CN109697258A - It is a kind of for determining the method and apparatus of the customization financial information of target user - Google Patents
It is a kind of for determining the method and apparatus of the customization financial information of target user Download PDFInfo
- Publication number
- CN109697258A CN109697258A CN201811611247.1A CN201811611247A CN109697258A CN 109697258 A CN109697258 A CN 109697258A CN 201811611247 A CN201811611247 A CN 201811611247A CN 109697258 A CN109697258 A CN 109697258A
- Authority
- CN
- China
- Prior art keywords
- user
- information
- financial information
- target user
- matching
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The purpose of the application is to provide a kind of method and apparatus for determining the customization financial information of target user, pre-processes to financial information to be processed, to obtain at least one first financial information;Matching operation is carried out to target user and first financial information;When the operating result of the matching operation includes successful match, first financial information is determined as to the customization financial information of the target user;Wherein, the matching operation includes implicit loopback matching operation, the implicit loopback matching operation includes: to determine corresponding implicit loopback matching user based on the user category information of the target user, and the user preference information of the target user is determined according to the association preference information of the implicit loopback matching user, matching operation is then carried out based on the user preference information.The application improves the accuracy of customization financial information, and the decision for user, especially new user offers convenience.
Description
Technical field
This application involves computer fields more particularly to a kind of for determining the skill of the customization financial information of target user
Art.
Background technique
For finance and economics reader, in the prior art, traditional information recommendation system can only provide better simply information,
But for reader, especially finance and economics reader, according to the variation of transaction situation, market conditions, relevant policies etc., needed for
Information be diversification and may it is fast-changing, traditional information recommendation system it is more difficult adapt to these variation.On the other hand, it is
The information customized is obtained, for field of finance and economics, environment (such as above-mentioned transaction situation, market conditions, relevant policies
Deng) quick variation will cause and customize the lag of information, bring obstacle to the decision of user.
Summary of the invention
The purpose of the application is to provide a kind of for determining the method and apparatus of the customization financial information of target user.
According to the one aspect of the application, a kind of method for determining the customization financial information of target user is provided,
Wherein, method includes the following steps:
A pre-processes financial information to be processed, to obtain at least one first financial information;
B carries out matching operation to target user and first financial information;
First financial information is determined as described by c when the operating result of the matching operation includes successful match
The customization financial information of target user;
Wherein, the matching operation includes implicit loopback matching operation, and the implicit loopback matching operation includes:
Corresponding implicit loopback matching user is determined based on the user category information of the target user, and according to described hidden
The association preference information of formula loopback matching user determines the user preference information of the target user, then inclined based on the user
Good information carries out matching operation;
The pretreatment includes following at least any one:
Extract keyword corresponding to the financial information to be processed;
Extract label corresponding to the financial information to be processed;
Duplicate removal is carried out to the extraction information of the financial information to be processed;
Determine the temperature information of the extraction information of the financial information to be processed.
According to further aspect of the application, provide a kind of for determining setting for the customization financial information of target user
It is standby, wherein the equipment includes:
First module, for being pre-processed to financial information to be processed, to obtain at least one first financial information;
Second module, for carrying out matching operation to target user and first financial information;
Third module, for when the operating result of the matching operation includes successful match, first finance and economics to be believed
Breath is determined as the customization financial information of the target user;
Wherein, the matching operation includes implicit loopback matching operation, and the implicit loopback matching operation includes:
Corresponding implicit loopback matching user is determined based on the user category information of the target user, and according to described hidden
The association preference information of formula loopback matching user determines the user preference information of the target user, then inclined based on the user
Good information carries out matching operation;
The pretreatment includes following at least any one:
Extract keyword corresponding to the financial information to be processed;
Extract label corresponding to the financial information to be processed;
Duplicate removal is carried out to the extraction information of the financial information to be processed;
Determine the temperature information of the extraction information of the financial information to be processed.
According to the one aspect of the application, a kind of equipment for determining the customization financial information of target user is provided,
The equipment includes:
Processor;And
It is arranged to the memory of storage computer executable instructions, the executable instruction makes the place when executed
Manage the operation that device executes approach described above.
