WO2020164332A1 - 一种保险业务的处理方法、装置及设备 - Google Patents

一种保险业务的处理方法、装置及设备 Download PDF

Info

Publication number
WO2020164332A1
WO2020164332A1 PCT/CN2020/070042 CN2020070042W WO2020164332A1 WO 2020164332 A1 WO2020164332 A1 WO 2020164332A1 CN 2020070042 W CN2020070042 W CN 2020070042W WO 2020164332 A1 WO2020164332 A1 WO 2020164332A1
Authority
WO
WIPO (PCT)
Prior art keywords
insurance business
business plan
plan
keyword
insurance
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/070042
Other languages
English (en)
French (fr)
Inventor
胡燚铠
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Alibaba Group Holding Ltd
Original Assignee
Alibaba Group Holding Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Alibaba Group Holding Ltd filed Critical Alibaba Group Holding Ltd
Publication of WO2020164332A1 publication Critical patent/WO2020164332A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance

Definitions

  • This manual relates to the field of computer technology, in particular to a method, device and equipment for processing insurance business.
  • users with insurance needs can view or consult related insurance business plans (ie insurance products, etc.) in relevant sales outlets or web pages to actively learn about relevant insurance business plans. Users use this method to choose which suits them. Insurance business plan.
  • insurance business clerk can also make phone recommendation or door-to-door recommendation to the user, so as to understand whether the user has insurance needs, and recommend a suitable insurance business plan to the user.
  • the purpose of the embodiments of this specification is to provide an insurance business processing method, device and equipment, so as to provide a more efficient screening or recommendation mechanism for insurance business solutions.
  • An embodiment of this specification provides a method for processing insurance business, the method includes:
  • the multiple insurance business plans are screened.
  • screening the insurance business plan according to the similarity includes:
  • the insurance business plan whose score value is greater than the predetermined score threshold is taken as the insurance business plan after the screening process.
  • the method further includes:
  • the keyword extraction for each insurance business plan through a pre-trained keyword extraction model to obtain the target keyword corresponding to each insurance business plan includes:
  • the liability clause information and/or exempt clause information contained in each insurance business plan is input into the keyword extraction model to obtain the target keyword corresponding to each insurance business plan.
  • the obtaining multiple insurance business solutions to be screened includes:
  • the selected insurance business plan is acquired as the insurance business plan to be screened.
  • the method further includes:
  • the keyword extraction model is obtained based on the word2vec algorithm or the doc2vec algorithm.
  • the determining the similarity between the target keyword corresponding to each of the insurance business schemes and a predetermined benchmark keyword includes:
  • An embodiment of this specification provides an insurance business processing device, the device includes:
  • the plan acquisition module is used to acquire multiple insurance business plans to be screened
  • the keyword extraction module is used to extract keywords for each insurance business plan through a pre-trained keyword extraction model to obtain the target keywords corresponding to each insurance business plan;
  • a similarity determination module configured to determine the similarity between the target keyword corresponding to each insurance business plan and a predetermined benchmark keyword
  • the screening module is used to screen the multiple insurance business plans according to the similarity.
  • the screening module includes:
  • a scoring unit configured to score each of the insurance business schemes according to the similarity to obtain a score value corresponding to each of the insurance business schemes
  • the screening unit is configured to use the insurance business plan with the score value greater than the predetermined score threshold as the insurance business plan after the screening process.
  • the device further includes:
  • the recommendation module is used to recommend the insurance business plan after the screening process to the user, and send the target keyword corresponding to the insurance business plan after the screening process to the user.
  • the keyword extraction module includes:
  • the information acquisition module is used to acquire the liability clause information and/or exempt clause information contained in each insurance business plan;
  • the keyword extraction unit is used to input the liability clause information and/or exempt clause information contained in each insurance business plan into the keyword extraction model to obtain the target keyword corresponding to each insurance business plan .
  • the solution acquisition module includes:
  • the selection unit is used to select a predetermined number of insurance business plans from a plurality of pre-selected insurance business plans according to a predetermined plan selection mechanism;
  • the plan acquisition unit is used to acquire the selected insurance business plan as the insurance business plan to be screened.
  • the device further includes:
  • the selection module is used to select suspect keywords from a predetermined target insurance business plan
  • the model initial building module is used to build an initial keyword extraction model according to the target insurance business plan including the suspected keyword;
  • the sample acquisition module is used to acquire a sample of an insurance business plan marked with keywords in advance;
  • the model training module is used to train the initial keyword extraction model through the insurance business plan sample to obtain the keyword extraction model.
  • the keyword extraction model is obtained based on the word2vec algorithm or the doc2vec algorithm.
  • the similarity determination module is configured to determine the similarity between the target keyword corresponding to each insurance business plan and a predetermined benchmark keyword through one or more of the following algorithms: Euclid Get distance, cosine similarity algorithm, Manhattan distance, Minkowski distance, Pearson correlation coefficient algorithm and Jaccard distance.
  • the embodiment of this specification provides a processing device for insurance business, and the processing device for insurance business includes:
  • a memory arranged to store computer-executable instructions which, when executed, cause the processor to:
  • the multiple insurance business plans are screened.
  • the embodiments of this specification obtain multiple insurance business plans to be screened, and extract keywords for each insurance business plan through a pre-trained keyword extraction model to obtain each Target keywords corresponding to insurance business plans, and then determine the similarity between the target keywords corresponding to each insurance business plan and the predetermined benchmark keywords.
  • multiple insurance business plans can be screened based on the above similarities Processing, in this way, by extracting keywords for each insurance business plan, the more important content of each insurance business plan can be extracted for the user to view, which not only reduces the time for users to view the content of the insurance business plan, but also Make users pay attention to the more important content of each insurance business plan; in addition, considering that there are often more insurance business plans to be screened for users to choose, users can still view the target keywords corresponding to each insurance business plan separately.
  • this program selects a certain number (such as one or five) of insurance business plans from the multiple insurance business plans to be screened, so as to Subsequent recommendations are given to users to view or select, thereby further reducing the time for users to view information, users can quickly find insurance business solutions that suit them, and improve the efficiency of screening or recommending insurance business solutions.
  • Figure 1 is an embodiment of an insurance business processing method in this specification
  • Figure 2 is an embodiment of another insurance business processing method in this manual
  • Figure 3 is a schematic diagram of a keyword extraction model in this specification.
  • Figure 4 is a schematic diagram of the processing logic of an insurance business in this manual
  • Figure 5 is an embodiment of an insurance business processing device in this specification
  • Figure 6 is an embodiment of an insurance business processing device in this specification.
  • the embodiments of this specification provide a method, device and equipment for processing insurance services.
  • the embodiment of this specification provides a method for processing insurance services.
  • the execution subject of the method can be a terminal device or a server, etc., where the terminal device can be a mobile terminal device such as a mobile phone or a tablet, or Equipment such as personal computers.
  • the server can be an independent server, or a server cluster composed of multiple servers, and the server can be a back-end server for insurance business, etc.
  • This method can be used in processing such as selecting a relatively optimal one from a plurality of different insurance business plans or selecting multiple relatively optimal insurance business plans.
  • the execution subject of this embodiment can be described with a server as an example.
  • a terminal device is used as the execution subject, refer to the following related content, which will not be repeated here.
  • the method may specifically include the following steps:
  • step S102 multiple insurance business plans to be screened are obtained.
  • the insurance business plan can be a combination of tangible insurance products and intangible insurance services provided by insurance institutions to the market or customers.
  • insurance business plans can be created by insurance institutions, available for customers to choose from, and available in the insurance market.
  • Financial instruments used for transactions can, in a broad sense, refer to all insurance products and insurance services that insurance institutions provide to the market or customers and can be obtained, used or consumed by customers.
  • Insurance business plans can include one or more insurance plans, such as hospital medical insurance plans or critical illness insurance plans. Different insurance business plans can have different contents.
  • hospital medical insurance plans can include the following: Premium : Xxx yuan/month; duration: 1 year; coverage: unlimited disease, unlimited treatment time; deductible: 0 deductible for major diseases, cumulative annual deductible for other diseases 10,000; reimbursement ratio: 100% reimbursement ;Waiting period: 30 days; Hospital scope: second-level and above public hospitals; Claim method: on-site payment of medical expenses/convenient online claims; Renewal: Renewable year by year, liability clause and exemption clause, etc.
  • the insurance business plan is designed by the actuarial product department of the insurance institution to design, price and file a unified business plan. Then, the insurance institution will put the insurance business plan on the market for sale. Since insurance business plans are usually designed uniformly by the actuarial product department in insurance institutions, and different users may have different own circumstances and have different requirements for insurance business. Generally, users with insurance needs can view or consult related insurance business plans (ie insurance products, etc.) in relevant sales outlets or web pages to actively learn about relevant insurance business plans. Users use this method to choose which suits them. Insurance business plan. In addition, usually, insurance business clerk can also make phone recommendation or door-to-door recommendation to the user, so as to understand whether the user has insurance needs, and recommend a suitable insurance business plan to the user.
  • related insurance business plans ie insurance products, etc.
  • insurance business clerk can also make phone recommendation or door-to-door recommendation to the user, so as to understand whether the user has insurance needs, and recommend a suitable insurance business plan to the user.
  • the embodiments of this specification provide a screening or recommendation mechanism for insurance business solutions, which may specifically include the following content:
  • each insurance business plan often contains more content.
  • the user checks the content of the insurance business plan one by one to select the appropriate insurance, the user will spend more time, but in actual situations, due to the insurance business plan
  • the content of is too much, and users often do not view all of the content, and this will cause users to miss important content (such as a certain exemption clause, etc.). Therefore, you can filter keywords in the insurance business plan for users.
  • keywords can make users pay attention to the more important content of the insurance business plan.
  • Different types of insurance business plans included in the business are used as insurance business plans to be screened, or multiple different insurance business plans contained in a certain type of insurance business can be obtained as insurance business plans to be screened, or,
  • An application program for insurance business is set, and the application program can be installed in the user's terminal device. Through this application, the user can obtain the insurance business plan contained in the insurance business from the server, the user's terminal device can display the obtained insurance business plan, and the user can select multiple insurance business plans as the insurance business plan to be screened, etc. .
  • step S104 keyword extraction is performed on each insurance business plan through the pre-trained keyword extraction model to obtain the target keyword corresponding to each insurance business plan.
  • the keyword extraction model can be a model constructed by a certain algorithm to extract keywords from a given content information.
  • One of the algorithms can be selected according to the actual situation, such as TF-IDF algorithm, TextRank algorithm or word2vec Algorithms, etc., which are not limited in the embodiments of this specification.
  • multiple insurance business plans can be selected in advance, suspect keywords between different insurance business plans can be selected from multiple insurance business plans, and the preliminary key can be constructed through the selected suspect keywords and the preset algorithm Word extraction model.
  • a certain amount of sample data can be selected to train the above preliminary keyword extraction model.
  • a large number of insurance business plans can be collected, and each insurance business can be labeled
  • the keywords included in the plan such as keywords related to insurance expenses, keywords related to exemption clauses, and related keywords of liability clauses, etc., so as to obtain a sample of the marked insurance business plan, and then pass the marked insurance business
  • the program sample trains the above-mentioned preliminary keyword extraction model to obtain specific values of related parameters in the keyword extraction model, and then obtains the trained keyword extraction model.
  • the information of each insurance business plan to be screened obtained in step S102 can be input into the trained keyword extraction model for keyword extraction, so as to obtain each The keyword corresponding to the insurance business plan (ie the target keyword).
  • the user can be reminded of which content in the corresponding insurance business plan is more important and which content requires the user to pay special attention (for example, under which circumstances need to be exempted, etc.), so that when the user views a certain insurance business plan, You only need to pay attention to the content corresponding to the corresponding keyword, which can greatly reduce the time it takes for users to view the insurance business plan lock.
  • the content corresponding to the keyword is often the more important content in the insurance business plan, allowing users to quickly confirm the insurance Whether the business plan is suitable for policyholders, etc.
  • each insurance business plan can be extracted for users to view, and considering that there are often more insurance business plans to be screened for users to choose, users can view each insurance business plan separately
