CN111353793A - CRM (customer relationship management) service recommendation method and device - Google Patents
CRM (customer relationship management) service recommendation method and device Download PDFInfo
- Publication number
- CN111353793A CN111353793A CN201811563505.3A CN201811563505A CN111353793A CN 111353793 A CN111353793 A CN 111353793A CN 201811563505 A CN201811563505 A CN 201811563505A CN 111353793 A CN111353793 A CN 111353793A
- Authority
- CN
- China
- Prior art keywords
- service
- crm
- cat
- services
- neighbor
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 241000282326 Felis catus Species 0.000 claims abstract description 83
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 71
- 238000001914 filtration Methods 0.000 claims abstract description 48
- 238000012545 processing Methods 0.000 claims description 17
- 230000006870 function Effects 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 8
- 238000004891 communication Methods 0.000 claims description 7
- 239000011159 matrix material Substances 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 5
- 230000000739 chaotic effect Effects 0.000 claims description 2
- 238000005457 optimization Methods 0.000 abstract description 9
- 230000008901 benefit Effects 0.000 description 11
- 238000004364 calculation method Methods 0.000 description 9
- 238000011156 evaluation Methods 0.000 description 6
- 238000004590 computer program Methods 0.000 description 5
- RWSOTUBLDIXVET-UHFFFAOYSA-N Dihydrogen sulfide Chemical compound S RWSOTUBLDIXVET-UHFFFAOYSA-N 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 230000001419 dependent effect Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000000977 initiatory effect Effects 0.000 description 2
- 239000008833 migu Substances 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Accounting & Taxation (AREA)
- Physics & Mathematics (AREA)
- Entrepreneurship & Innovation (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention provides a CRM service recommendation method and a CRM service recommendation device, wherein the method comprises the following steps: acquiring CRM (customer relationship management) service currently handled by a user as a reference service, and recording a characteristic value of the reference service; according to the characteristic values of the reference business and the characteristic values of other CRM businesses, calculating the similarity between the other CRM businesses and the reference business through a collaborative filtering algorithm, selecting a plurality of CRM businesses as target businesses according to the sequence of the similarity from large to small, and displaying the target businesses to the user; and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO. The embodiment of the invention has high intelligent degree, the collaborative filtering algorithm introduces the cat swarm algorithm to carry out core parameter optimization, the operation precision is improved, the complexity of the service recommendation rule is high, the data is accurately matched with the CRM collaborative filtering algorithm after being normalized, and the method has excellent adaptability.
Description
Technical Field
The embodiment of the invention relates to the technical field of business recommendation, in particular to a CRM business recommendation method and device.
Background
The traditional CRM system is used as a core support system for customer service and marketing, mainly has the main functions of carrying out corresponding service query and handling according to the service requirements of customers, does not support methods such as intelligent recommendation and intelligent navigation, and the CRM system finishes handling services according to the service requirements proposed by the customers and pertinently handles the services. The requirement on business personnel is high, business logic needs to be understood very deeply, the functions and layout of each business interface are very clear, and the business interface is quickly positioned and processed.
The existing big data platform mainly aims at making customer portraits and marketing activities according to customer service data and situation data, potential association relations among services are not mined, management and recommendation of potential characteristic relations of the services are lacked, meanwhile, strong data mining capacity of big data is not associated with a CRM system, and support work of intelligent service recommendation is the key point of future development.
In summary, the prior art has the following disadvantages:
(1) with the increasing complexity of the business, the requirement on the business proficiency of business personnel is also improved, the customer appeal is varied, the business personnel must know which interface can handle which business to quickly respond to the appeal of the service marketing, the workload of front-line personnel is increased invisibly, the problem of mishandling and missed handling is easy to occur, and the customer satisfaction is further influenced.
(2) With the rapid development of the internet and the deepening of the electronic informatization degree, the forms of the binary structure business of the voice and the flow of the client are rapidly changed, the marketing mode of the client service is increasingly diversified, the traditional question-answering business handling mode cannot meet the requirements of the client, the traditional service mode is changed, the requirements of the user on intellectualization and convenience are met, and the mode becomes a choice which must be made by an operator. The method has the advantages that the efficiency is improved, the cost is reduced, and the service and business recommendation capability of a front-line business worker of the mobile company facing the mobile internet is improved by creating the service through the media capability of the internet, so that the method can follow the new era step of the transformation development of customer service.
Disclosure of Invention
Embodiments of the present invention provide a CRM service recommendation method and apparatus that overcome the above problems or at least partially solve the above problems.
In a first aspect, an embodiment of the present invention provides a CRM service recommendation method, including:
acquiring CRM (customer relationship management) business currently handled by a user as a reference business, and carrying out normalization processing on a plurality of characteristics of the reference business to obtain characteristic values of the plurality of characteristics to form a project configuration file;
calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the project configuration files of the reference service and the project configuration files of the other CRM services;
selecting a plurality of CRM services as target services according to the sequence of similarity from large to small, and recommending the target services to the user;
and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO.