According to the one aspect of the application, a kind of computer-readable medium of store instruction is provided, described instruction is in quilt
System is made to execute the operation of approach described above when execution.
Compared with prior art, the application can according to the variation of transaction situation, market conditions, relevant policies etc., for
Family, which provides, timely and effectively customizes financial information, and the accuracy of customization financial information is improved by implicit loopback matching, is
User, especially new user decision offer convenience.For example, in some embodiments, in order to reach best cold starting effect and
User's conversion ratio, the professional finance and economics editor of the optional access of this system is as user concealed good friend's intervention module with guiding in real time user.
Detailed description of the invention
By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other
Feature, objects and advantages will become more apparent upon:
Fig. 1 shows a kind of method for determining the customization financial information of target user according to the application one embodiment
Flow chart;
Fig. 2 shows according to a kind of for determining setting for the customization financial information of target user of the application another embodiment
Standby functional block diagram;
Fig. 3 shows a kind of example devices functional block diagram according to the application one embodiment.
The same or similar appended drawing reference represents the same or similar component in attached drawing.
Specific embodiment
The application is described in further detail with reference to the accompanying drawing.
In a typical configuration of this application, terminal, the equipment of service network and trusted party include one or more
Processor (CPU), input/output interface, network interface and memory.
Memory may include the non-volatile memory in computer-readable medium, random access memory (RAM) and/or
The forms such as Nonvolatile memory, such as read-only memory (ROM) or flash memory (flash RAM).Memory is computer-readable medium
Example.
Computer-readable medium includes permanent and non-permanent, removable and non-removable media can be by any method
Or technology come realize information store.Information can be computer readable instructions, data structure, the module of program or other data.
The example of the storage medium of computer includes, but are not limited to phase change memory (PRAM), static random access memory (SRAM), moves
State random access memory (DRAM), other kinds of random access memory (RAM), read-only memory (ROM), electric erasable
Programmable read only memory (EEPROM), flash memory or other memory techniques, read-only disc read only memory (CD-ROM) (CD-ROM),
Digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or
Any other non-transmission medium, can be used for storage can be accessed by a computing device information.
The application meaning equipment includes but is not limited to that user equipment, the network equipment or user equipment and the network equipment pass through
Network is integrated constituted equipment.The user equipment includes but is not limited to that any one can carry out human-computer interaction with user
The mobile electronic product, such as smart phone, tablet computer etc. of (such as human-computer interaction is carried out by touch tablet), the mobile electricity
Sub- product can use any operating system, such as android operating system, iOS operating system.Wherein, the network equipment
The electronic equipment of numerical value calculating and information processing can be carried out automatically according to the instruction for being previously set or storing including a kind of,
Hardware includes but is not limited to microprocessor, specific integrated circuit (ASIC), programmable logic device (PLD), field programmable gate
Array (FPGA), digital signal processor (DSP), embedded device etc..The network equipment includes but is not limited to computer, net
The cloud that network host, single network server, multiple network server collection or multiple servers are constituted;Here, cloud is by based on cloud
The a large number of computers or network servers for calculating (Cloud Computing) is constituted, wherein cloud computing is the one of distributed computing
Kind, a virtual supercomputer consisting of a loosely coupled set of computers.The network includes but is not limited to interconnect
Net, wide area network, Metropolitan Area Network (MAN), local area network, VPN network, wireless self-organization network (Ad Hoc network) etc..Preferably, the equipment
Can also be run on the user equipment, the network equipment or user equipment and the network equipment, the network equipment, touch terminal or
The network equipment and touch terminal are integrated the program in constituted equipment by network.
Certainly, those skilled in the art will be understood that above equipment is only for example, other are existing or are likely to occur from now on
Equipment be such as applicable to the application, should also be included within the application protection scope, and be incorporated herein by reference.
In the description of the present application, the meaning of " plurality " is two or more, unless otherwise specifically defined.