  • the corresponding target keywords will still consume a lot of time. For this reason, a certain number (such as one or five, etc.) of insurance business plans can be selected from the multiple insurance business plans to be screened and recommended to the user to view or select. See the processing of step S106 and step S108 below.
  • step S106 the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword is determined.
  • the benchmark keywords can be set in a variety of ways. For example, according to the actual situation, the technical staff can choose a standard form for the description of a certain thing. In practical applications, there may be more Different description methods, such as insurance amount and insurance amount, etc., and other description methods are often different from benchmark description methods. For this reason, keywords and benchmarks can be judged by the similarity between keywords and benchmark keywords Whether the keywords are the same. Specifically, after the target keywords corresponding to each insurance business plan are obtained through the processing of the above step S104, the target keywords corresponding to each insurance business plan and the corresponding benchmark keywords can be calculated for similarity to obtain each insurance business plan. The similarity value corresponding to the target keyword in the business plan.
  • step S108 the above-mentioned multiple insurance business plans are screened according to the above-mentioned similarity.
  • the insurance business plan can be used as a unit, and the target contained in it can be calculated for each insurance business plan.
  • the sum of similarity values corresponding to keywords For example, if an insurance business plan includes 3 target keywords, and the similarity values corresponding to the 3 target keywords are 0.9, 0.8, and 0.9, respectively, the insurance business plan corresponds to similarity
  • the sum of the degree values is 2.6.
  • the sum of the similarity values corresponding to each insurance business plan can be calculated, and the corresponding insurance business plans can be sorted according to the relationship of the sum of the similarity values, and the sum of the similarity values can be arranged from large to small
  • the insurance business plans arranged in sequence can be obtained, and then, the insurance business plans that are ranked first or a predetermined number (such as ranked in the top 5) can be selected from the insurance business plans arranged in sequence, So as to get the insurance business plan after screening.
  • the selected insurance business plan may include the complete content of the insurance business plan, and may also include the target keywords corresponding to the selected insurance business plan. If it is necessary to recommend the filtered insurance business plan to the user, the target keyword corresponding to the filtered insurance business plan can be sent to the user.
  • the embodiment of this specification provides an insurance business processing method. By acquiring multiple insurance business plans to be screened, and using a pre-trained keyword extraction model to extract keywords for each insurance business plan, each insurance business plan is obtained Then, determine the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword. Finally, you can screen multiple insurance business plans based on the above similarity.
  • this solution selects a certain number (such as one or five) of insurance business solutions from the multiple insurance business solutions to be screened for subsequent recommendation to The user can view or select, thereby further reducing the time for the user to view information, and the user can quickly find an insurance business plan that suits him, which improves the efficiency of screening or recommending insurance business plans.
  • the embodiment of this specification provides a method for processing insurance services.
  • the execution subject of the method can be a terminal device or a server, etc., where the terminal device can be a mobile terminal device such as a mobile phone or a tablet computer, or Equipment such as personal computers.
  • the server can be an independent server, or a server cluster composed of multiple servers, and the server can be a back-end server for insurance business, etc.
  • This method can be used in processing such as selecting a relatively optimal one from a plurality of different insurance business plans or selecting multiple relatively optimal insurance business plans.
  • the execution subject of this embodiment can be described with a server as an example.
  • a terminal device is used as the execution subject, refer to the following related content, which will not be repeated here.
  • the method may specifically include the following steps:
  • step S202 a suspected keyword is selected from a predetermined target insurance business plan.
  • the target insurance business plan may be any insurance business plan, for example, an insurance business plan corresponding to the insurance purchased by the user.
  • Suspected keywords can be keywords or not.
  • a small amount of insurance business plans can be collected (which can be used as target insurance business plans), and then the target can be determined by comparing the contents of multiple different business plans.
  • step S204 an initial keyword extraction model is constructed according to the target insurance business plan including the above-mentioned suspected keywords.
  • a preliminary keyword extraction model can be constructed through the above-mentioned target insurance business plan containing suspected keywords and a preset algorithm.
  • the preset algorithm may include one or more parameters, and the target Insurance business plan and the meaning keywords contained in it, get the equation (or equation set) about the above parameters, you can solve the equation (or equation set), get the value of each parameter, and determine the initial key based on the obtained value of each parameter Word extraction model, so that the initial keyword extraction model can be constructed through the above method.
  • the relevant information of the target insurance business plan can be used as input data, and the input data can be input into the doc2vec algorithm for calculation, and the corresponding output result (ie Keywords), and then compare the output results with the suspected keywords in the target insurance business plan, and then adjust the values of the relevant parameters in the doc2vec algorithm, and finally get the initial keyword extraction model.
  • the output result ie Keywords
  • a certain amount of sample data can be selected to train the above-mentioned preliminary keyword extraction model. For details, please refer to the processing of step S206 and step S208 below.
  • step S206 a sample of an insurance business plan marked with keywords in advance is obtained.
  • a larger number of insurance business plans can be collected as sample data (ie, insurance business plan samples), and each insurance business plan sample can be marked Include keywords, such as keywords for insurance costs, keywords for insurance amounts, keywords for claims ratio, keywords related to exemption clauses, and keywords related to liability clauses, etc., so as to obtain a sample of the marked insurance business plan.
  • step S208 the initial keyword extraction model is trained through the insurance business plan sample to obtain the keyword extraction model.
  • the keyword extraction model can be obtained based on the word2vec algorithm or the doc2vec algorithm.
  • the word2vec algorithm can use large-scale corpus to learn the algorithm of vector representation of words
  • the doc2vec algorithm can use large-scale corpus, the algorithm of learning the vector representation of sentences, etc.
  • the keyword extraction model can be trained in a variety of ways.
  • the training process based on the word2vec algorithm can be seen as training the N-gram language model through the neural network learning algorithm , And the method to find the vector corresponding to the keyword in the training process.
  • it can be divided into CBOW (Continuous Bag-of-Word Model) and Skip-gram models.
  • CBOW Continuous Bag-of-Word Model
  • Skip-gram models According to the two methods for reducing training complexity, it can be divided into Hierarchical Softmax and Negative Sampling. The above two modes and two methods are combined, and there are actually four implementation methods.
  • CBOW also known as continuous bag of words model, is a three-layer neural network model.
  • the first layer is the input layer.
  • the current word is Wi, and a K-dimensional vector is randomly initialized for each word.
  • the input of the CBOW model is the word vector of the context of the current word.
  • the middle layer (or hidden layer) accumulates the vectors of context words to obtain the middle vector, and the third layer is a Huffman tree.
  • the corresponding training process may include: establishing a vocabulary according to the corpus, initializing a K-dimensional vector for all words in the vocabulary, and constructing a Huffman tree according to the word frequency. Train the texts in the corpus one by one.
  • the correct path from the root node to the leaf node can be determined, and the predictions made on all classifiers (non-leaf nodes) on the path can be determined.
  • the vector representation of each word can be obtained.
  • the input of the Skip-gram model is no longer multiple word vectors, but a word vector.
  • the specific training process is similar to the processing process of the above-mentioned CBOW model, and will not be repeated here.
  • the keyword extraction model based on the doc2vec algorithm can be divided into a DM (Distributed Memory) model and a DBOW (distributed bag of words) model.
  • the DM model is similar to the CBOW model in the word2vec algorithm, and the DBOW model is similar to the word2vec algorithm.
  • the Skip-gram model is similar. For details, please refer to the above related content, which will not be repeated here.
  • step S210 multiple insurance business plans to be screened are obtained.
  • the process of obtaining multiple insurance business solutions to be screened can include multiple implementation methods, which can be implemented as the processing method of step S102 in the first embodiment above, and can also be implemented in the following ways, specifically including the following steps 1 and Step two.
  • Step 1 Select a predetermined number of insurance business plans from multiple pre-selected insurance business plans according to a predetermined plan selection mechanism.
  • the scheme selection mechanism can be set in advance, which can be specifically set according to actual conditions. For example, multiple pre-selected insurance business schemes can be scored by means of scoring, and then the insurance to be screened can be selected based on the relationship between the score value Business plans, specifically, multiple pre-selected insurance business plans can be obtained, and then each pre-selected insurance business plan can be scored by means of scoring, specifically based on the pre-selected insurance business plan including the number of certain designated keywords, etc. For scoring, if the number of specified keywords is contained, the corresponding scoring value is larger, and if the number of specified keywords is contained, the corresponding scoring value is smaller.
  • a predetermined number such as 10, etc.
  • Step 2 Obtain the selected insurance business plan as an insurance business plan to be screened.
  • step S212 keyword extraction is performed on each insurance business plan through the pre-trained keyword extraction model to obtain the target keyword corresponding to each insurance business plan.
  • step S212 For the processing of the foregoing step S212, reference may be made to the related content in the step S104 in the foregoing embodiment 1, which will not be repeated here.
  • keyword extraction can be performed on the corresponding liability clause information and/or exempt clause information, which can specifically include the following steps 1 and 2.
  • Step 1 Obtain information on liability clauses and/or exempt clauses contained in each insurance business plan.
  • the insurance business plan will contain liability clauses and exemption clauses, and the content contained in the liability clauses and exemption clauses may play an important role in the subsequent insurance claims of users. Therefore, the insurance business plan can be Liability clauses and exemption clauses are used as key content to extract keywords. For this reason, liability clauses and exempt clauses can be extracted from each insurance business plan, so as to obtain information on liability clauses and exemption clauses in each insurance business plan.
  • Step 2 Input the liability clause information and/or exempt clause information contained in each insurance business plan into the keyword extraction model to obtain the target keyword corresponding to each insurance business plan.
  • the liability clause information and/or exempt clause information of the insurance business plan can be input into the keyword extraction model for calculation to obtain the target keyword corresponding to the insurance business plan.
  • the liability clause and/or exemption clause can be divided into multiple different sentences, and the specific division method can be Set according to the actual situation, for example, considering that the liability clause and the exemption clause are often composed of multiple different clauses, therefore, each clause can be regarded as a division part, thereby dividing the liability clause and the exemption clause into multiple different Statement. Then, input each sentence into the keyword extraction model as shown in Figure 3 to obtain the corresponding target keyword.
  • step S214 the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword is determined.
  • the trained keyword extraction model can be obtained through the processing of the above steps S202 to S208, and in the process of training the keyword extraction model, the benchmark keywords for the insurance business plan can also be obtained.
  • the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword can be determined through one or more of the following algorithms: Euclidean distance, cosine similarity algorithm, Manhattan distance, Minkov Skye distance, Pearson correlation coefficient algorithm and Jaccard distance. Take the cosine similarity algorithm as an example.
  • the target keyword corresponding to each insurance business plan can be obtained.
  • the keyword extraction model based on word2vec algorithm or doc2vec algorithm can also get each target
  • the vector corresponding to the keyword correspondingly, in the process of training the keyword extraction model, the vector corresponding to the benchmark keyword can also be obtained.
  • the specific processing process for determining the similarity between the target keyword and the benchmark keyword can include, for example, the algorithm used is Euclidean distance, which can be based on the vector corresponding to the target keyword Determine the corresponding coordinate point, and determine the corresponding coordinate point based on the vector corresponding to the reference keyword, and then determine the coordinate point corresponding to the Euclidean distance between the target keyword and the reference keyword, and then the above-mentioned coordinate point Substitute it into the Euclidean distance formula for calculation to obtain the similarity between the target keyword and the corresponding benchmark keyword.
  • the algorithm used is Euclidean distance, which can be based on the vector corresponding to the target keyword Determine the corresponding coordinate point, and determine the corresponding coordinate point based on the vector corresponding to the reference keyword, and then determine the coordinate point corresponding to the Euclidean distance between the target keyword and the reference keyword, and then the above-mentioned coordinate point Substitute it into the Euclidean distance formula for calculation to obtain the similarity between the target keyword and the
  • step S216 each insurance business plan is scored according to the above-mentioned similarity, and a score value corresponding to each insurance business plan is obtained.
  • the scoring mechanism can be preset, which can be set according to the actual situation. In practical applications, different scoring values can be set according to different similarity values, for example, the similarity value can be multiplied by With a fixed value, the corresponding score value is obtained.
  • the similarity between the target keyword corresponding to the insurance business plan and the predetermined benchmark keyword can be scored, and the corresponding score value can be obtained.
  • the score values in the same insurance business plan can be added to obtain the insurance The score value corresponding to the business plan.