In a second aspect, an embodiment of the present invention provides a CRM service recommendation apparatus, including:
the system comprises a preprocessing module, a data processing module and a data processing module, wherein the preprocessing module is used for acquiring a CRM (customer relationship management) service currently handled by a user as a reference service, normalizing a plurality of characteristics of the reference service to acquire characteristic values of the plurality of characteristics and form a project configuration file;
the collaborative filtering module is used for calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the project configuration file of the reference service and the project configuration files of the other CRM services;
the recommendation module is used for selecting a plurality of CRM services as target services according to the sequence of similarity from large to small and recommending the target services to the user;
and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO.
In a third aspect, an embodiment of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the method provided in the first aspect when executing the program.
In a fourth aspect, an embodiment of the present invention provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the method as provided in the first aspect.
The CRM service recommendation method and device provided by the embodiment of the invention have the advantages that firstly, the intelligent degree is high, the association relationship between preferential, service and platform services, which is regular/has universal applicability and periodicity, is found based on the potential weak association service relationship, scientific evaluation is carried out, the usability and the correctness of the rules are determined, after business personnel successfully transact source services through CRM, the dependent services with the weak association relationship with the services are immediately recommended, secondly, a collaborative filtering algorithm is introduced into a cat swarm algorithm for optimizing core parameters, the operation precision is improved, finally, the complexity of the service recommendation rules is high, data needs to be accurately matched with the CRM collaborative filtering algorithm after normalization, and the CRM service recommendation method and device have excellent adaptability.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a CRM service recommendation method provided in an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a CRM service recommendation device according to an embodiment of the present invention;
fig. 3 is a schematic physical structure diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Generally, a group of services, i.e. a collection of several independent benefits, services and platform services, is common to services transacted by a customer in a business hall, the internet and other channels, and then the initiation time of a service work order (standing book) of the several benefits, services and platform services should be in a very short range, ideally, the interval is only several seconds to 1 minute, and the maximum time is not more than 10 minutes. The service initiation in a short time can be regarded as a group of associated services for realizing the customer requirements, and the service at an interval of more than 10 minutes has a weak association relationship, so that the combined requirements of the customers cannot be reflected.
In order to overcome the above problems in the prior art, an embodiment of the present invention provides a CRM service recommendation method, which includes: based on the potential weak association business relationship in the scene, finding out association relations with general applicability and periodicity among preferential, service and platform businesses, carrying out scientific evaluation, and determining the availability and correctness of the rules. After a customer transacts a business, the related business obtained by collaborative filtering calculation of big data according to the proposal is immediately pushed, and the business transaction success rate and the intelligent degree of a CRM system are improved. And after the previous business is transacted, the next N businesses are actively recommended and intelligently guided through the rules calculated by a big data collaborative filtering algorithm.
In addition, in order to evaluate the rule correctness more accurately, a CSO (cat swarm algorithm) collaborative filtering algorithm is introduced to carry out business rule combing and scoring, characteristic rules with the strongest business association relationship are mined, and a whole set of CRM business recommendation system is formed by combining the inherent strong association relationship (account deduction, default opening, dependency relationship rules and the like) of CRM.
Fig. 1 is a schematic flow chart of a CRM service recommendation method provided in an embodiment of the present invention, as shown in fig. 1, including:
s101, CRM business currently handled by a user is obtained and used as reference business, a plurality of characteristics of the reference business are normalized, characteristic values of the characteristics are obtained, and a project configuration file is formed.
Specifically, the CRM service in the embodiment of the present invention may include a platform service, for example, a sp _ code service and a biz _ code service, and may also be a preference for a service, for example, a quarterly package, a half-year package, a year package, and the like. The characteristics of the service include basic information of the user (e.g., age, sex, occupation, place of residence), service information, user rating, and the like. The project profile is a comprehensive reflection of the project characteristics. The configuration file of the project is the basis for calculating the distance and the similarity between the projects, and after the configuration file is established, the nearest neighbor candidate set needs to be obtained, and then the similarity relation between the projects is calculated.
It should be appreciated that CRM services are highly complex, and the service categories include: the number of various service categories is accumulated to 10 ten thousand, so that the potential association relation before various services is mined, the most suitable service is recommended to the client, and the method is a great challenge. The method is characterized in that a business potential characteristic relation is rapidly calculated through a CSO collaborative filtering algorithm, and various metadata are integrated into a mode suitable for algorithm rule processing through data normalization processing and data combing. In addition, the normalized standard data source required by the collaborative filtering algorithm needs to process data of CRM business ledger tables to input the model, and the matching of the method and the original model is very important.
S102, calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the project configuration file of the reference service and the project configuration files of the other CRM services, and optimizing the average absolute deviation (MAE) index of the collaborative filtering algorithm through a cat swarm algorithm (CSO).
It should be noted that the collaborative filtering is to simply recommend information interested by a user by using the preferences of a group with a certain interest engagement and common experience, and an individual gives a considerable response (such as scoring) to the information through a collaborative mechanism and records the response to achieve the purpose of filtering, so as to help others to filter the information. In the embodiment of the invention, the similarity is calculated through the project configuration file of the CRM service. After the core parameters of the collaborative filtering algorithm are optimized through a cat swarm algorithm (CSO), the processing speed and the combing precision are greatly improved.