This application provides a kind of information to determine equipment, which determines that equipment is associated with wealth for providing it for target user
Through information.Equipment is determined below based on the information, and the application is described in detail.
According to the one aspect of the application, the method for determining the customization financial information of target user is provided.Wherein,
The customization financial information includes the recommendation information for being provided to user in some embodiments.With reference to Fig. 1, this method includes
Step S10, step S20 and step S30.
In step slo, information determines that equipment pre-processes the financial information to be processed about target user, to obtain
Obtain at least one first financial information about target user.The information to be processed is obtained by system, for for system into one
Required information, such as above-mentioned first financial information are obtained after step processing.In some embodiments, system prepare first it is above-mentioned to
Information is handled, such as obtains financial information to be processed from least one financial information source.Specifically, at least one described finance and economics
Information source is in some embodiments a variety of different websites, for example, these websites include but is not limited to one below or
It is multinomial: social site;News site;Financial information website;Government information website.Here, those skilled in the art will be understood that
These financial information sources are only for example without carrying out any restriction, other wealth that are existing or being likely to occur from now on to the application
It such as can be suitably used for the application through information source, be also contained in the protection scope of the application, and be incorporated herein by reference.
Wherein, convenient into one for the financial information to be processed to be processed into the pretreatment of financial information to be processed
The form of step data processing.In some embodiments, to the pretreatment of financial information to be processed, including every processing as described below
One of mode is several:
Extract keyword corresponding to the financial information to be processed, such as the abstract part based on current financial information
Perhaps full text is compared with popular vocabulary to determine the popular vocabulary hit as keyword or plucking current financial information
The higher vocabulary of the frequency of occurrences in part or full text is wanted, as keyword;
Label corresponding to the financial information to be processed is extracted, such as system provides a user current financial information
At least one extracts information (such as keyword for user's selection and as label), and according to the selection operation of user, determines
It is at least one of described to extract at least one in information as label corresponding to the current financial information;
Duplicate removal is carried out to the financial information (or it extracts information) to be processed, such as is calculated in information and database
Other information similarity, and merger repeat information, in some embodiments can be by the preparatory distance of vector space without prison
Superintend and direct learning algorithm progress;
Determine the temperature information of the financial information to be processed (or it extracts information), such as by platform interior
Popular information be monitored gained.
Wherein, in some embodiments, a certain item extracts information, is from current financial information (title and/or abstract
And/or text) in extract a keyword.
In step S20, information determines that equipment carries out matching operation to target user and first financial information;This
One financial information will act as determining the basis of the information provided a user.In some embodiments, above-mentioned matching operation is to be based on
What machine learning carried out, specific implementation algorithm can be based on including but not limited to BP neural network, deep neural network, supporting vector
The data models such as machine, random forest are realized.Those skilled in the art will be understood that data model listed herewith is only for example,
And it is not construed as any restriction to the application;Other data models that are existing or being likely to occur from now on, such as can be suitably used for
The application is also contained in the protection scope of the application, and is incorporated herein by reference.
In step s 30, information determines that equipment, will be described when the operating result of the matching operation includes successful match
At least one first financial information is provided to the target user.In some embodiments, which includes implicit loopback
Matching operation.Wherein, in some embodiments, implicit loopback matching operation, comprising: the user class based on the target user
Other information determines corresponding implicit loopback matching user, and is determined according to the association preference information of the implicit loopback matching user
The user preference information of the target user then carries out matching operation based on the user preference information.For example, described implicit
Loopback matches the finance and economics editor that user is profession;Recommend information according to the related information of client and other clients;Finance and economics editor with
Machine becomes the implicit good friend of client;Finance and economics editor makes professional judge to the recommendation of relevent information data, is transmitted back to information recommendation
System.Hot spot article seen by user will be consistent with its implicit good friend.Finance and economics is edited specialized recommendation and is participated in, and is protecting
Under the premise of demonstrate,proving personalized recommendation, focus incident is allowed to push to user at the moment at the first time.