  • step S2128 an insurance business plan with a score greater than a predetermined score threshold is used as the insurance business plan after the screening process.
  • the predetermined scoring threshold can be set according to the actual situation. Different scoring thresholds may result in a different number of insurance business solutions after screening. If the scoring threshold is larger, the number of insurance business solutions after screening will be smaller. If the scoring threshold is small, the number of insurance business plans after the screening process will be greater. Among them, if the scoring threshold is only less than one of the above scoring values, the insurance business plan after the screening process will only contain one insurance business plan. The business plan can be considered as a relatively superior or optimal insurance business plan after screening and processing.
  • step S220 the insurance business plan after the screening process is recommended to the user, and the target keyword corresponding to the insurance business plan after the screening process is sent to the user.
  • the relatively superior or optimal insurance business plan obtained through the above screening process can be sent to the user.
  • the insurance business plan after the screening process can be The target keyword is sent to the user, so that the user can determine whether the corresponding insurance business plan is suitable for the user through only the relevant information of the target keyword.
  • the embodiment of this specification provides an insurance business processing method. By acquiring multiple insurance business plans to be screened, and using a pre-trained keyword extraction model to extract keywords for each insurance business plan, each insurance business plan is obtained Then, determine the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword. Finally, you can screen multiple insurance business plans based on the above similarity.
  • each insurance business plan By extracting keywords for each insurance business plan, the more important content of each insurance business plan can be extracted for users to view, which not only reduces the time for users to view the content of insurance business plans, but also allows users to pay attention The more important content of each insurance business plan; in addition, considering that there are often more insurance business plans to be screened for users to choose, it will still consume a lot of time for users to separately view the target keywords corresponding to each insurance business plan.
  • this solution selects a certain number (such as one or five) of insurance business solutions from the multiple insurance business solutions to be screened for subsequent recommendation to The user can view or select, thereby further reducing the time for the user to view information, and the user can quickly find an insurance business plan that suits him, which improves the efficiency of screening or recommending insurance business plans.
  • the embodiment of this specification also provides an apparatus for processing insurance business, as shown in FIG.
  • the insurance business processing device includes: a solution acquisition module 501, a keyword extraction module 502, a similarity determination module 503, and a screening module 504, wherein:
  • the plan obtaining module 501 is used to obtain multiple insurance business plans to be screened;
  • the keyword extraction module 502 is configured to extract keywords for each insurance business plan through a pre-trained keyword extraction model to obtain a target keyword corresponding to each insurance business plan;
  • the similarity determination module 503 is configured to determine the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword;
  • the screening module 504 is configured to perform screening processing on the multiple insurance business schemes according to the similarity.
  • the screening module 504 includes:
  • a scoring unit configured to score each of the insurance business schemes according to the similarity to obtain a score value corresponding to each of the insurance business schemes
  • the screening unit is configured to use the insurance business plan with the score value greater than the predetermined score threshold as the insurance business plan after the screening process.
  • the device further includes:
  • the recommendation module is used to recommend the insurance business plan after the screening process to the user, and send the target keyword corresponding to the insurance business plan after the screening process to the user.
  • the keyword extraction module 502 includes:
  • the information acquisition module is used to acquire the liability clause information and/or exempt clause information contained in each insurance business plan;
  • the keyword extraction unit is used to input the liability clause information and/or exempt clause information contained in each insurance business plan into the keyword extraction model to obtain the target keyword corresponding to each insurance business plan .
  • the solution acquisition module 501 includes:
  • the selection unit is used to select a predetermined number of insurance business plans from a plurality of pre-selected insurance business plans according to a predetermined plan selection mechanism;
  • the plan acquisition unit is used to acquire the selected insurance business plan as the insurance business plan to be screened.
  • the device further includes:
  • the selection module is used to select suspect keywords from a predetermined target insurance business plan
  • the model initial building module is used to build an initial keyword extraction model according to the target insurance business plan including the suspected keyword;
  • the sample acquisition module is used to acquire a sample of an insurance business plan marked with keywords in advance;
  • the model training module is used to train the initial keyword extraction model through the insurance business plan sample to obtain the keyword extraction model.
  • the keyword extraction model is obtained based on the word2vec algorithm or the doc2vec algorithm.
  • the similarity determination module 503 is configured to determine the similarity between the target keyword corresponding to each insurance business plan and a predetermined benchmark keyword through one or more of the following algorithms: Euclidean distance, cosine similarity algorithm, Manhattan distance, Minkowski distance, Pearson correlation coefficient algorithm and Jaccard distance.
  • the embodiment of this specification provides an insurance business processing device. By acquiring multiple insurance business plans to be screened, and using a pre-trained keyword extraction model to extract keywords for each insurance business plan, each insurance business plan is obtained Then, determine the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword. Finally, you can screen multiple insurance business plans based on the above similarity.
  • each insurance business plan By extracting keywords for each insurance business plan, the more important content of each insurance business plan can be extracted for users to view, which not only reduces the time for users to view the content of insurance business plans, but also allows users to pay attention The more important content of each insurance business plan; in addition, considering that there are often more insurance business plans to be screened for users to choose, it will still consume a lot of time for users to separately view the target keywords corresponding to each insurance business plan.
  • this solution selects a certain number (such as one or five) of insurance business solutions from the multiple insurance business solutions to be screened for subsequent recommendation to The user can view or select, thereby further reducing the time for the user to view information, and the user can quickly find an insurance business plan that suits him, which improves the efficiency of screening or recommending insurance business plans.
  • the embodiment of this specification also provides an insurance business processing device, as shown in FIG. 6.
  • the processing device for the insurance business may be the server provided in the foregoing embodiment.
  • Insurance business processing equipment may have relatively large differences due to different configurations or performances, and may include one or more processors 601 and memory 602, and the memory 602 may store one or more storage applications or data. Among them, the memory 602 may be short-term storage or persistent storage.
  • the application program stored in the memory 602 may include one or more modules (not shown in the figure), and each module may include a series of computer-executable instructions in a processing device for insurance services.
  • the processor 601 may be configured to communicate with the memory 602, and execute a series of computer-executable instructions in the memory 602 on the processing device of the insurance business.
  • the insurance business processing equipment may also include one or more power supplies 603, one or more wired or wireless network interfaces 604, one or more input and output interfaces 605, and one or more keyboards 606.
  • the insurance business processing device includes a memory and one or more programs, one or more programs are stored in the memory, and one or more programs may include one or more modules, and Each module may include a series of computer-executable instructions in a processing device for insurance services, and the one or more programs configured to be executed by one or more processors include the following computer-executable instructions:
  • the multiple insurance business plans are screened.
  • the screening process of the insurance business plan according to the similarity includes:
  • the insurance business plan whose score value is greater than the predetermined score threshold is taken as the insurance business plan after the screening process.
  • the method further includes:
  • the keyword extraction for each insurance business plan through the pre-trained keyword extraction model to obtain the target keyword corresponding to each insurance business plan includes:
  • the liability clause information and/or exempt clause information contained in each insurance business plan is input into the keyword extraction model to obtain the target keyword corresponding to each insurance business plan.
  • the obtaining multiple insurance business solutions to be screened includes:
  • the selected insurance business plan is acquired as the insurance business plan to be screened.
  • the keyword extraction model is obtained based on the word2vec algorithm or the doc2vec algorithm.
  • the determining the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword includes:
  • the embodiment of this specification provides a processing device for insurance business. By obtaining multiple insurance business plans to be screened, and extracting keywords for each insurance business plan through a pre-trained keyword extraction model, each insurance business plan is obtained. Then, determine the similarity between the target keyword corresponding to each insurance business plan and the predetermined benchmark keyword. Finally, you can screen multiple insurance business plans based on the above similarity.
  • each insurance business plan By extracting keywords for each insurance business plan, the more important content of each insurance business plan can be extracted for users to view, which not only reduces the time for users to view the content of insurance business plans, but also allows users to pay attention The more important content of each insurance business plan; in addition, considering that there are often more insurance business plans to be screened for users to choose, it will still consume a lot of time for users to separately view the target keywords corresponding to each insurance business plan.
  • this solution selects a certain number (such as one or five) of insurance business solutions from the multiple insurance business solutions to be screened for subsequent recommendation to The user can view or select, thereby further reducing the time for the user to view information, and the user can quickly find an insurance business plan that suits him, which improves the efficiency of screening or recommending insurance business plans.
  • a programmable logic device Programmable Logic Device, PLD
  • PLD Programmable Logic Device
  • FPGA Field Programmable Gate Array
  • HDL Hardware Description Language
  • ABEL Advanced Boolean Expression Language
  • AHDL Altera Hardware Description Language
  • HDCal JHDL
  • Lava Lava
  • Lola MyHDL
  • PALASM RHDL
  • VHDL Very-High-Speed Integrated Circuit Hardware Description Language
  • Verilog Verilog
  • the controller can be implemented in any suitable manner.
  • the controller can take the form of, for example, a microprocessor or a processor and a computer-readable medium storing computer-readable program codes (such as software or firmware) executable by the (micro)processor. , Logic gates, switches, application specific integrated circuits (ASICs), programmable logic controllers and embedded microcontrollers.
  • controllers include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20 and Silicon Labs C8051F320, the memory controller can also be implemented as a part of the memory control logic.
  • controller in addition to implementing the controller in a purely computer-readable program code manner, it is entirely possible to program the method steps to make the controller use logic gates, switches, application specific integrated circuits, programmable logic controllers and embedded The same function can be realized in the form of a microcontroller, etc. Therefore, such a controller can be regarded as a hardware component, and the devices included in it for implementing various functions can also be regarded as a structure within the hardware component. Or even, the device for realizing various functions can be regarded as both a software module for realizing the method and a structure within a hardware component.
  • a typical implementation device is a computer.
  • the computer may be, for example, a personal computer, a laptop computer, a cell phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or Any combination of these devices.
  • one or more embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this specification may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, one or more embodiments of this specification may adopt a computer program implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The form of the product.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device.
  • the device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.
  • These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment.
  • the instructions provide steps for implementing functions specified in a flow or multiple flows in the flowchart and/or a block or multiple blocks in the block diagram.
  • the computing device includes one or more processors (CPU), input/output interfaces, network interfaces, and memory.
  • processors CPU
  • input/output interfaces network interfaces
  • memory volatile and non-volatile memory
  • the memory may include non-permanent memory in computer readable media, random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). Memory is an example of computer readable media.
  • RAM random access memory
  • ROM read-only memory
  • flash RAM flash memory
  • Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be realized by any method or technology.
  • the information can be computer-readable instructions, data structures, program modules, or other data.
  • Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, Magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media can be used to store information that can be accessed by computing devices. According to the definition in this article, computer-readable media does not include transitory media, such as modulated data signals and carrier waves.
  • one or more embodiments of this specification can be provided as a method, a system, or a computer program product. Therefore, one or more embodiments of this specification may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, one or more embodiments of this specification may adopt a computer program implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The form of the product.
  • computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • One or more embodiments of this specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules.
  • program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types.
  • One or more embodiments of this specification can also be practiced in distributed computing environments. In these distributed computing environments, tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Engineering & Computer Science (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