S103, selecting a plurality of CRM services as target services according to the sequence of similarity from large to small, and recommending the target services to the user;
the embodiment of the invention has the following beneficial effects:
1) the intelligent degree is high, based on the potential weak association business relationship, the association relationship with the preferential, service and platform business regularity/universal applicability and periodicity is found, scientific evaluation is carried out, the usability and the correctness of the rule are determined, and after business personnel transact the successful source business in the CRM, the dependent business with the weak association relationship with the business is recommended immediately.
2) The collaborative filtering algorithm introduces a cat swarm algorithm to optimize core parameters, and the operation precision is improved.
3) The service recommendation rule is high in complexity, the data needs to be accurately matched with the CRM collaborative filtering algorithm after being normalized, and the service recommendation rule has excellent adaptability.
On the basis of the above embodiments, according to the feature value of the reference service and the feature values of the other CRM services, calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm, specifically:
s201, acquiring a neighbor candidate set of the reference service from the other CRM services, wherein the neighbor candidate set comprises a plurality of neighbor services which have common user scores with the reference service and are most closely related.
It should be noted that the selection of the nearest neighbor is a very important step in the personalized recommendation system, and the recommendation quality of the collaborative filtering system depends on the effect of the algorithm to find the neighbor.
S202, setting initial weight values of all characteristic values in the project configuration file to establish and construct a weighted cosine similarity function between the reference service and the neighbor service.
It should be noted that, after the selection of the neighbor candidate set is completed, the weighted cosine similarity calculation is performed subsequently. First, the similarity between the reference traffic and the neighbor traffic is measured. Commonly used indicators are cosine similarity or modified cosine similarity between profiles and pearson correlation coefficients. The embodiment of the invention utilizes the weighted cosine similarity between configuration files, and adopts a CSO algorithm to optimize in a training set when determining the nearest weight value.
S203, optimizing the weight value in the training set through a cat swarm algorithm CSO, obtaining a nearest weight value when the average absolute deviation MAE reaches the minimum, and calculating a nearest similarity solution of the reference service and the neighbor service according to the nearest weight value.
It should be noted that, since the project profiles contain many attributes, many of which are sparse or incomplete, in order to overcome these problems, the embodiments of the present invention use stochastic and heuristic models to speed up and improve the quality of user profile matching, and by applying the cat swarm algorithm CSO to the matching of the project profiles, the solution to such problems is improved.
In the prediction process, a user which does not score the current service and scores the neighbor project is selected, and the score of the current project is predicted by taking the nearest similarity value between the current project configuration file and the neighbor project configuration file as a weight according to a previously established nearest neighbor candidate set.
On the basis of the foregoing embodiments, obtaining the neighbor candidate set of the reference service from the other CRM services specifically includes:
and extracting a plurality of other CRM services of the users having the most common scores with the reference service by using a maximum communication method and utilizing matrix operation as a neighbor candidate set.
It should be noted that, when determining the data set selected by the neighbor, most scholars at home and abroad perform traversal calculation and selection on the whole database, that is, search between all configuration files in the database and the current project configuration file, and select the neighbor with the highest similarity as the nearest neighbor. The traversal selection has the following problems that firstly, data sparsity exists, and some projects rarely share common scores with the current project, so that the similarity of the two projects cannot be calculated; secondly, when the project and user data are very huge, it is not practical to perform full traversal search. The method and the device are selected according to the principle of similarity of user behaviors, and the more common user evaluation is performed on the same items, the better similarity of the items on the user behaviors is shown, so that neighbors with common user evaluation with the current items are arranged in a descending order, and a plurality of items are selected as candidate sets selected by the nearest neighbors.
When using the collaborative filtering algorithm, it is a common practice to directly adopt the scoring information of the users as a measure of the similarity between items or users. However, in the similarity calculation, the similarity between two projects or users is not measured only by scores, and is influenced by personal characteristics of the users in addition to the factors of the projects. Therefore, when calculating the similarity, it is necessary to comprehensively consider the information of the user and the information of the item. Also, gender may be weighted higher relative to occupation since each factor may affect the score to a different degree, for example, for a user, perhaps gender factors may affect the score more than occupation. In the embodiment of the invention, a weight (representing the weight of the characteristic y) is added in front of each influence characteristic, and in the experimental process, the weight is optimized through a CSO algorithm to reach an optimal solution.
Let w (a) denote a potential solution to the current item profile feature weight value, a being denoted as a set of weights. Wherein the weights of the features f are represented, and each feature set comprises K weight values, which represent the positions of the particles in a K-dimensional space. As the particles move in this K-dimensional space, these feature values will also continually adjust until a set of weights is found that recently describe the active user preferences.
On the basis of the above embodiments, the weighted cosine similarity function between the reference service and the neighbor service specifically includes:
wherein, A represents the project configuration file of the reference service, j represents the project configuration file of the neighbor service, W is K × K diagonal matrix, K is the number of weight values, each value of the diagonal matrix identifies each characteristic value but is present, and n represents the total number of the reference service and the neighbor service directly used for jointly scoring the users.
