information, transportation information and weather information in the logistics order data.

[0129] Clustering units are used to cluster historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters.

[0130] The search unit is used to select the order cluster that is most similar to the logistics order data from several order clusters based on the insurance feature vector, and use it as the target cluster.

[0131] The search unit is also used to use the insurance feature vector to determine the set of similar orders corresponding to the logistics order data from the target cluster.

[0132] The insurance planning unit is used to generate the optimal transportation insurance plan for the target goods based on the transportation insurance information of each similar order in the similar order set, and send it to the user terminal. The optimal transportation insurance plan includes cargo transportation insurance pricing information and insurance plan information.

[0133] The insurance unit is used to respond to the user's human-computer interaction operation for insurance application, and to send the optimal transportation insurance plan to the insurance client to complete the transportation insurance application for the target goods.

[0134] The working process, working details and technical effects of the system provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0135] like Figure 3 As shown, the third aspect of this embodiment provides a cargo transportation insurance pricing and underwriting device based on logistics end-to-end data. Taking the device as an electronic device as an example, it includes: a memory, a processor, and a transceiver connected in sequence. The memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer programs and execute the cargo transportation insurance pricing and underwriting method based on logistics end-to-end data as described in the first aspect of the embodiment.

[0136] For specific examples, the memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; specifically, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented using at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), PLA (Programmable Logic Array). The processor may also include a main processor and a coprocessor. The main processor, also known as the CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state.

[0137] In some embodiments, the processor may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. For example, the processor may not be limited to microprocessors of the STM32F105 series, reduced instruction set computer (RISC) microprocessors, x86 architecture processors, or processors with integrated neural network processing units (NPUs). The transceiver may be, but is not limited to, a Wi-Fi transceiver, a Bluetooth transceiver, a General Packet Radio Service (GPRS) transceiver, a ZigBee (a low-power LAN protocol based on the IEEE 802.15.4 standard) transceiver, a 3G transceiver, a 4G transceiver, and / or a 5G transceiver. Furthermore, the device may also include, but is not limited to, a power module, a display screen, and other necessary components.

[0138] The working process, working details and technical effects of the electronic device provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0139] The fourth aspect of this embodiment provides a storage medium that stores instructions containing the cargo transportation insurance pricing and underwriting method based on the full-chain logistics data described in the first aspect of the embodiment. That is, the storage medium stores instructions, and when the instructions are run on a computer, the cargo transportation insurance pricing and underwriting method based on the full-chain logistics data described in the first aspect of the embodiment is executed.

[0140] The storage medium refers to a carrier for storing data, which may include, but is not limited to, floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or memory sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0141] The working process, working details and technical effects of the storage medium provided in this embodiment can be found in the first aspect of the embodiment, and will not be repeated here.

[0142] The fifth aspect of this embodiment provides a computer program product containing instructions that, when executed on a computer, cause the computer to perform the cargo transportation insurance pricing and underwriting method based on logistics end-to-end data as described in the first aspect of this embodiment. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0143] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for pricing and underwriting cargo transportation insurance based on end-to-end logistics data, characterized in that, include: Obtain logistics order data for the target goods and a historical cargo transportation insurance database. The logistics order data includes cargo information, transportation information, and weather information, and the historical cargo transportation insurance database contains several historical logistics order data and corresponding transportation insurance information for each historical logistics order data. Based on cargo information, transportation information, and weather information in logistics order data, an insurance feature vector for the target cargo is constructed. Clustering is performed on historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters; Based on the insurance feature vector, the order cluster most similar to the logistics order data is selected from several order clusters and used as the target cluster; Using the insurance feature vector, a set of similar orders corresponding to the logistics order data is determined from the target cluster; Based on the transportation insurance information of each similar order in the similar order set, the optimal transportation insurance plan for the target goods is generated and sent to the user terminal. The optimal transportation insurance plan includes cargo transportation insurance pricing information and insurance plan information. In response to the user's interactive operation for insurance application, the optimal transportation insurance plan is sent to the insurance client to complete the transportation insurance application for the target goods.

2. The method according to claim 1, characterized in that, Based on cargo information, transportation information, and weather information from logistics order data, an insurance feature vector for the target cargo is constructed, including: Based on the cargo information, the cargo type, unit price, and attributes are determined, and cargo characteristics are generated using the cargo type, unit price, and attributes. Based on the transportation information, the transportation mode, transportation distance, transportation time and carrier information are determined; Using the aforementioned transportation mode, transportation distance, transportation duration, and carrier information, transportation characteristics are generated; Based on the weather information, generate the weather characteristics of the target cargo during transportation; The insurance feature vector is generated using the cargo features, the transportation features, and the weather features.

3. The method according to claim 1, characterized in that, Clustering was performed on historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters, including: For any historical logistics order data in the historical cargo transportation insurance database, obtain the k nearest neighbor historical logistics order data of any historical logistics order data; Calculate the similarity between any historical logistics order data and each of its nearest neighboring historical logistics order data, and calculate the cluster density of any historical logistics order data based on each similarity. Based on the k nearest neighbor historical logistics order data, the cluster distance of any historical logistics order data is calculated; Based on the cluster density and the cluster distance, the cluster decision value of any historical logistics order data is determined, and after all historical logistics order data has been queried, the cluster decision value of each historical logistics order data is obtained. Based on the clustering decision values ​​of various historical logistics order data, several cluster centers are determined; Using several cluster centers, all historical logistics order data are clustered to obtain several order clusters.