一种保险业务的处理方法、装置及设备，所述方法包括：获取待筛选的多个保险业务方案（S102）；通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字（S104）；确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度（S106）；根据所述相似度，对所述多个保险业务方案进行筛选处理（S108）。

Description

一种保险业务的处理方法、装置及设备 技术领域
本说明书涉及计算机技术领域，尤其涉及一种保险业务的处理方法、装置及设备。
背景技术
随着保险知识的普及人们对保险理念的加深，越来越多的人选择保险作为自己进行理财规划与风险管理的重要工具。而随着保险业务的种类以及项目的不断增多，如何在众多的保险业务方案中选择适合自己的保险业务方案成为需要解决的问题。
通常，具有保险需求的用户可以在相关的销售网点或网络页面中查看或咨询相关的保险业务方案(即保险产品等)，以主动了解相关的保险业务方案，用户采用这种方法来选择适合自己的保险业务方案进行投保。此外，通常，保险业务的业务员还可以向用户进行电话推荐或登门推荐，从而了解用户是否具有保险需求，并向用户推荐合适的保险业务方案。
然而，通过上述方式筛选保险业务方案或向用户推荐保险业务方案的处理中，需要用户花费大量的时间去查询和了解各个保险业务方案的内容、责任条款和免责条款等，而且业务员向用户推荐保险业务方案也会耗费业务员的大量时间，使得保险业务方案的筛选或向用户推荐的效率较低，因此，需要提供一种效率更高的保险业务方案的筛选或推荐机制。
发明内容
本说明书实施例的目的是提供一种保险业务的处理方法、装置及设备，以提供一种效率更高的保险业务方案的筛选或推荐机制。
为了实现上述技术方案，本说明书实施例是这样实现的：
本说明书实施例提供的一种保险业务的处理方法，所述方法包括：
获取待筛选的多个保险业务方案；
通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
根据所述相似度，对所述多个保险业务方案进行筛选处理。
可选地，根据所述相似度，对所述保险业务方案进行筛选处理，包括：
根据所述相似度对每个所述保险业务方案进行评分，得到每个所述保险业务方案对应的评分值；
将所述评分值大于预定评分阈值的保险业务方案作为筛选处理后的保险业务方案。
可选地，根据所述相似度，对所述保险业务方案进行筛选处理之后，所述方法还包括：
将筛选处理后的保险业务方案推荐给用户，并将所述筛选处理后的保险业务方案对应的目标关键字发送给所述用户。
可选地，所述通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字，包括：
获取每个所述保险业务方案中包含的责任条款信息和/或免责条款信息；
将每个所述保险业务方案中包含的责任条款信息和/或免责条款信息输入到所述关键字提取模型中，得到每个所述保险业务方案对应的目标关键字。
可选地，所述获取待筛选的多个保险业务方案，包括：
根据预定的方案选取机制从多个预选保险业务方案中选取预定数量的保险业务方案；
将选取的保险业务方案获取为待筛选的保险业务方案。
可选地，所述方法还包括：
从预定的目标保险业务方案中选取疑似关键字；
根据包括所述疑似关键字的所述目标保险业务方案构建初始关键字提取模型；
获取预先标注关键字的保险业务方案样本；
通过所述保险业务方案样本对所述初始关键字提取模型进行训练，得到所述关键字提取模型。
可选地，所述关键字提取模型基于word2vec算法或doc2vec算法得到。
可选地，所述确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，包括：
通过以下一种或多种算法，确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度：欧几里得距离、余弦相似度算法、曼哈顿距离、明可夫斯基距离、皮尔森相关系数算法和Jaccard距离。
本说明书实施例提供的一种保险业务的处理装置，所述装置包括：
方案获取模块，用于获取待筛选的多个保险业务方案；
关键字提取模块，用于通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
相似度确定模块，用于确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
筛选模块，用于根据所述相似度，对所述多个保险业务方案进行筛选处理。
可选地，所述筛选模块，包括：
评分单元，用于根据所述相似度对每个所述保险业务方案进行评分，得到每个所述保险业务方案对应的评分值；
筛选单元，用于将所述评分值大于预定评分阈值的保险业务方案作为筛选处理后的保险业务方案。
可选地，所述装置还包括：
推荐模块，用于将筛选处理后的保险业务方案推荐给用户，并将所述筛选处理后的保险业务方案对应的目标关键字发送给所述用户。
可选地，所述关键字提取模块，包括：
信息获取模块，用于获取每个所述保险业务方案中包含的责任条款信息和/或免责条款信息；
关键字提取单元，用于将每个所述保险业务方案中包含的责任条款信息和/或免责条款信息输入到所述关键字提取模型中，得到每个所述保险业务方案对应的目标关键字。
可选地，所述方案获取模块，包括：
选取单元，用于根据预定的方案选取机制从多个预选保险业务方案中选取预定数 量的保险业务方案；
方案获取单元，用于将选取的保险业务方案获取为待筛选的保险业务方案。
可选地，所述装置还包括：
选取模块，用于从预定的目标保险业务方案中选取疑似关键字；
模型初建模块，用于根据包括所述疑似关键字的所述目标保险业务方案构建初始关键字提取模型；
样本获取模块，用于获取预先标注关键字的保险业务方案样本；
模型训练模块，用于通过所述保险业务方案样本对所述初始关键字提取模型进行训练，得到所述关键字提取模型。
可选地，所述关键字提取模型基于word2vec算法或doc2vec算法得到。
可选地，所述相似度确定模块，用于通过以下一种或多种算法，确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度：欧几里得距离、余弦相似度算法、曼哈顿距离、明可夫斯基距离、皮尔森相关系数算法和Jaccard距离。
本说明书实施例提供的一种保险业务的处理设备，所述保险业务的处理设备包括：
处理器；以及
被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器：
获取待筛选的多个保险业务方案；
通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
根据所述相似度，对所述多个保险业务方案进行筛选处理。
由以上本说明书实施例提供的技术方案可见，本说明书实施例通过获取待筛选的多个保险业务方案，并通过预先训练的关键字提取模型对每个保险业务方案进行关键字提取，得到每个保险业务方案对应的目标关键字，然后，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，最后，可以根据上述相似度，对多个保险业务方案进行筛选处理，这样，通过对每个保险业务方案进行关键字提取，可以为用 户将每个保险业务方案中较重要的内容提取出来供用户查看，不仅可以减少用户查看保险业务方案内容的时间，还可以使得用户关注每个保险业务方案中较重要的内容；另外，考虑到能够供用户选择的待筛选的保险业务方案往往较多，用户分别查看每一个保险业务方案对应的目标关键字仍然会消耗大量的时间，为此，本方案基于目标关键字与基准关键字之间的相似度，从待筛选的多个保险业务方案中筛选出一定数量(如一个或五个等)的保险业务方案，以便后续推荐给用户查看或选择，从而进一步减少用户查看信息的时间，用户可以快速找到适合自己的保险业务方案，提高了保险业务方案的筛选或推荐效率。
附图说明
为了更清楚地说明本说明书实施例或现有技术中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本说明书中记载的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本说明书一种保险业务的处理方法实施例；
图2为本说明书另一种保险业务的处理方法实施例；
图3为本说明书一种关键字提取模型的示意图；
图4为本说明书一种保险业务的处理逻辑的示意图；
图5为本说明书一种保险业务的处理装置实施例；
图6为本说明书一种保险业务的处理设备实施例。
具体实施方式
本说明书实施例提供一种保险业务的处理方法、装置及设备。
为了使本技术领域的人员更好地理解本说明书中的技术方案，下面将结合本说明书实施例中的附图，对本说明书实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本说明书一部分实施例，而不是全部的实施例。基于本说明书中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都应当属于本说明书保护的范围。
实施例一
如图1所示，本说明书实施例提供一种保险业务的处理方法，该方法的执行主体可以为终端设备或服务器等，其中，该终端设备可以如手机或平板电脑等移动终端设备，还可以如个人计算机等设备。该服务器可以是独立的服务器，也可以是由多个服务器组成的服务器集群，而且，该服务器可以是保险业务的后台服务器等。该方法可以用于从多个不同的保险业务方案中选取相对最优的一个或选取相对较优的多个保险业务方案等处理中。为了提高处理效率，本实施例的执行主体可以以服务器为例进行说明，对于以终端设备作为执行主体的情况，可以参见下述相关内容，在此不再赘述。该方法具体可以包括以下步骤：
在步骤S102中，获取待筛选的多个保险业务方案。
其中，保险业务方案可以是保险机构为市场或客户提供的有形保险产品和无形保险服务的综合体，保险业务方案在狭义上可以是指由保险机构创造、可供客户选择，并可在保险市场进行交易的金融工具，在广义上可以是指保险机构向市场或客户提供并可由客户取得、利用或消费的一切保险产品和保险服务等。保险业务方案可以包括某一项或多项保险方案，例如住院医疗保险方案或重疾险方案，不同的保险业务方案，其内容可以不同，以住院医疗保险方案为例，可以包括以下内容：保费：xxx元/月；期限：1年；保障范围：不限病种、不限治疗时间；免赔额：重大疾病0免赔，其它病种年度累计免赔1万；报销比例：100％报销；等待期：30天；医院范围：二级及以上公立医院；理赔方式：上门垫付医药费/在线便捷理赔；续保：可逐年续保，责任条款和免责条款等。
在实施中，随着保险知识的普及人们对保险理念的加深，越来越多的人选择保险作为自己进行理财规划与风险管理的重要工具。而随着保险业务的种类以及项目的不断增多，如何在众多的保险业务方案中选择适合自己的保险业务方案成为需要解决的问题。
通常，保险业务方案是由保险机构中的精算产品部门进行统一的业务方案设计、定价和报备的，然后，保险机构会将保险业务方案投放到市场中进行销售。由于保险业务方案通常是由保险机构中的精算产品部门进行统一设计，而且不同的用户，其自身的情况可能不同，对保险业务的需求也会不同。通常，具有保险需求的用户可以在相关的销售网点或网络页面中查看或咨询相关的保险业务方案(即保险产品等)，以主动了解相关的保险业务方案，用户采用这种方法来选择适合自己的保险业务方案进行投保。此外，通常，保险业务的业务员还可以向用户进行电话推荐或登门推荐，从而了解用户是否具有保险需求，并向用户推荐合适的保险业务方案。
然而，通过上述方式筛选保险业务方案或向用户推荐保险业务方案的处理中，需要用户花费大量的时间去查询和了解各个保险业务方案的内容、责任条款和免责条款等，而且业务员向用户推荐保险业务方案也会耗费业务员的大量时间，使得保险业务方案的推荐效率较低，因此，需要提供一种效率更高的保险业务方案的筛选或推荐机制。为此，本说明书实施例提供一种保险业务方案的筛选或推荐机制，具体可以包括以下内容：