In the experiment, the weight of the characteristic value of each item is assumed to be stable, so that only one optimal solution of the weight value exists in the current item in the optimization process, namely, only one single adaptive function exists. This fitness function is defined as the average of the scoring prediction errors between two items. Thus, the goal of the optimization is to calculate the W for each current item (one fitness function for each W).
In the embodiment of the present invention, the expression of the mean absolute deviation MAE is:
wherein n is the number of users with common scores between the reference service A and the neighbor service j, meanARepresenting the average value of the scores of the reference service A, similar (A, j) is the similarity between the reference service A and the neighbor service j, vote (j, i) represents the score of the i user to the neighbor service j, meaniRepresents the average of all the scores of the i-subscriber, and vote (a, i) represents the actual score of the i-subscriber for the reference service a.
In Cat Swarm algorithm (Cat Swarm Optimization), cats are the feasible solutions to the problem to be optimized. The cat swarm algorithm divides the behavior of the cat into two modes, wherein the mode is called as a search mode when the cat is in a lazy state and a surrounding state; the other is the state when tracking a dynamic object, called tracking mode. In the cat swarm algorithm, a part of cats execute a search mode, the rest execute a tracking mode, the two modes interact through a Mix-ture Ratio (MR), the MR represents the proportion of the number of cats in the tracking mode in the whole cat swarm, and the MR is a smaller value during calculation. The optimization problem is solved by using a cat swarm algorithm, and the number of individuals participating in optimization calculation, namely the number of cats, needs to be determined. Attributes for each cat (including its own position consisting of M dimensions), velocity in each dimension, adaptation to the reference function, and identification value indicating whether the cat is in search mode or tracking mode. After the cat has performed the search mode and the tracking mode, its fitness is calculated according to the fitness function and the best solution in the current population is retained. Then, the cat group is randomly divided into cats in a searching part and cats in a tracking part according to the combination rate, and iterative calculation is carried out by the method until the preset iterative times are reached.
On the basis of the above embodiments, as an optional embodiment, the weight values are optimized in the training set by the cat swarm algorithm CSO, specifically:
s301, adjusting the mode allocation proportion MR of the cats in the cat swarm algorithm through the dynamically changing linear factor;
according to experience, the overall search range can be enlarged by increasing the proportion of searched cats in the early stage of the algorithm, the proportion of tracked cats in the population is increased by adjusting the MR, the local optimization precision of the cat population can be improved, and the rapid convergence in the later stage of the algorithm can be ensured. The expression for determining cat pool search pattern assignment based on linear Mixture Ratio (MRL) is as follows:
wherein, MRmaxAnd MRminRespectively, an upper limit value and a lower limit value of the Mixing Ratio (MR), t is the current iteration number, ITmaxIs the maximum number of iterations. MRLThe mode allocation ratio of the cat can be adjusted by a dynamically changing linear factor, thereby realizing the optimization of the MR.
S302, defining the number of the cat groups according to the number of the characteristic values, simultaneously defining SMP (memory pool search) and CDC (dimension change count), and randomly initializing the positions and the speeds of the cat groups;
s303, calculating the position information of each cat, and defining the cat with the best position information as XgbestUnder the current iteration, carrying out mode distribution according to the mode distribution proportion MR, updating the current position of the cat in the search mode, and carrying out cat group demarcation on the cat in the tracking mode;
s304, evaluating the updated position of the cat and updating the global optimal solution Xgbest(ii) a Updating the position of the population based on the chaotic sequence;
s305, judging whether the maximum iteration times is reached, and if the maximum iteration times is reached, outputting the global optimal solution X of the last iterationgbest。
Updating the current position of the cat in the search mode, specifically:
generating copies of K population individuals based on SMP to form an initial population xi；
One cat in K populations was selected, the current position was maintained, and the remaining copies (K-1) were mutated under the influence of orthogonal variation. x is the number ofiRepresenting the initial population, the updated individual resulting from the variation is xi mThe expression is as follows:
xi m＝xi+(SRD*xi*N(0,1))
wherein N (0,1) is a standard normally distributed random number with zero mean and standard deviation, N (0,1) is associated with the variable domains SRD and xiAs a sudden change in the current dimension, i.e. SRD xi*N(0,1)；
Fitness values were evaluated for all replicates, including the original position in the K populations and the variant position of replicate (K-1). Obtaining the optimal individual in the K copies according to the optimal adaptive value, and taking the optimal individual as an optimal solution;
replacing the current position of the cat with the optimal solution, and resetting the varied individual position as the initial group position x if the individual fitness after normal variation is betteri；
The cat group definition of the cats in the tracking mode specifically comprises the following steps:
the solution space is D-dimensional, and the velocity v of the ith cat is definedi,dIs v isi,d＝(vi,1,vi,2,vi,3...,vi,D) Position xi,dIs xi,d＝(xi,1,xi,2,xi,3...,xi,D) Global optimum position x for cat groupg,dIs denoted by xg,d＝(xg,1,xg,2,xg,3...,xg,D)；
Calculating the motion vector of a new individual in the cat group, and using the motion vector of the ith cat in each dimension by using a formula vi,d＝w*vi,d+c*r*(xg,d-xi,d) Calculating;
according to the formula xi,d n＝xi,d+vi,dCalculating the new position of the ith cat;
if the new position of the ith cat in any dimension is outside the search space range, the velocity of the current dimension is set to a boundary value and the process is restarted with a reverse search.