In some embodiments, information determines equipment according to financial information Matching Model, to target user and described first
Financial information carries out matching operation, wherein the financial information Matching Model is developed based on deep neural network.Certainly, this field
Technical staff will be understood that above-mentioned financial information Matching Model is developed based on deep neural network, be only for example, without that should regard
For any restrictions to the application;In fact, data model that is any existing or being likely to occur from now on, as can for developing
Above-mentioned financial information Matching Model and be suitable for the application, be also contained in the protection scope of the application, and by reference
It is incorporated herein.For example, in some embodiments, above-mentioned financial information Matching Model is based on decision tree, random forest, BP nerve net
Network and/or support vector machines etc. are realized.
Wherein, above-described financial information Matching Model includes at least two layers full articulamentum, it preferably includes at least four
The full articulamentum of layer, every layer of full articulamentum are handled through batch normalization.Here, the batch normalization for being applied to neural network operates
It can make the average value zero of data and variance is units, so that various features are in same magnitude.Because in nerve net
In the data transmission of network, sometimes some features can be very big or very small, therefore batch normalization can make network more preferable
Learning characteristic weight (W) and deviation (B).Compared to such as only to data input when pre-process, to each layer and
Neural network enables to neural network model more accurate using batch normalization.Specifically, the step of the batch normalization
Suddenly it specifically includes that
1) mean value and variance of batching data are calculated;
2) standardizing as a result, being done to the batch data according to above-mentioned processing;
3) it is multiplied by scaled matrix, and adds translation vector.
The input changes in distribution of hidden layer is excessive in order to prevent, can allow each hidden node activation input distribution narrow down to (-
1,1) on the one hand section can reduce the input space, to reduce tune ginseng difficulty;On the other hand, it can prevent gradient from exploding/disappearing
It loses, to accelerate network convergence.
Wherein, the quantity of neuron determines the accuracy of model in neural network, can be by every in control neural network
The quantity of the neuron of one level carrys out the fitting degree of Controlling model.Specifically, very little neuron makes neural network not
It can learn the connection between outputting and inputting well, too many neuron will lead to the over-fitting of model.Pass through profession
Neural network design and adjustment, can guarantee that recommender system reaches high precision.Pass through training, adjustment and the knowledge mapping of profession
Cooperation, above-mentioned neural network is capable of the hobby of point-device matching behavior of user, feature and user, thus by content standard
Really recommend target user.
Further, in some embodiments, the variance of the output of at least two layers full articulamentum is equal.For example, making
Ensure that the variable in neural network can be initialized preferably with Xavier initial method.Basically, at the beginning of Xavier
Beginningization can guarantee that the variable of the neural network of each level is initialized in same range, this initial method it is main
Purposes is to avoid the loss and explosion of gradient, neural network can be allowed to restrain faster using Xavier initialization.
In some embodiments, above-mentioned steps S20 includes sub-step S21 (not shown), sub-step S22 (not shown) and son
Step S23 (not shown).In sub-step S21, information determines user's operation and the user of the equipment according to target user
The operation corresponding operating time determines at least one preferences purpose preference in the user preference information of the target user
Weight.For example, such as user it is recent behavior it is important, for a long time before behavior it is then relatively secondary;The preference weight is used for
Measure user behavior significance level for a user, pay a price bigger behavior its weight of user is higher.Then, in sub-step
In rapid S22, information determines that equipment is based on the preference information and the preference weight, determine the feature of the target user to
Amount;In sub-step S23, information determines that equipment is based on described eigenvector to the target user and first financial information
Carry out matching operation.In this way, system in combination with user behavior property (such as behavior occur time or user
The effort paid needed for carrying out the behavior) for user personalized information is provided, so that recommendation information more meets user currently
Instant demand, and depend not only on a large amount of historical behaviors of user.