考虑到用户在选择保险业务方案时，能够给用户做选择的保险业务方案往往包括多个，例如，用户需要选择一款健康类保险，则相应的保险业务方案可以包括住院医疗保险方案、重疾险方案、门诊保险方案等。而且，每个保险业务方案中包含的内容往往较多，这样，如果用户逐个查看保险业务方案的内容来选择合适的保险，则用户会花费较多时间，而在实际情况下，由于保险业务方案的内容过多，用户往往不会查看其中的全部内容，而这样又会使得用户漏掉其中的重要内容(如某项免责条款等)，因此，可以为用户筛选保险业务方案中的关键字，通过关键字可以使得用户关注该保险业务方案中比较重要的内容，另外还可以通过关键字从众多的保险业务方案中筛选出相对较优的保险业务方案提供给用户，基于此，可以获取某保险业务中包含的不同种类的保险业务方案作为待筛选的保险业务方案，或者，可以获取某一种类的保险业务中包含的多个不同的保险业务方案作为待筛选的保险业务方案，再或者，可以设置有保险业务的应用程序，用户的终端设备中可以安装有该应用程序。通过该应用程序，用户可以从服务器中获取该保险业务中包含的保险业务方案，用户的终端设备可以显示获取的保险业务方案，用户可以从中选择多个保险业务方案作为待筛选的保险业务方案等。
在步骤S104中，通过预先训练的关键字提取模型对每个保险业务方案进行关键字提取，得到每个保险业务方案对应的目标关键字。
其中，关键字提取模型可以是通过某算法构建的用于从给定的内容信息中提取出关键字的模型，其中的某算法可以根据实际情况进行选取，例如TF-IDF算法、TextRank算法或word2vec算法等，本说明书实施例对此不做限定。
在实施中，可以预先选取多个保险业务方案，可以从多个保险业务方案中选取不同保险业务方案之间的疑似关键字，可以通过选取的疑似关键字和预先设定的算法构建初步的关键字提取模型。为了提升关键字提取模型查找到关键字的准确度，可以选取一定数量的样本数据对上述初步的关键字提取模型进行训练，具体地，可以搜集大量的保险业务方案，并可以标记每个保险业务方案中包含的关键字，例如保险费用的关键字、免责条款相关的关键字和责任条款的相关关键字等，从而得到已经打标的保险业务方案 样本，然后，可以通过已打标的保险业务方案样本对上述初步的关键字提取模型进行训练，得到关键字提取模型中的相关参数的具体数值，进而得到训练后的关键字提取模型。
通过上述方式得到训练后的关键字提取模型后，可以将上述步骤S102中得到的待筛选的每个保险业务方案的信息输入到训练后的关键字提取模型中进行关键字提取，从而得到每个保险业务方案对应的关键字(即目标关键字)。通过得到的关键字，可以提醒用户相应的保险业务方案中哪些内容比较重要，哪些内容需要用户特别关注(例如哪种情况下需要进行免责等)，这样，当用户查看某保险业务方案时，可以只需要关注相应的关键字对应的内容，从而可以大大缩减用户查看保险业务方案锁耗费的时间，而且，关键字对应的内容往往是保险业务方案中比较重要的内容，可以使得用户快速确认该保险业务方案是否适合投保人等。
通过上述方式，可以为用户将每个保险业务方案中较重要的内容提取出来供用户查看，而考虑到能够供用户选择的待筛选的保险业务方案往往较多，用户分别查看每一个保险业务方案对应的目标关键字，仍然会消耗大量的时间，为此，可以从待筛选的多个保险业务方案中选取一定数量(如一个或五个等)的保险业务方案推荐给用户查看或选择，具体可以参见下述步骤S106和步骤S108的处理。
在步骤S106中，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度。
在实施中，可以通过多种方式设定基准关键字，例如，可以针对实际情况，由技术人员对某一事物的描述选择一种标准的形式，在实际应用中，对于同一事物可能会有多种不同的描述方式，例如保额和保险金额等，而其它描述方式与基准描述方式之间往往存在不同，为此，可以通过关键字与基准关键字之间的相似度来判断关键字与基准关键字之间是否相同。具体地，通过上述步骤S104的处理得到每个保险业务方案对应的目标关键字后，可以将每个保险业务方案对应的目标关键字与相应的基准关键字进行相似度计算，从而得到每个保险业务方案中的目标关键字对应的相似度数值。
在步骤S108中，根据上述相似度，对上述多个保险业务方案进行筛选处理。
在实施中，通过上述步骤S106的处理，得到每个保险业务方案中的目标关键字对应的相似度数值后，可以以保险业务方案为单位，针对每个保险业务方案，可以计算其中包含的目标关键字对应的相似度数值的和，例如，某保险业务方案中包括3个目标关键字，3个目标关键字对应的相似度数值分别为0.9、0.8和0.9，则该保险业务方案对应 的相似度数值的和为2.6。通过上述方式可以计算每个保险业务方案对应的相似度数值的和，可以根据相似度数值的和的大小关系，对相应的保险业务方案进行排序，可以将相似度数值的和由大到小排列，相应的，可以得到依次排列的保险业务方案，然后，可以从依次排列的保险业务方案中选取排在第一位或排列在前的预定数量(如排列在前5位)的保险业务方案，从而得到筛选后的保险业务方案。其中，需要说明的是，筛选后的保险业务方案可以包括保险业务方案的完整内容，还可以包括筛选后的保险业务方案对应的目标关键字等。如果需要将筛选后的保险业务方案推荐给用户，则可以将筛选后的保险业务方案对应的目标关键字发送给用户。
本说明书实施例提供一种保险业务的处理方法，通过获取待筛选的多个保险业务方案，并通过预先训练的关键字提取模型对每个保险业务方案进行关键字提取，得到每个保险业务方案对应的目标关键字，然后，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，最后，可以根据上述相似度，对多个保险业务方案进行筛选处理，这样，通过对每个保险业务方案进行关键字提取，可以为用户将每个保险业务方案中较重要的内容提取出来供用户查看，不仅可以减少用户查看保险业务方案内容的时间，还可以使得用户关注每个保险业务方案中较重要的内容；另外，考虑到能够供用户选择的待筛选的保险业务方案往往较多，用户分别查看每一个保险业务方案对应的目标关键字仍然会消耗大量的时间，为此，本方案基于目标关键字与基准关键字之间的相似度，从待筛选的多个保险业务方案中筛选出一定数量(如一个或五个等)的保险业务方案，以便后续推荐给用户查看或选择，从而进一步减少用户查看信息的时间，用户可以快速找到适合自己的保险业务方案，提高了保险业务方案的筛选或推荐效率。
实施例二
如图2所示，本说明书实施例提供一种保险业务的处理方法，该方法的执行主体可以为终端设备或服务器等，其中，该终端设备可以如手机或平板电脑等移动终端设备，还可以如个人计算机等设备。该服务器可以是独立的服务器，也可以是由多个服务器组成的服务器集群，而且，该服务器可以是保险业务的后台服务器等。该方法可以用于从多个不同的保险业务方案中选取相对最优的一个或选取相对较优的多个保险业务方案等处理中。为了提高处理效率，本实施例的执行主体可以以服务器为例进行说明，对于以终端设备作为执行主体的情况，可以参见下述相关内容，在此不再赘述。该方法具体可以包括以下步骤：
在步骤S202中，从预定的目标保险业务方案中选取疑似关键字。
其中，目标保险业务方案可以是任意的保险业务方案，例如，用户已购买的保险对应的保险业务方案等。疑似关键字可以是关键字，也可以不是关键字。
在实施中，为了能够构建一个初始的关键字提取模型，可以搜集少量保险业务方案(可以作为目标保险业务方案)，然后，可以通过多个不同的业务方案中的内容比对等方式，确定目标保险业务方案中可能包含的关键字，由于通过上述方式确定的关键字可以是实际中的关键字，也可以不是实际中的关键字，因此，可以将得到的关键字称为疑似关键字，具体如，可以从每个目标保险业务方案中提取免责信息，然后，可以比对不同的目标保险业务方案的免责信息，可以将2个或更多个目标保险业务方案的免责信息中均包含的关键字作为疑似关键字等。通过上述方式，可以从预定的目标保险业务方案中选取疑似关键字。
在步骤S204中，根据包括上述疑似关键字的目标保险业务方案构建初始关键字提取模型。
在实施中，可以通过包含疑似关键字的上述目标保险业务方案和预先设定的算法构建初步的关键字提取模型，具体地，预先设定的算法中可以包括一个或多个参数，可以通过目标保险业务方案和其中包含的意思关键字，得到关于上述参数的方程(或方程组)，可以求解该方程(或方程组)，得到各个参数的数值，基于得到的各个参数的数值，确定初始关键字提取模型，从而通过上述方式可以构建初始关键字提取模型。
例如，以上述算法为doc2vec算法为例，如图3所示，可以将目标保险业务方案的相关信息作为输入数据，将可以该输入数据输入到doc2vec算法中进行计算，得到相应的输出结果(即关键字)，然后，再通过目标保险业务方案中的疑似关键字对输出结果进行比对，进而调整doc2vec算法中的相关参数的数值，最终可以得到初始关键字提取模型。
为了提升关键字提取模型查找到关键字的准确度，可以选取一定数量的样本数据对上述初步的关键字提取模型进行训练，具体可以参见下述步骤S206和步骤S208的处理。
在步骤S206中，获取预先标注关键字的保险业务方案样本。
在实施中，为了使得关键字提取模型查找到关键字的准确度更高，可以搜集较多数量的保险业务方案作为样本数据(即保险业务方案样本)，并可以标记每个保险业务方案样本中包含的关键字，例如保险费用的关键字、保险金额的关键字、理赔比例的关 键字、免责条款相关的关键字和责任条款的相关关键字等，从而得到已打标的保险业务方案样本。
在步骤S208中，通过保险业务方案样本对初始关键字提取模型进行训练，得到关键字提取模型。
其中，关键字提取模型可以基于word2vec算法或doc2vec算法得到。其中的word2vec算法可以通过大规模语料，学习词语的向量表示的算法，其中的doc2vec算法可以通过大规模语料，学习语句的向量表示的算法等。
在实施中，可以通过多种方式对关键字提取模型进行训练，针对基于word2vec算法得到的关键字提取模型，基于word2vec算法的训练过程可以看做是通过神经网络学习算法来训练N-gram语言模型，并在训练过程中求出关键字所对应的向量的方法。根据语言模型的不同，又可分为CBOW(Continuous Bag-of-Word Model)和Skip-gram两种模型。而根据两种降低训练复杂度的方法又可分为Hierarchical Softmax和Negative Sampling，上述两种模式和两种方法进行组合，实际上可以有四种实现方式。CBOW又称连续词袋模型，是一个三层神经网络模型。其中，第一层是输入层，首先选定输入词的数量，当前词为Wi，每一个词随机初始化一个K维向量，则CBOW模型的输入是当前词的上下文的词的词向量。中间层(或隐层)将上下文词的向量累加得到中间向量，第三层是一颗哈夫曼树。相应的训练过程可以包括：根据语料库建立词汇表，词汇表中的所有词均初始化一个K维向量，并根据词频构建哈夫曼树。将语料库中的文本依次进行训练，根据已知词的哈夫曼编码，可以确定从根节点到叶节点的正确路径，也确定路径上所有分类器(非叶节点)上作出的预测，通过上述过程最终即可可得到每个词的向量表示。
Skip-gram模型的输入不再是多个词向量，而是一个词向量。具体训练过程与上述CBOW模型的处理过程类似，在此不再赘述。
针对基于doc2vec算法得到的关键字提取模型，可以分为DM(Distributed Memory)模型和DBOW(distributed bag of words)模型，其中DM模型与上述word2vec算法中的CBOW模型类似，DBOW模型与上述word2vec算法中的Skip-gram模型类似，具体可以参见上述相关内容，在此不再赘述。
在步骤S210中，获取待筛选的多个保险业务方案。