It should be noted that the score of the existing item in the experimental set is predicted according to the score received by the optimal neighbor service and the nearest similarity solution obtained from the feature value after the optimization between the two services. In the prediction process, a user which does not score the current service and scores the neighbor project is selected, and the score of the current project is predicted by taking the nearest similarity value between the current project configuration file and the neighbor project configuration file as a weight according to a previously established nearest neighbor candidate set.
After the user-item scoring matrix is filled, items with higher user prediction scores are selected to recommend the current user. And recommending the preference with the strongest association relation to a CRM foreground interface to complete service recommendation.
Specific application processes of the CRM service recommendation method provided by the embodiment of the present invention are described below for different application scenarios.
(1) Recommending offers after handling platform business services
Platform services are handled by a group of sp _ code and biz _ code services, the services charge fees according to monthly rents, and customers have the requirements of handling quarter packages, half-year packages and year packages after handling the platform services, so that associated preferential benefits are cooperatively recommended, and the service is an important service scene.
The data about 15000 days after 7 months and 10 days is calculated, and a platform service id, a platform service name, a preference id and a preference name are listed, the number of the preference is handled immediately after the platform service is handled, the number of the preference is handled within 10 minutes after the platform service is handled, and the number of the preference is handled at the same time after the service is handled. Therefore, a preferential recommendation rule of the platform service is formed, and after the platform service is transacted, the associated preferential is actively recommended to the client.
And finally, processing the data by using a collaborative filtering algorithm through a collaborative filtering scoring program developed by python, and circularly outputting 3 preferential IDs which can be handled recommended by each platform service ID so as to control the recommended service range.
And finally, processing the data by using a collaborative filtering algorithm through a collaborative filtering scoring program developed by python, and circularly outputting 3 preferential IDs which can be handled recommended by each platform service ID so as to control the recommended service range.
Note: in order to prove the failure relevance of the business recommendation rule, two indexes are further defined: the number of the special offers is handled immediately after the platform service is handled, and the number of the special offers is handled within 10 minutes after the platform service is handled. Ideally, the associated service transaction intervals are very short and theoretically not more than 1 minute, if the number of services handled within 10 minutes after the platform service is handled is greatly different from the number of services handled immediately, the set of rules are proved to be possibly inaccurate, and if the coincidence degree is high, the set of rules have very good timeliness, the relationship between the source service and the recommended service is strong, and the recommendation score is high.
The output result shows that the number of the preferential offers are handled immediately after the platform service is handled, the number of the preferential offers are handled within 10 minutes after the platform service is handled, the contact ratio is more than 95%, and the rule output by the recommendation algorithm is excellent and feasible in the aspect of time efficiency.
(2) Recommending associated offers after handling offers
A mainstream interface for transacting the benefits by the CRM system is 'product change', after transacting the benefit packets, a client may transact an extension packet, an associated benefit packet or a service packet of the same type with a potential intention, at the moment, a benefit recommendation rule corresponding to the benefits is calculated through a big data collaborative filtering algorithm, and the method is also helpful for service-guided intelligent recommendation and the method is the same as (1).
And (3) calculating a rule: immediately or within 10 minutes after the discovery is processed, the next related discovery is related and inquired, and the preferential id, the preferential name, the recommended preferential id, the recommended preferential name, the immediate processing times and the processing times within 10 minutes are also output.
The calculated source data amount is about 30980, which is a preferential treatment record for one day.
(3) Business rule checking
After the platform service recommendation offer and the preferential recommendation offer calculation are completed, the occurrence probability of the rule is also considered, the rules with low probability, low score and small service volume are screened out, and only the most common service recommendation rule is reserved.
The accessory gives service recommendation universality data and service recommendation universality data, the data are collected according to the number of initial service id recommendation rules and then sorted in a descending order, the recommended preferential ids are collected according to the recommendation rule times and then sorted in a descending order, and an output rule list shows that: the service recommendation triggering times of mainstream platform services such as migu, mobile game, 139 mailbox and the like are very many, and the rule recommendation triggering with less service volume is less; the corresponding preferential recommendation rules, such as migu members, data traffic packets, 139 mailboxes, music members, etc., are triggered more times. Therefore, according to the universality of the rules, the rules with few service recommendation scenes and small recommendation probability are screened, the service recommendation performance of the CRM system can be improved, and the system risk is reduced.
(4) Implementing service recommendation in CRM foreground
4.1, a business prompt box of the CRM system is provided with a guide button for recommending the business.
4.2, after the business personnel click the 'business recommendation button', the current customer acceptance number is automatically transmitted into the CRM frame and is brought into a menu 'product change' recommended by the target business, and automatic inquiry of customer information is executed.
And 4.3, actively inquiring and displaying the target discount package on a product change interface according to a service recommendation rule table obtained by the methods of calculation of a collaborative filtering recommendation algorithm, service rule combing, and the like, and guiding business personnel to handle recommended discount preferentially.
4.4, but the checking is not forced or acquiescent, but the checking is a display and reminding function, and business personnel can still inquire other businesses independently.