In some embodiments, the pretreatment includes the temperature letter for the extraction information for determining the financial information to be processed
Breath.Correspondingly, in sub-step S21, information determines that equipment obtains user's operation and the user's operation institute of target user
The corresponding operating time;If the user's operation is associated with the extraction information, according to the user's operation of target user, described
Operating time corresponding to user's operation and the temperature information, determine in the user preference information of the target user extremely
One item missing preferences purpose preference weight, wherein the preference weight and the temperature information are negatively correlated.For example, in general,
The recent behavior of user is important, and user for a long time before behavior it is relatively secondary.Therefore, if user read certain recently
Piece article, then the corresponding feature of this article will have relatively high weight.Furthermore the number of user behavior, is used sometimes
Family can generate behavior many times to an article.Therefore the number that the same behavior of same information occurs for user also reflects
For user to the interest of information, the corresponding feature weight of the information of behavior often is higher.Finally, if user is to an awfully hot door
Topic produce behavior, tend not to the individual character for representing user, may be to the topic not because user may be to follow the wind
There is too big interest, as soon as especially more illustrate to use when user produces unessential behavior several times once in a while to a popular information
Family may may be simply because this topic and be everywhere, it is easy to see to this information without what interest.Instead
It, if as soon as user produces behavior to a not popular topic, illustrate the individual needs of user.To by above
The mode, the application can be provided in conjunction with customer priorities more accurately recommends information.
Further, in some embodiments, the pretreatment includes the extraction letter for determining the financial information to be processed
The temperature information of breath;The above method further includes step S40 (not shown) and step S50 (not shown).In step s 40, information
Determine that equipment determines the degree of association of the target user Yu at least one pre-set user;In step s 50, information determines that equipment is worked as
The degree of association meets preset trigger condition, and the second financial information associated with the pre-set user is provided to the target
User.Wherein, in some embodiments, the pre-set user corresponds to " old user " or " professional user " of system, they
Opinion will be higher to the reference value of similar other users, especially novice users.Therefore, these pre-set users with
When target user reaches certain correlation degree, according to the preference of these pre-set users, provide to target user for reference
Financial information (referred herein to " the second financial information "), will provide help significantly for target user.
Based on the above, further, in some embodiments, preset trigger condition described above includes below one
Item is multinomial:
The degree of association of at least one pre-set user is higher than the degree of association of other pre-set users, such as chooses the degree of association
The related content of highest several pre-set users pushes to target user;
The degree of association of at least one pre-set user is higher than a default degree of association threshold value, such as meets the requirements default
The related content of user is pushed to target user;
The speedup of the degree of association of at least one pre-set user be higher than a speedup threshold value, such as target user have " connect
Closely " the trend of certain pre-set user illustrates the transfer tendency of target user's attention, can be predicted simultaneously target user
The possible interested content of push, to improve customer experience.
According to further aspect of the application, the equipment for determining the customization financial information of target user is provided.Its
In, the customization financial information includes the recommendation information for being provided to user in some embodiments.With reference to Fig. 1, the equipment
Including the first module 10, the second module 20 and third module 30.
First module 10 pre-processes the financial information to be processed about target user, to obtain about target user
At least one of first financial information.The information to be processed is obtained by system, for obtaining institute after being further processed for system
The information needed, such as above-mentioned first financial information.In some embodiments, system prepares above-mentioned information to be processed first, such as
Financial information to be processed is obtained from least one financial information source.Specifically, at least one described financial information source, some
It is a variety of different websites in embodiment, such as these websites include but is not limited to one or more below: social site;
News site;Financial information website;Government information website.Here, those skilled in the art will be understood that these financial information sources
It is only for example without carrying out any restriction to the application, other existing or financial information sources for being likely to occur from now on are as can suitable
It for the application, is also contained in the protection scope of the application, and is incorporated herein by reference.
Wherein, convenient into one for the financial information to be processed to be processed into the pretreatment of financial information to be processed
The form of step data processing.In some embodiments, to the pretreatment of financial information to be processed, including every processing as described below
One of mode is several:
Extract keyword corresponding to the financial information to be processed, such as the abstract part based on current financial information
Perhaps full text is compared with popular vocabulary to determine the popular vocabulary hit as keyword or plucking current financial information
The higher vocabulary of the frequency of occurrences in part or full text is wanted, as keyword;
Label corresponding to the financial information to be processed is extracted, such as system provides a user current financial information
At least one extracts information (such as keyword for user's selection and as label), and according to the selection operation of user, determines
It is at least one of described to extract at least one in information as label corresponding to the current financial information;
Duplicate removal is carried out to the financial information (or it extracts information) to be processed, such as is calculated in information and database
Other information similarity, and merger repeat information, in some embodiments can be by the preparatory distance of vector space without prison
Superintend and direct learning algorithm progress;
Determine the temperature information of the financial information to be processed (or it extracts information), such as by platform interior
Popular information be monitored gained.