在实施中，获取待筛选的多个保险业务方案的处理可以包括多种实现方式，可以 如上述实施例一中步骤S102的处理方式实现，还可以通过以下方式实现，具体可以包括以下步骤一和步骤二。
步骤一，根据预定的方案选取机制从多个预选保险业务方案中选取预定数量的保险业务方案。
在实施中，可以预先设定方案选取机制，具体可以根据实际情况设定，例如可以通过打分的方式为多个预选保险业务方案进行打分，然后，通过打分分值的大小关系选取待筛选的保险业务方案，具体地，可以获取多个预选保险业务方案，然后，可以通过打分的方式，对每个预选保险业务方案进行打分，具体可以基于预选保险业务方案包括某些指定关键字的数量等进行打分，如果包含指定关键字的数量较多，则相应的打分分值较大，如果包含指定关键字的数量较少，则相应的打分分值较小。然后，可以选取打分分值超过预定阈值的预选保险业务方案，或者，根据打分分值由大到小，对多个预选保险业务方案进行排序，在根据预先设定的选取数量(如10个等)，从上述排序结果中选取排列在前的选取数量的预选保险业务方案。
步骤二，将选取的保险业务方案获取为待筛选的保险业务方案。
在步骤S212中，通过预先训练的关键字提取模型对每个保险业务方案进行关键字提取，得到每个保险业务方案对应的目标关键字。
上述步骤S212的处理可以参见上述实施例一中步骤S104中的相关内容，在此不再赘述。
此外，对于保险业务方案中包括责任条款和/或免责条款的情况，可以对相应的责任条款信息和/或免责条款信息进行关键字提取，具体可以包括以下步骤一和步骤二的处理。
步骤一，获取每个保险业务方案中包含的责任条款信息和/或免责条款信息。
在实施中，通常，保险业务方案中会包含责任条款和免责条款，而责任条款和免责条款中包含的内容对于用户后续进行保险理赔可能会起到重要作用，因此，可以将保险业务方案中的责任条款和免责条款作为重点内容进行关键字提取，为此，可以从每个保险业务方案中提取责任条款和免责条款，从而得到每个保险业务方案中的责任条款信息和免责条款信息。
步骤二，将每个保险业务方案中包含的责任条款信息和/或免责条款信息输入到关键字提取模型中，得到每个保险业务方案对应的目标关键字。
在实施中，针对任一个保险业务方案，可以将该保险业务方案的责任条款信息和/或免责条款信息输入到关键字提取模型进行计算，得到该保险业务方案对应的目标关键字。另外，考虑到保险业务方案中的责任条款和/或免责条款的全部内容可能会较多，为了提高处理效率，可以将责任条款和/或免责条款划分为多个不同的语句，具体划分方式可以根据实际情况设定，例如，考虑到责任条款和免责条款中往往是由多个不同的条款构成，因此，可以将每一个条款作为一个划分部分，从而将责任条款和免责条款划分为多个不同的语句。然后，分别将每一个语句输入到如图3所示的关键字提取模型中，得到相应的目标关键字。
在步骤S214中，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度。
在实施中，在通过上述步骤S202～步骤S208的处理可以得到训练后的关键字提取模型，而在训练关键字提取模型的过程中，还可以得到针对保险业务方案的基准关键字。
可以通过以下一种或多种算法，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度：欧几里得距离、余弦相似度算法、曼哈顿距离、明可夫斯基距离、皮尔森相关系数算法和Jaccard距离。以余弦相似度算法为例，通过上述关键字提取模型，可以得到每个保险业务方案对应的目标关键字，不仅如此，基于word2vec算法或doc2vec算法等得到的关键字提取模型还可以得到每个目标关键字对应的向量，相应的，在训练关键字提取模型的过程中，还可以得到基准关键字对应的向量。然后，可以任意选取一个目标关键字，并获取该目标关键字对应的向量，并且获取相应的基准关键字对应的向量，计算该目标关键字对应的向量与相应的基准关键字对应的向量之间的夹角的余弦值，从而得到该目标关键字与相应的基准关键字之间的相似度。通过上述方式可以得到每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度。
需要说明的是，不同的算法，确定目标关键字与基准关键字之间的相似度的具体处理过程可以包括，例如，采用的算法为欧几里得距离，则可以基于目标关键字对应的向量确定相应的坐标点，并可以基于基准关键字对应的向量确定相应的坐标点，进而确定目标关键字与基准关键字之间的欧几里得距离对应的坐标点，然后，可以将上述坐标点代入到欧几里得距离公式中进行计算，得到目标关键字与相应的基准关键字之间的相似度等。
在步骤S216中，根据上述相似度对每个保险业务方案进行评分，得到每个保险业务方案对应的评分值。
在实施中，如图4所示，可以预先设定评分机制，具体可以根据实际情况设定，在实际应用中，可以根据不同的相似度数值设置不同的评分值，例如可以将相似度数值乘以固定数值，得到相应的评分值等。通过上述评分机制，可以对保险业务方案对应的目标关键字与预定的基准关键字之间的相似度进行评分，得到相应的评分值，可以将同一保险业务方案中的评分值相加得到该保险业务方案对应的评分值。
在步骤S218中，将评分值大于预定评分阈值的保险业务方案作为筛选处理后的保险业务方案。
其中，预定评分阈值可以根据实际情况设定，不同的评分阈值，最终得到的筛选处理后的保险业务方案的数量可能不同，如果评分阈值较大，则筛选处理后的保险业务方案的数量会较少，如果评分阈值较小，则筛选处理后的保险业务方案的数量会较多，其中，如果评分阈值仅小于上述评分值中的一个评分值，则筛选处理后的保险业务方案仅包含一个保险业务方案，可以认为筛选处理后的保险业务方案为相对较优或最优的保险业务方案。
在步骤S220中，将筛选处理后的保险业务方案推荐给用户，并将筛选处理后的保险业务方案对应的目标关键字发送给该用户。
在实施中，可以将通过上述筛选处理得到的相对较优或最优的保险业务方案发送给用户，同时，为了减少用户查看保险业务方案内容的时间，可以将筛选处理后的保险业务方案对应的目标关键字发送给该用户，这样，用户可以仅通过目标关键字的相关信息确定相应的保险业务方案是否适合该用户。
本说明书实施例提供一种保险业务的处理方法，通过获取待筛选的多个保险业务方案，并通过预先训练的关键字提取模型对每个保险业务方案进行关键字提取，得到每个保险业务方案对应的目标关键字，然后，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，最后，可以根据上述相似度，对多个保险业务方案进行筛选处理，这样，通过对每个保险业务方案进行关键字提取，可以为用户将每个保险业务方案中较重要的内容提取出来供用户查看，不仅可以减少用户查看保险业务方案内容的时间，还可以使得用户关注每个保险业务方案中较重要的内容；另外，考虑到能够供用户选择的待筛选的保险业务方案往往较多，用户分别查看每一个保险业务方案对应的目标关键字仍然会消耗大量的时间，为此，本方案基于目标关键字与基准关键字之间的相似度，从待筛选的多个保险业务方案中筛选出一定数量(如一个或五个等)的保险业务方案，以便后续推荐给用户查看或选择，从而进一步减少用户查看信息的时间，用 户可以快速找到适合自己的保险业务方案，提高了保险业务方案的筛选或推荐效率。
实施例三
以上为本说明书实施例提供的保险业务的处理方法，基于同样的思路，本说明书实施例还提供一种保险业务的处理装置，如图5所示。
该保险业务的处理装置包括：方案获取模块501、关键字提取模块502、相似度确定模块503和筛选模块504，其中：
方案获取模块501，用于获取待筛选的多个保险业务方案；
关键字提取模块502，用于通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
相似度确定模块503，用于确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
筛选模块504，用于根据所述相似度，对所述多个保险业务方案进行筛选处理。
本说明书实施例中，所述筛选模块504，包括：
评分单元，用于根据所述相似度对每个所述保险业务方案进行评分，得到每个所述保险业务方案对应的评分值；
筛选单元，用于将所述评分值大于预定评分阈值的保险业务方案作为筛选处理后的保险业务方案。
本说明书实施例中，所述装置还包括：
推荐模块，用于将筛选处理后的保险业务方案推荐给用户，并将所述筛选处理后的保险业务方案对应的目标关键字发送给所述用户。
本说明书实施例中，所述关键字提取模块502，包括：
信息获取模块，用于获取每个所述保险业务方案中包含的责任条款信息和/或免责条款信息；
关键字提取单元，用于将每个所述保险业务方案中包含的责任条款信息和/或免责条款信息输入到所述关键字提取模型中，得到每个所述保险业务方案对应的目标关键字。
本说明书实施例中，所述方案获取模块501，包括：
选取单元，用于根据预定的方案选取机制从多个预选保险业务方案中选取预定数 量的保险业务方案；
方案获取单元，用于将选取的保险业务方案获取为待筛选的保险业务方案。
本说明书实施例中，所述装置还包括：
选取模块，用于从预定的目标保险业务方案中选取疑似关键字；
模型初建模块，用于根据包括所述疑似关键字的所述目标保险业务方案构建初始关键字提取模型；
样本获取模块，用于获取预先标注关键字的保险业务方案样本；
模型训练模块，用于通过所述保险业务方案样本对所述初始关键字提取模型进行训练，得到所述关键字提取模型。
本说明书实施例中，所述关键字提取模型基于word2vec算法或doc2vec算法得到。
本说明书实施例中，所述相似度确定模块503，用于通过以下一种或多种算法，确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度：欧几里得距离、余弦相似度算法、曼哈顿距离、明可夫斯基距离、皮尔森相关系数算法和Jaccard距离。
本说明书实施例提供一种保险业务的处理装置，通过获取待筛选的多个保险业务方案，并通过预先训练的关键字提取模型对每个保险业务方案进行关键字提取，得到每个保险业务方案对应的目标关键字，然后，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，最后，可以根据上述相似度，对多个保险业务方案进行筛选处理，这样，通过对每个保险业务方案进行关键字提取，可以为用户将每个保险业务方案中较重要的内容提取出来供用户查看，不仅可以减少用户查看保险业务方案内容的时间，还可以使得用户关注每个保险业务方案中较重要的内容；另外，考虑到能够供用户选择的待筛选的保险业务方案往往较多，用户分别查看每一个保险业务方案对应的目标关键字仍然会消耗大量的时间，为此，本方案基于目标关键字与基准关键字之间的相似度，从待筛选的多个保险业务方案中筛选出一定数量(如一个或五个等)的保险业务方案，以便后续推荐给用户查看或选择，从而进一步减少用户查看信息的时间，用户可以快速找到适合自己的保险业务方案，提高了保险业务方案的筛选或推荐效率。
实施例四
以上为本说明书实施例提供的保险业务的处理装置，基于同样的思路，本说明书 实施例还提供一种保险业务的处理设备，如图6所示。
所述保险业务的处理设备可以为上述实施例提供的服务器。
保险业务的处理设备可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上的处理器601和存储器602，存储器602中可以存储有一个或一个以上存储应用程序或数据。其中，存储器602可以是短暂存储或持久存储。存储在存储器602的应用程序可以包括一个或一个以上模块(图示未示出)，每个模块可以包括对保险业务的处理设备中的一系列计算机可执行指令。更进一步地，处理器601可以设置为与存储器602通信，在保险业务的处理设备上执行存储器602中的一系列计算机可执行指令。保险业务的处理设备还可以包括一个或一个以上电源603，一个或一个以上有线或无线网络接口604，一个或一个以上输入输出接口605，一个或一个以上键盘606。