Fig. 2 is a schematic structural diagram of a CRM service recommendation device according to an embodiment of the present invention, and as shown in fig. 2, the CRM service recommendation device includes: a preprocessing module 201, a collaborative filtering module 202, and a recommendation module 203, wherein:
the system comprises a preprocessing module 201, a data processing module and a data processing module, wherein the preprocessing module 201 is used for acquiring a CRM (customer relationship management) service currently handled by a user as a reference service, normalizing a plurality of characteristics of the reference service to acquire characteristic values of the plurality of characteristics and form a project configuration file;
a collaborative filtering module 202, configured to calculate, according to the project configuration file of the reference service and the project configuration files of other CRM services, a similarity between the other CRM services and the reference service through a collaborative filtering algorithm; and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO.
The recommending module 203 is used for selecting a plurality of CRM services as target services according to the sequence of similarity from large to small, and recommending the target services to the user;
the CRM service recommendation apparatus provided in the embodiment of the present invention specifically executes the flows of the CRM service recommendation method embodiments, and please refer to the contents of the CRM service recommendation method embodiments in detail, which is not described herein again. The CRM service recommendation device provided by the embodiment of the invention has the following beneficial effects:
1) the intelligent degree is high, based on the potential weak association business relationship, the association relationship with the preferential, service and platform business regularity/universal applicability and periodicity is found, scientific evaluation is carried out, the usability and the correctness of the rule are determined, and after business personnel transact the successful source business in the CRM, the dependent business with the weak association relationship with the business is recommended immediately.
2) The collaborative filtering algorithm introduces a cat swarm algorithm to optimize core parameters, and the operation precision is improved.
3) The service recommendation rule is high in complexity, the data needs to be accurately matched with the CRM collaborative filtering algorithm after being normalized, and the service recommendation rule has excellent adaptability.
Fig. 3 is a schematic entity structure diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 3, the electronic device may include: a processor (processor)310, a communication Interface (communication Interface)320, a memory (memory)330 and a communication bus 340, wherein the processor 310, the communication Interface 320 and the memory 330 communicate with each other via the communication bus 340. The processor 310 may invoke a computer program stored on the memory 330 and executable on the processor 310 to perform the CRM service recommendation methods provided by the various embodiments described above, including, for example: acquiring CRM (customer relationship management) business currently handled by a user as a reference business, and carrying out normalization processing on a plurality of characteristics of the reference business to obtain characteristic values of the plurality of characteristics to form a project configuration file; calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the project configuration files of the reference service and the project configuration files of the other CRM services; selecting a plurality of CRM services as target services according to the sequence of similarity from large to small, and recommending the target services to the user; and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO.
In addition, the logic instructions in the memory 330 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. Based on such understanding, the technical solutions of the embodiments of the present invention may be essentially implemented or make a contribution to the prior art, or may be implemented in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
An embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented by a processor to execute the CRM service recommendation method provided in the foregoing embodiments, for example, the method includes: acquiring CRM (customer relationship management) business currently handled by a user as a reference business, and carrying out normalization processing on a plurality of characteristics of the reference business to obtain characteristic values of the plurality of characteristics to form a project configuration file; calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the project configuration files of the reference service and the project configuration files of the other CRM services; selecting a plurality of CRM services as target services according to the sequence of similarity from large to small, and recommending the target services to the user; and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A CRM service recommendation method, comprising:
acquiring CRM (customer relationship management) business currently handled by a user as a reference business, and carrying out normalization processing on a plurality of characteristics of the reference business to obtain characteristic values of the plurality of characteristics to form a project configuration file;
calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the project configuration files of the reference service and the project configuration files of the other CRM services;
selecting a plurality of CRM services as target services according to the sequence of similarity from large to small, and recommending the target services to the user;
and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO.
2. The method according to claim 1, wherein the calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the eigenvalue of the reference service and the eigenvalues of the other CRM services specifically comprises:
acquiring a neighbor candidate set of the reference service from the other CRM services, wherein the neighbor candidate set comprises a plurality of neighbor services which have common user scores with the reference service and are most closely related;
setting initial weight values of all characteristic values in a project configuration file to establish a weighted cosine similarity function between the reference service and the neighbor service;
and optimizing the weight value in the training set by a cat swarm algorithm CSO, obtaining a nearest weight value when the average absolute deviation MAE reaches the minimum, and calculating a nearest similarity solution of the reference service and the neighbor service according to the nearest weight value.
3. The method according to claim 2, wherein the obtaining of the neighbor candidate set of the reference service from the other CRM services specifically includes:
and extracting a plurality of other CRM services of the users having the most common scores with the reference service by using a maximum communication method and utilizing matrix operation as a neighbor candidate set.
4. The method according to claim 2, wherein the weighted cosine similarity function between the reference service and the neighbor service is specifically:
wherein, A represents the project configuration file of the reference service, j represents the project configuration file of the neighbor service, W is K × K diagonal matrix, K is the number of weight values, each value of the diagonal matrix identifies each characteristic value but is present, and n represents the total number of the reference service and the neighbor service directly used for jointly scoring the users.