Wherein, in some embodiments, a certain item extracts information, is from current financial information (title and/or abstract
And/or text) in extract a keyword.
Second module 20 carries out matching operation to target user and first financial information;First financial information will be used
Make the basis of the determining information provided a user.In some embodiments, above-mentioned matching operation is carried out based on machine learning,
Implementing algorithm can be based on the including but not limited to number such as BP neural network, deep neural network, support vector machines, random forest
According to model realization.Those skilled in the art will be understood that data model listed herewith is only for example, and be not construed as to this Shen
Any restriction please;Other data models that are existing or being likely to occur from now on, such as can be suitably used for the application, be also contained in this
In the protection scope of application, and it is incorporated herein by reference.
Third module 30, will at least one described first wealth when the operating result of the matching operation includes successful match
The target user is provided to through information.In some embodiments, which includes implicit loopback matching operation.Wherein,
In some embodiments, implicit loopback matching operation, comprising: correspondence is determined based on the user category information of the target user
Implicit loopback match user, and determine the target user's according to the association preference information of the implicit loopback matching user
User preference information then carries out matching operation based on the user preference information.For example, the implicit loopback matching user is
The finance and economics editor of profession;Recommend information according to the related information of client and other clients;Finance and economics editor is at random as the hidden of client
Formula good friend;Finance and economics editor makes professional judge to the recommendation of relevent information data, is transmitted back to information recommendation system.Seen by user
Hot spot article will be consistent with its implicit good friend.Finance and economics is edited specialized recommendation and is participated in, and is guaranteeing personalized recommendation
Under the premise of, allow focus incident to push to user at the moment at the first time.
In some embodiments, information determines equipment according to financial information Matching Model, to target user and described first
Financial information carries out matching operation, wherein the financial information Matching Model is developed based on deep neural network.Certainly, this field
Technical staff will be understood that above-mentioned financial information Matching Model is developed based on deep neural network, be only for example, without that should regard
For any restrictions to the application;In fact, data model that is any existing or being likely to occur from now on, as can for developing
Above-mentioned financial information Matching Model and be suitable for the application, be also contained in the protection scope of the application, and by reference
It is incorporated herein.For example, in some embodiments, above-mentioned financial information Matching Model is based on decision tree, random forest, BP nerve net
Network and/or support vector machines etc. are realized.
Wherein, above-described financial information Matching Model includes at least two layers full articulamentum, it preferably includes at least four
The full articulamentum of layer, every layer of full articulamentum are handled through batch normalization.Here, the batch normalization for being applied to neural network operates
It can make the average value zero of data and variance is units, so that various features are in same magnitude.Because in nerve net
In the data transmission of network, sometimes some features can be very big or very small, therefore batch normalization can make network more preferable
Learning characteristic weight (W) and deviation (B).Compared to such as only to data input when pre-process, to each layer and
Neural network enables to neural network model more accurate using batch normalization.Specifically, the step of the batch normalization
Suddenly it specifically includes that
1) mean value and variance of batching data are calculated;
2) standardizing as a result, being done to the batch data according to above-mentioned processing;
3) it is multiplied by scaled matrix, and adds translation vector.
The input changes in distribution of hidden layer is excessive in order to prevent, can allow each hidden node activation input distribution narrow down to (-
1,1) on the one hand section can reduce the input space, to reduce tune ginseng difficulty;On the other hand, it can prevent gradient from exploding/disappearing
It loses, to accelerate network convergence.