具体在本实施例中，保险业务的处理设备包括有存储器，以及一个或一个以上的程序，其中一个或者一个以上程序存储于存储器中，且一个或者一个以上程序可以包括一个或一个以上模块，且每个模块可以包括对保险业务的处理设备中的一系列计算机可执行指令，且经配置以由一个或者一个以上处理器执行该一个或者一个以上程序包含用于进行以下计算机可执行指令：
获取待筛选的多个保险业务方案；
通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
根据所述相似度，对所述多个保险业务方案进行筛选处理。
本说明书实施例中，所述根据所述相似度，对所述保险业务方案进行筛选处理，包括：
根据所述相似度对每个所述保险业务方案进行评分，得到每个所述保险业务方案对应的评分值；
将所述评分值大于预定评分阈值的保险业务方案作为筛选处理后的保险业务方案。
本说明书实施例中，所述根据所述相似度，对所述保险业务方案进行筛选处理之后，还包括：
将筛选处理后的保险业务方案推荐给用户，并将所述筛选处理后的保险业务方案 对应的目标关键字发送给所述用户。
本说明书实施例中，所述通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字，包括：
获取每个所述保险业务方案中包含的责任条款信息和/或免责条款信息；
将每个所述保险业务方案中包含的责任条款信息和/或免责条款信息输入到所述关键字提取模型中，得到每个所述保险业务方案对应的目标关键字。
本说明书实施例中，所述获取待筛选的多个保险业务方案，包括：
根据预定的方案选取机制从多个预选保险业务方案中选取预定数量的保险业务方案；
将选取的保险业务方案获取为待筛选的保险业务方案。
本说明书实施例中，还包括：
从预定的目标保险业务方案中选取疑似关键字；
根据包括所述疑似关键字的所述目标保险业务方案构建初始关键字提取模型；
获取预先标注关键字的保险业务方案样本；
通过所述保险业务方案样本对所述初始关键字提取模型进行训练，得到所述关键字提取模型。
本说明书实施例中，所述关键字提取模型基于word2vec算法或doc2vec算法得到。
本说明书实施例中，所述确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，包括：
通过以下一种或多种算法，确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度：欧几里得距离、余弦相似度算法、曼哈顿距离、明可夫斯基距离、皮尔森相关系数算法和Jaccard距离。
本说明书实施例提供一种保险业务的处理设备，通过获取待筛选的多个保险业务方案，并通过预先训练的关键字提取模型对每个保险业务方案进行关键字提取，得到每个保险业务方案对应的目标关键字，然后，确定每个保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，最后，可以根据上述相似度，对多个保险业务方案进行筛选处理，这样，通过对每个保险业务方案进行关键字提取，可以为用户将每个保险 业务方案中较重要的内容提取出来供用户查看，不仅可以减少用户查看保险业务方案内容的时间，还可以使得用户关注每个保险业务方案中较重要的内容；另外，考虑到能够供用户选择的待筛选的保险业务方案往往较多，用户分别查看每一个保险业务方案对应的目标关键字仍然会消耗大量的时间，为此，本方案基于目标关键字与基准关键字之间的相似度，从待筛选的多个保险业务方案中筛选出一定数量(如一个或五个等)的保险业务方案，以便后续推荐给用户查看或选择，从而进一步减少用户查看信息的时间，用户可以快速找到适合自己的保险业务方案，提高了保险业务方案的筛选或推荐效率。
上述对本说明书特定实施例进行了描述。其它实施例在所附权利要求书的范围内。在一些情况下，在权利要求书中记载的动作或步骤可以按照不同于实施例中的顺序来执行并且仍然可以实现期望的结果。另外，在附图中描绘的过程不一定要求示出的特定顺序或者连续顺序才能实现期望的结果。在某些实施方式中，多任务处理和并行处理也是可以的或者可能是有利的。
在20世纪90年代，对于一个技术的改进可以很明显地区分是硬件上的改进(例如，对二极管、晶体管、开关等电路结构的改进)还是软件上的改进(对于方法流程的改进)。然而，随着技术的发展，当今的很多方法流程的改进已经可以视为硬件电路结构的直接改进。设计人员几乎都通过将改进的方法流程编程到硬件电路中来得到相应的硬件电路结构。因此，不能说一个方法流程的改进就不能用硬件实体模块来实现。例如，可编程逻辑器件(Programmable Logic Device，PLD)(例如现场可编程门阵列(Field Programmable Gate Array，FPGA))就是这样一种集成电路，其逻辑功能由用户对器件编程来确定。由设计人员自行编程来把一个数字系统“集成”在一片PLD上，而不需要请芯片制造厂商来设计和制作专用的集成电路芯片。而且，如今，取代手工地制作集成电路芯片，这种编程也多半改用“逻辑编译器(logic compiler)”软件来实现，它与程序开发撰写时所用的软件编译器相类似，而要编译之前的原始代码也得用特定的编程语言来撰写，此称之为硬件描述语言(Hardware Description Language，HDL)，而HDL也并非仅有一种，而是有许多种，如ABEL(Advanced Boolean Expression Language)、AHDL(Altera Hardware Description Language)、Confluence、CUPL(Cornell University Programming Language)、HDCal、JHDL(Java Hardware Description Language)、Lava、Lola、MyHDL、PALASM、RHDL(Ruby Hardware Description Language)等，目前最普遍使用的是VHDL(Very-High-Speed Integrated Circuit Hardware Description Language)与Verilog。本领域技术人员也应该清楚，只需要将方法流程用上述几种硬件描述语言稍作逻辑编程并编程到集成电路中，就可以很容易得到实现该逻辑方法流程的硬件电路。
控制器可以按任何适当的方式实现，例如，控制器可以采取例如微处理器或处理器以及存储可由该(微)处理器执行的计算机可读程序代码(例如软件或固件)的计算机可读介质、逻辑门、开关、专用集成电路(Application Specific Integrated Circuit，ASIC)、可编程逻辑控制器和嵌入微控制器的形式，控制器的例子包括但不限于以下微控制器：ARC 625D、Atmel AT91SAM、Microchip PIC18F26K20以及Silicone Labs C8051F320，存储器控制器还可以被实现为存储器的控制逻辑的一部分。本领域技术人员也知道，除了以纯计算机可读程序代码方式实现控制器以外，完全可以通过将方法步骤进行逻辑编程来使得控制器以逻辑门、开关、专用集成电路、可编程逻辑控制器和嵌入微控制器等的形式来实现相同功能。因此这种控制器可以被认为是一种硬件部件，而对其内包括的用于实现各种功能的装置也可以视为硬件部件内的结构。或者甚至，可以将用于实现各种功能的装置视为既可以是实现方法的软件模块又可以是硬件部件内的结构。
上述实施例阐明的系统、装置、模块或单元，具体可以由计算机芯片或实体实现，或者由具有某种功能的产品来实现。一种典型的实现设备为计算机。具体的，计算机例如可以为个人计算机、膝上型计算机、蜂窝电话、相机电话、智能电话、个人数字助理、媒体播放器、导航设备、电子邮件设备、游戏控制台、平板计算机、可穿戴设备或者这些设备中的任何设备的组合。
为了描述的方便，描述以上装置时以功能分为各种单元分别描述。当然，在实施本说明书一个或多个实施例时可以把各单元的功能在同一个或多个软件和/或硬件中实现。
本领域内的技术人员应明白，本说明书的实施例可提供为方法、系统、或计算机程序产品。因此，本说明书一个或多个实施例可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本说明书一个或多个实施例可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本说明书的实施例是参照根据本说明书实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方 框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
在一个典型的配置中，计算设备包括一个或多个处理器(CPU)、输入/输出接口、网络接口和内存。
内存可能包括计算机可读介质中的非永久性存储器，随机存取存储器(RAM)和/或非易失性内存等形式，如只读存储器(ROM)或闪存(flash RAM)。内存是计算机可读介质的示例。
计算机可读介质包括永久性和非永久性、可移动和非可移动媒体可以由任何方法或技术来实现信息存储。信息可以是计算机可读指令、数据结构、程序的模块或其他数据。计算机的存储介质的例子包括，但不限于相变内存(PRAM)、静态随机存取存储器(SRAM)、动态随机存取存储器(DRAM)、其他类型的随机存取存储器(RAM)、只读存储器(ROM)、电可擦除可编程只读存储器(EEPROM)、快闪记忆体或其他内存技术、只读光盘只读存储器(CD-ROM)、数字多功能光盘(DVD)或其他光学存储、磁盒式磁带，磁带磁磁盘存储或其他磁性存储设备或任何其他非传输介质，可用于存储可以被计算设备访问的信息。按照本文中的界定，计算机可读介质不包括暂存电脑可读媒体(transitory media)，如调制的数据信号和载波。
还需要说明的是，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、商品或者设备不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、商品或者设备所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括所述要素的过程、方法、商品或者设备中还存在另外的相同要素。
本领域技术人员应明白，本说明书的实施例可提供为方法、系统或计算机程序产 品。因此，本说明书一个或多个实施例可采用完全硬件实施例、完全软件实施例或结合软件和硬件方面的实施例的形式。而且，本说明书一个或多个实施例可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本说明书一个或多个实施例可以在由计算机执行的计算机可执行指令的一般上下文中描述，例如程序模块。一般地，程序模块包括执行特定任务或实现特定抽象数据类型的例程、程序、对象、组件、数据结构等等。也可以在分布式计算环境中实践本说明书一个或多个实施例，在这些分布式计算环境中，由通过通信网络而被连接的远程处理设备来执行任务。在分布式计算环境中，程序模块可以位于包括存储设备在内的本地和远程计算机存储介质中。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例重点说明的都是与其他实施例的不同之处。尤其，对于系统实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。
以上所述仅为本说明书的实施例而已，并不用于限制本说明书。对于本领域技术人员来说，本说明书可以有各种更改和变化。凡在本说明书的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本说明书的权利要求范围之内。