4. The method according to claim 3, characterized in that, Calculate the similarity between any historical logistics order data and each of its nearest neighboring historical logistics order data, including: For any neighboring historical logistics order data, similar historical order data corresponding to the neighboring historical logistics order data are determined from the historical cargo transportation insurance database, and a first set is formed using the neighboring historical logistics order data and the similar historical order data; Find the intersection between the first set and the second set to obtain the order intersection, wherein the second set contains any historical logistics order data and the k nearest neighbor historical logistics order data corresponding to any historical logistics order data; Based on the intersection of the orders, the similarity between any historical logistics order data and any nearest neighbor historical logistics order data is calculated.

5. The method according to claim 4, characterized in that, Based on the intersection of the orders, the similarity between any historical logistics order data and any nearest neighbor historical logistics order data is calculated, including: Determine whether any of the historical logistics order data and any of the nearest neighbor historical logistics order data exist in the order intersection; If not, then determine whether any of the historical logistics order data or any of the nearest historical logistics order data exists in the intersection of the orders; If so, then obtain the first historical insurance feature vector corresponding to any historical logistics order data, and the second historical insurance feature vector corresponding to any nearest neighbor historical logistics order data; Based on the first historical insurance feature vector and the second historical insurance feature vector, the vector distance between any historical logistics order data and any nearest historical logistics order data is calculated; Based on the total number of data points in the order intersection and the vector distance, the similarity between any historical logistics order data and any nearest neighbor historical logistics order data is calculated.

6. The method according to claim 3, characterized in that, Based on the k nearest neighbor historical logistics order data, the clustering distance of any historical logistics order data is calculated, including: Obtain similar historical order data from each neighboring historical logistics order database; Calculate the first distance between any historical logistics order data and each of its nearest historical logistics order data, and sum the obtained first distances to obtain the total first distance; For the a-th nearest historical logistics order data, calculate the second distance between the a-th nearest historical logistics order data and each similar historical order data corresponding to the a-th nearest historical logistics order data; Sum the individual second distances to obtain the total second distance; Summing the first and second total distances yields the third total distance. The initial clustering distance is obtained by multiplying the first distance between any historical logistics order data and the a-th nearest neighbor historical logistics order data by the sum of the third distances. Increment 'a' by 1 and recalculate the second distance between the a-th nearest historical logistics order data and each similar historical order data corresponding to the a-th nearest historical logistics order data, until a equals k, to obtain multiple initial clustering distances, where the initial value of a is 1; From multiple initial cluster distances, the smallest initial cluster distance is selected as the cluster distance for any historical logistics order data.

7. The method according to claim 1, characterized in that, Based on the clustering decision values ​​of various historical logistics order data, several cluster centers were determined, including: The historical logistics order data are sorted in descending order of clustering decision values ​​to obtain a sorted sequence; Based on the total number of historical logistics order data, a data truncation threshold is determined. Based on the data truncation threshold, several historical logistics order data are extracted from the sorted sequence to form a specified set; The mutation threshold is calculated based on the clustering decision value of each historical logistics order data in the specified set; For the q-th historical logistics order data in the specified set, calculate the difference between the clustering decision value of the q-th historical logistics order data and the clustering decision value of the (q-1)-th historical logistics order data, and use it as the first mutation intermediate value; Determine whether the absolute value of the first mutation intermediate value is greater than or equal to the mutation threshold; If so, record q in a one-dimensional array, increment q by 1, and recalculate the difference between the clustering decision value of the q-th historical logistics order data and the clustering decision value of the (q-1)-th historical logistics order data until q equals n, thus obtaining the final one-dimensional array. The initial value of q is 2, and n is the total number of data in the specified set. From the final one-dimensional array, select the element with the largest value, and use the historical logistics order data corresponding to the element with the largest value as the cluster mutation point; Historical logistics order data that occur before the cluster mutation point in the sorted sequence are used as cluster centers.

8. The method according to claim 1, characterized in that, Each order cluster has a corresponding historical insurance feature vector at its cluster center. Based on this insurance feature vector, the order clusters most similar to the logistics order data are selected from several order clusters as target clusters, including: Calculate the distance between the insurance feature vector and the historical insurance feature vector of the distance center of each order cluster, and sort the order clusters in ascending order of distance to obtain the cluster sequence; The order clusters that rank u first in the sorted cluster sequence are taken as the target clusters, where u is a positive integer.

9. A cargo transportation insurance pricing and underwriting system based on end-to-end logistics data, characterized in that, include: The acquisition unit is used to acquire logistics order data of the target goods and historical cargo transportation insurance database. The logistics order data includes cargo information, transportation information and weather information, and the historical cargo transportation insurance database contains several historical logistics order data and transportation insurance information corresponding to each historical logistics order data. The feature construction unit is used to construct the insurance feature vector of the target goods based on the cargo information, transportation information and weather information in the logistics order data; Clustering units are used to cluster historical logistics order data in the historical cargo transportation insurance database to obtain several order clusters. The search unit is used to select the order cluster that is most similar to the logistics order data from several order clusters based on the insurance feature vector, and use it as the target cluster. The search unit is also used to use the insurance feature vector to determine the set of similar orders corresponding to the logistics order data from the target cluster; The insurance planning unit is used to generate the best transportation insurance plan for the target goods based on the transportation insurance information of each similar order in the similar order set, and send it to the user terminal. The best transportation insurance plan includes cargo transportation insurance pricing information and insurance plan information. The insurance unit is used to respond to the user's human-computer interaction operation for insurance application, and to send the optimal transportation insurance plan to the insurance client to complete the transportation insurance application for the target goods.

10. An electronic device, characterized in that, include: The system comprises a memory, a processor, and a transceiver connected in sequence for communication. The memory is used to store computer programs, the transceiver is used to send and receive messages, and the processor is used to read the computer programs and execute the cargo transportation insurance pricing and underwriting method based on logistics end-to-end data as described in any one of claims 1 to 8.

Images