5. The method of claim 2, wherein the mean absolute deviation MAE is expressed as:
wherein n is the number of users with common scores between the reference service A and the neighbor service j, meanARepresenting the average value of the scores of the reference service A, similar (A, j) is the similarity between the reference service A and the neighbor service j, vote (j, i) represents the score of the i user to the neighbor service j, meaniRepresents the average of all the scores of the i-subscriber, and vote (a, i) represents the actual score of the i-subscriber for the reference service a.
6. The method according to claim 2, wherein the weight values are optimized in the training set by the cat swarm algorithm CSO, specifically:
adjusting the mode allocation proportion MR of the cats in the cat swarm algorithm through the dynamically changing linear factor; defining the number of the cat groups according to the number of the characteristic values, simultaneously defining SMP (symmetric multi-processor) and CDC (performance data center) parameters, and randomly initializing the positions and the speeds of the cat groups;
calculating the position information of each cat, and defining the cat with the best position information as XgbestUnder the current iteration, carrying out mode distribution according to the mode distribution proportion MR, updating the current position of the cat in the search mode, and carrying out cat group demarcation on the cat in the tracking mode;
evaluating the updated cat position and updating the global optimal solution Xgbest(ii) a Updating the position of the population based on the chaotic sequence;
judging whether the maximum iteration times is reached, if so, outputting the global optimal solution X of the last iterationgbest。
7. The method according to claim 6, wherein the updating of the current location of the cat in search mode is performed by:
generating copies of K population individuals based on SMP to form an initial population xi；
One cat in the K populations was selected,keeping the current position unchanged, and mutating the rest copies (K-1) under the influence of orthogonal variation; x is the number ofiRepresenting the initial population, the updated individual resulting from the variation is xi mThe expression is as follows:
xi m＝xi+(SRD*xi*N(0,1))
wherein N (0,1) is a standard normally distributed random number with zero mean and standard deviation, N (0,1) is associated with the variable domains SRD and xiAs a sudden change in the current dimension, i.e. SRD xi*N(0,1)；
Evaluating fitness values of all the copies, including original positions in K copies of the population and variation positions of the copies (K-1), obtaining the best individual of the K copies according to the best fitness value, and taking the best individual as the best solution;
replacing the current position of the cat with the optimal solution, and resetting the varied individual position as the initial group position x if the individual fitness after normal variation is betteri；
The cat group definition of the cats in the tracking mode specifically comprises the following steps:
the solution space is D-dimensional, and the velocity v of the ith cat is definedi,dIs v isi,d＝(vi,1,vi,2,vi,3...,vi,D) Position xi,dIs xi,d＝(xi,1,xi,2,xi,3...,xi,D) Global optimum position x for cat groupg,dIs denoted by xg,d＝(xg,1,xg,2,xg,3...,xg,D)；
Using the motion vector of the ith cat in each dimension by the formula vi,d＝w*vi,d+c*r*(xg,d-xi,d) Calculating;
according to the formula xi,d n＝xi,d+vi,dCalculating the new position of the ith cat;
if the new position of the ith cat in any dimension exceeds the range of the search space, setting the speed of the current dimension as a boundary value, and restarting the process by reverse search;
fitness values were evaluated for all cats.
8. A CRM service recommendation apparatus, comprising:
the system comprises a preprocessing module, a data processing module and a data processing module, wherein the preprocessing module is used for acquiring a CRM (customer relationship management) service currently handled by a user as a reference service, normalizing a plurality of characteristics of the reference service to acquire characteristic values of the plurality of characteristics and form a project configuration file;
the collaborative filtering module is used for calculating the similarity between the other CRM services and the reference service through a collaborative filtering algorithm according to the project configuration file of the reference service and the project configuration files of the other CRM services;
the recommendation module is used for selecting a plurality of CRM services as target services according to the sequence of similarity from large to small and recommending the target services to the user;
and optimizing the average absolute deviation MAE index of the collaborative filtering algorithm through a cat swarm algorithm CSO.
9. An electronic device, comprising:
at least one processor; and
at least one memory communicatively coupled to the processor, wherein:
the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the CRM service recommendation method of any of claims 1-7.
10. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the CRM service recommendation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811563505.3A CN111353793A (en) | 2018-12-20 | 2018-12-20 | CRM (customer relationship management) service recommendation method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811563505.3A CN111353793A (en) | 2018-12-20 | 2018-12-20 | CRM (customer relationship management) service recommendation method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111353793A true CN111353793A (en) | 2020-06-30 |
Family
ID=71193665
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811563505.3A Pending CN111353793A (en) | 2018-12-20 | 2018-12-20 | CRM (customer relationship management) service recommendation method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111353793A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112016958A (en) * | 2020-08-18 | 2020-12-01 | 中移（杭州）信息技术有限公司 | Product selection method and device, electronic equipment and storage medium |
| CN112613906A (en) * | 2020-12-18 | 2021-04-06 | 北京思特奇信息技术股份有限公司 | Marketing recommendation evaluation method and device based on mobile communication network and computer equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101540874A (en) * | 2009-04-23 | 2009-09-23 | 中山大学 | Interactive TV program recommendation method based on collaborative filtration |
| CN104899246A (en) * | 2015-04-12 | 2015-09-09 | 西安电子科技大学 | Collaborative filtering recommendation method of user rating neighborhood information based on fuzzy mechanism |
| CN104933156A (en) * | 2015-06-25 | 2015-09-23 | 西安理工大学 | Collaborative filtering method based on shared neighbor clustering |
| CN106096718A (en) * | 2016-06-06 | 2016-11-09 | 华侨大学 | A kind of normalization PSO Stochastic Optimization Algorithms and collaborative filtering recommending method |
| CN106815347A (en) * | 2017-01-13 | 2017-06-09 | 沈阳工学院 | Improvement slope one Collaborative Filtering Recommendation Algorithms based on commodity similarity |
| CN108052639A (en) * | 2017-12-21 | 2018-05-18 | 中国联合网络通信集团有限公司 | Industry user based on carrier data recommends method and device |
-
2018
- 2018-12-20 CN CN201811563505.3A patent/CN111353793A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101540874A (en) * | 2009-04-23 | 2009-09-23 | 中山大学 | Interactive TV program recommendation method based on collaborative filtration |
| CN104899246A (en) * | 2015-04-12 | 2015-09-09 | 西安电子科技大学 | Collaborative filtering recommendation method of user rating neighborhood information based on fuzzy mechanism |
| CN104933156A (en) * | 2015-06-25 | 2015-09-23 | 西安理工大学 | Collaborative filtering method based on shared neighbor clustering |
| CN106096718A (en) * | 2016-06-06 | 2016-11-09 | 华侨大学 | A kind of normalization PSO Stochastic Optimization Algorithms and collaborative filtering recommending method |
| CN106815347A (en) * | 2017-01-13 | 2017-06-09 | 沈阳工学院 | Improvement slope one Collaborative Filtering Recommendation Algorithms based on commodity similarity |
| CN108052639A (en) * | 2017-12-21 | 2018-05-18 | 中国联合网络通信集团有限公司 | Industry user based on carrier data recommends method and device |
Non-Patent Citations (1)
| Title |
|---|
| DONGMING ZHAO 等: "A pitch distribution in slow-wave structure of STWT using Cauchy mutated cat swarm optimization with gravitional search operator", 《JOURNAL OF THE CHINESE INSTITUTE OF ENGINEERS》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112016958A (en) * | 2020-08-18 | 2020-12-01 | 中移（杭州）信息技术有限公司 | Product selection method and device, electronic equipment and storage medium |
| CN112016958B (en) * | 2020-08-18 | 2024-03-22 | 中移（杭州）信息技术有限公司 | Product selection method, device, electronic equipment and storage medium |
| CN112613906A (en) * | 2020-12-18 | 2021-04-06 | 北京思特奇信息技术股份有限公司 | Marketing recommendation evaluation method and device based on mobile communication network and computer equipment |
| CN112613906B (en) * | 2020-12-18 | 2024-04-30 | 北京思特奇信息技术股份有限公司 | Marketing recommendation evaluation method and device based on mobile communication network and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20220215032A1 (en) | Ai-based recommendation method and apparatus, electronic device, and storage medium | |
| US10963541B2 (en) | Systems, methods, and apparatuses for implementing a related command with a predictive query interface | |
| EP3623956A1 (en) | Method and apparatus for obtaining recommendation information, and electronic device | |
| US20160191450A1 (en) | Recommendations Engine in a Layered Social Media Webpage | |
| CN103038769B (en) | System and method for content to be directed into social network engine user | |
| WO2018040069A1 (en) | Information recommendation system and method | |
| JP4911536B2 (en) | Regional information retrieval device, regional information retrieval device control method, regional information retrieval system, and regional information retrieval system control method | |
| WO2021135562A1 (en) | Feature validity evaluation method and apparatus, and electronic device and storage medium | |
| US20150006414A1 (en) | Social network for employment search | |
| US11144594B2 (en) | Search method, search apparatus and non-temporary computer-readable storage medium for text search | |
| US8660901B2 (en) | Matching of advertising sources and keyword sets in online commerce platforms | |
| CN108648058B (en) | Product sorting method and device, electronic equipment and storage medium | |
| US20110093457A1 (en) | Method for calculating resource points of resource information and distributing points | |
| US10896458B2 (en) | Method, system, and computer-readable medium for product and vendor selection | |
| CN106663117A (en) | Constructing a graph that facilitates provision of exploratory suggestions | |
| US20160092771A1 (en) | Analysis of social media messages | |
| CN108399191B (en) | Personalized recommendation method for bidding information | |
| CN111310032A (en) | Resource recommendation method and device, computer equipment and readable storage medium | |
| CN113407854A (en) | Application recommendation method, device and equipment and computer readable storage medium | |
| CN111078997B (en) | Information recommendation method and device | |
| JP2012150563A (en) | Product recommendation device, method, and program | |
| CN111353793A (en) | CRM (customer relationship management) service recommendation method and device | |
| CN111428100A (en) | Data retrieval method and device, electronic equipment and computer-readable storage medium | |
| CN111160699A (en) | Expert recommendation method and system | |
| CN117421355A (en) | Search recall method, device and equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200630 |