Claims (17)

  1. 一种保险业务的处理方法，所述方法包括：
    获取待筛选的多个保险业务方案；
    通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
    确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
    根据所述相似度，对所述多个保险业务方案进行筛选处理。
  2. 根据权利要求1所述的方法，根据所述相似度，对所述保险业务方案进行筛选处理，包括：
    根据所述相似度对每个所述保险业务方案进行评分，得到每个所述保险业务方案对应的评分值；
    将所述评分值大于预定评分阈值的保险业务方案作为筛选处理后的保险业务方案。
  3. 根据权利要求1或2所述的方法，根据所述相似度，对所述保险业务方案进行筛选处理之后，所述方法还包括：
    将筛选处理后的保险业务方案推荐给用户，并将所述筛选处理后的保险业务方案对应的目标关键字发送给所述用户。
  4. 根据权利要求1所述的方法，通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字，包括：
    获取每个所述保险业务方案中包含的责任条款信息和/或免责条款信息；
    将每个所述保险业务方案中包含的责任条款信息和/或免责条款信息输入到所述关键字提取模型中，得到每个所述保险业务方案对应的目标关键字。
  5. 根据权利要求1所述的方法，获取待筛选的多个保险业务方案，包括：
    根据预定的方案选取机制从多个预选保险业务方案中选取预定数量的保险业务方案；
    将选取的保险业务方案获取为待筛选的保险业务方案。
  6. 根据权利要求1所述的方法，所述方法还包括：
    从预定的目标保险业务方案中选取疑似关键字；
    根据包括所述疑似关键字的所述目标保险业务方案构建初始关键字提取模型；
    获取预先标注关键字的保险业务方案样本；
    通过所述保险业务方案样本对所述初始关键字提取模型进行训练，得到所述关键字提取模型。
  7. 根据权利要求1所述的方法，所述关键字提取模型基于word2vec算法或doc2vec算法得到。
  8. 根据权利要求1所述的方法，确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度，包括：
    通过以下一种或多种算法，确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度：欧几里得距离、余弦相似度算法、曼哈顿距离、明可夫斯基距离、皮尔森相关系数算法和Jaccard距离。
  9. 一种保险业务的处理装置，所述装置包括：
    方案获取模块，用于获取待筛选的多个保险业务方案；
    关键字提取模块，用于通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
    相似度确定模块，用于确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
    筛选模块，用于根据所述相似度，对所述多个保险业务方案进行筛选处理。
  10. 根据权利要求9所述的装置，所述筛选模块，包括：
    评分单元，用于根据所述相似度对每个所述保险业务方案进行评分，得到每个所述保险业务方案对应的评分值；
    筛选单元，用于将所述评分值大于预定评分阈值的保险业务方案作为筛选处理后的保险业务方案。
  11. 根据权利要求9或10所述的装置，所述装置还包括：
    推荐模块，用于将筛选处理后的保险业务方案推荐给用户，并将所述筛选处理后的保险业务方案对应的目标关键字发送给所述用户。
  12. 根据权利要求9所述的装置，所述关键字提取模块，包括：
    信息获取模块，用于获取每个所述保险业务方案中包含的责任条款信息和/或免责条款信息；
    关键字提取单元，用于将每个所述保险业务方案中包含的责任条款信息和/或免责条款信息输入到所述关键字提取模型中，得到每个所述保险业务方案对应的目标关键字。
  13. 根据权利要求9所述的装置，所述方案获取模块，包括：
    选取单元，用于根据预定的方案选取机制从多个预选保险业务方案中选取预定数量的保险业务方案；
    方案获取单元，用于将选取的保险业务方案获取为待筛选的保险业务方案。
  14. 根据权利要求9所述的装置，所述装置还包括：
    选取模块，用于从预定的目标保险业务方案中选取疑似关键字；
    模型初建模块，用于根据包括所述疑似关键字的所述目标保险业务方案构建初始关键字提取模型；
    样本获取模块，用于获取预先标注关键字的保险业务方案样本；
    模型训练模块，用于通过所述保险业务方案样本对所述初始关键字提取模型进行训练，得到所述关键字提取模型。
  15. 根据权利要求9所述的装置，所述关键字提取模型基于word2vec算法或doc2vec算法得到。
  16. 根据权利要求9所述的装置，所述相似度确定模块，用于通过以下一种或多种算法，确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度：欧几里得距离、余弦相似度算法、曼哈顿距离、明可夫斯基距离、皮尔森相关系数算法和Jaccard距离。
  17. 一种保险业务的处理设备，所述保险业务的处理设备包括：
    处理器；以及
    被安排成存储计算机可执行指令的存储器，所述可执行指令在被执行时使所述处理器：
    获取待筛选的多个保险业务方案；
    通过预先训练的关键字提取模型对每个所述保险业务方案进行关键字提取，得到每个所述保险业务方案对应的目标关键字；
    确定每个所述保险业务方案对应的目标关键字与预定的基准关键字之间的相似度；
    根据所述相似度，对所述多个保险业务方案进行筛选处理。
PCT/CN2020/070042 2019-02-12 2020-01-02 一种保险业务的处理方法、装置及设备 Ceased WO2020164332A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910110784.6 2019-02-12
CN201910110784.6A CN110033382B (zh) 2019-02-12 2019-02-12 一种保险业务的处理方法、装置及设备

Publications (1)

Publication Number Publication Date
WO2020164332A1 true WO2020164332A1 (zh) 2020-08-20

Family

ID=67235651

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/070042 Ceased WO2020164332A1 (zh) 2019-02-12 2020-01-02 一种保险业务的处理方法、装置及设备

Country Status (2)

Country Link
CN (1) CN110033382B (zh)
WO (1) WO2020164332A1 (zh)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110033382B (zh) * 2019-02-12 2020-09-04 阿里巴巴集团控股有限公司 一种保险业务的处理方法、装置及设备
CN110879855A (zh) * 2019-10-28 2020-03-13 世纪保众(北京)网络科技有限公司 一种保险项目的对比方法、装置、设备和存储介质
CN112288585B (zh) * 2020-11-20 2024-05-28 中国人寿保险股份有限公司 保险业务精算数据处理方法、装置及电子设备
CN112541831B (zh) * 2020-12-16 2024-06-28 中国人寿保险股份有限公司 一种医疗保险的风险识别方法、装置、介质及电子设备
US11615493B2 (en) * 2020-12-16 2023-03-28 International Business Machines Corporation Contextual comparison of semantics in conditions of different policies
CN114691837B (zh) * 2022-06-01 2022-09-20 深圳市奕亮科技有限公司 一种基于大数据的保险业务数据处理方法和处理系统
CN116128292A (zh) * 2022-12-29 2023-05-16 南方电网数字平台科技(广东)有限公司 一种基于工作内容的作业风险等级的智能研判方法及装置
CN116467439A (zh) * 2023-02-17 2023-07-21 保腾网络科技有限公司 保险产品关键字的提取方法、装置、存储介质及服务器

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105808648A (zh) * 2016-02-25 2016-07-27 焦点科技股份有限公司 一种基于r语言程序的个性化推荐方法
CN107330719A (zh) * 2017-06-09 2017-11-07 上海新概念保险经纪有限公司 一种保险产品推荐方法及系统
CN108629697A (zh) * 2018-03-30 2018-10-09 平安科技（深圳）有限公司 保险产品配置方法、装置、计算机设备及存储介质
US20190018658A1 (en) * 2014-03-10 2019-01-17 Anthony Martello System and method for manipulating and displaying data
CN110033382A (zh) * 2019-02-12 2019-07-19 阿里巴巴集团控股有限公司 一种保险业务的处理方法、装置及设备

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002108865A (ja) * 2000-09-29 2002-04-12 Hitachi Kokusai Electric Inc データ検索システム
CN104463630B (zh) * 2014-12-11 2015-08-26 新一站保险代理有限公司 一种基于网购保险产品特性的产品推荐方法及系统
CN107704512B (zh) * 2017-08-31 2021-08-24 平安科技（深圳）有限公司 基于社交数据的金融产品推荐方法、电子装置及介质
CN109035032A (zh) * 2018-06-11 2018-12-18 中国平安人寿保险股份有限公司 数据结构化处理方法、装置、计算机设备及存储介质
CN108959618B (zh) * 2018-07-18 2021-12-24 北京欣欣苹果网络科技有限公司 互联网信息收集及处理方法和装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190018658A1 (en) * 2014-03-10 2019-01-17 Anthony Martello System and method for manipulating and displaying data
CN105808648A (zh) * 2016-02-25 2016-07-27 焦点科技股份有限公司 一种基于r语言程序的个性化推荐方法
CN107330719A (zh) * 2017-06-09 2017-11-07 上海新概念保险经纪有限公司 一种保险产品推荐方法及系统
CN108629697A (zh) * 2018-03-30 2018-10-09 平安科技（深圳）有限公司 保险产品配置方法、装置、计算机设备及存储介质
CN110033382A (zh) * 2019-02-12 2019-07-19 阿里巴巴集团控股有限公司 一种保险业务的处理方法、装置及设备

Also Published As

Publication number Publication date
CN110033382A (zh) 2019-07-19
CN110033382B (zh) 2020-09-04

Similar Documents

Publication Publication Date Title
CN110033382B (zh) 一种保险业务的处理方法、装置及设备
Lin et al. A deep learning architecture for semantic address matching
CN110321482B (zh) 一种信息的推荐方法、装置及设备
CN109858970B (zh) 一种用户行为预测方法、装置及存储介质
TWI612488B (zh) 用於預測商品的市場需求的計算機裝置與方法
US9613024B1 (en) System and methods for creating datasets representing words and objects
Zhang et al. Combining sentiment analysis with a fuzzy kano model for product aspect preference recommendation
US20190286978A1 (en) Using natural language processing and deep learning for mapping any schema data to a hierarchical standard data model (xdm)
CN115309864B (zh) 评论文本的情感智能分类方法、装置、电子设备及介质
CN110032728A (zh) 疾病名称标准化的转换方法和装置
Angadi et al. Multimodal sentiment analysis using reliefF feature selection and random forest classifier
CN110020427A (zh) 策略确定方法和装置
US20250200274A1 (en) Apparatus and method for generating annotations for electronic records
CN111061876A (zh) 事件舆情数据分析方法及装置
Vaissnave et al. Modeling of automated glowworm swarm optimization based deep learning model for legal text summarization
Kumari et al. Importance of data standardization methods on stock indices prediction accuracy
Zhang et al. Multi-head attention model for aspect level sentiment analysis
Agrawal et al. RETRACTED: A hybrid multi-source data fusion for word, sentence, aspect, and document-level sentiment analysis on real-time databases
US11822609B2 (en) Prediction of future prominence attributes in data set
CN116167371A (zh) 产品的推荐方法及装置、处理器和电子设备
Stevens et al. Predicting real estate price using text mining
Shukla et al. RETRACTED: A comprehensive survey on sentiment analysis: Challenges and future insights
US20240005231A1 (en) Methods and systems for holistic medical student and medical residency matching
Zhu et al. A named entity recognition model based on ensemble learning
Yuniari et al. Environment Sentiment Analysis of Bali Coffee Shop Visitors Using Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-trained Transformer 2 (GPT2) Model

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20756541

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20756541

Country of ref document: EP

Kind code of ref document: A1