Wherein, the quantity of neuron determines the accuracy of model in neural network, can be by every in control neural network
The quantity of the neuron of one level carrys out the fitting degree of Controlling model.Specifically, very little neuron makes neural network not
It can learn the connection between outputting and inputting well, too many neuron will lead to the over-fitting of model.Pass through profession
Neural network design and adjustment, can guarantee that recommender system reaches high precision.Pass through training, adjustment and the knowledge mapping of profession
Cooperation, above-mentioned neural network is capable of the hobby of point-device matching behavior of user, feature and user, thus by content standard
Really recommend target user.
Further, in some embodiments, the variance of the output of at least two layers full articulamentum is equal.For example, making
Ensure that the variable in neural network can be initialized preferably with Xavier initial method.Basically, at the beginning of Xavier
Beginningization can guarantee that the variable of the neural network of each level is initialized in same range, this initial method it is main
Purposes is to avoid the loss and explosion of gradient, neural network can be allowed to restrain faster using Xavier initialization.
In some embodiments, above-mentioned second module 20 (is not shown including 21 (not shown) of first unit, second unit 22
Out) and 23 (not shown) of third unit.First unit 21 is right according to the user's operation of target user and user's operation institute
The operating time answered determines at least one preferences purpose preference weight in the user preference information of the target user.Example
Such as, for example, user it is recent behavior it is important, for a long time before behavior it is then relatively secondary;The preference weight is for measuring user
The significance level of behavior for a user, pay a price bigger behavior its weight of user are higher.Then, second unit 22 is based on
The preference information and the preference weight, determine the feature vector of the target user;Third unit 23 is based on the spy
It levies vector and matching operation is carried out to the target user and first financial information.In this way, system is in combination with use
The behavior property (such as the effort paid needed for the time or user's implementation behavior of behavior generation) at family provides a for user
Property information so that recommendation information more meets the current instant demand of user, and depend not only on a large amount of history of user
Behavior.
In some embodiments, the pretreatment includes the temperature letter for the extraction information for determining the financial information to be processed
Breath.Correspondingly, first unit 21 obtains the operating time corresponding to the user's operation and the user's operation of target user;If
The user's operation is associated with the extraction information, according to corresponding to the user's operation of target user, the user's operation
Operating time and the temperature information determine at least one preferences purpose in the user preference information of the target user
Preference weight, wherein the preference weight and the temperature information are negatively correlated.For example, in general, the recent behavior ratio of user
It is more important, and user for a long time before behavior it is relatively secondary.Therefore, if user read certain article recently, this
The corresponding feature of article will have relatively high weight.Furthermore the number of user behavior, user can produce an article sometimes
Raw behavior many times.Therefore the number that the same behavior of same information occurs for user also reflects user to the emerging of information
Interest, the corresponding feature weight of the information of behavior often are higher.Finally, if user produces row to the topic of an awfully hot door
For, the individual character for representing user is tended not to, it, may be special to the topic there is no too big interest because user may be to follow the wind
As soon as not being more to illustrate that user can to this information when user produces unessential behavior several times once in a while to a popular information
It can may be simply because this topic without what interest and be everywhere, it is easy to see., whereas if user is to one
A not popular topic produces behavior, just illustrates the individual needs of user.To pass through manner discussed above, this Shen
It please can be provided in conjunction with customer priorities and more accurately recommend information.
Further, in some embodiments, the pretreatment includes the extraction letter for determining the financial information to be processed
The temperature information of breath;Above equipment further includes 40 (not shown) of the 4th module and 50 (not shown) of the 5th module.4th module 40
Determine the degree of association of the target user Yu at least one pre-set user;5th module 50 presets triggering when the degree of association meets
The second financial information associated with the pre-set user is provided to the target user by condition.Wherein, in some embodiments
In, the pre-set user corresponds to " old user " or " professional user " of system, their opinion is to other similar use
The reference value of family, especially novice users will be higher.Therefore, reach certain with target user in these pre-set users to be associated with
When degree, according to the preference of these pre-set users, financial information (referred herein to " the second wealth for reference is provided to target user
Through information "), will help be provided for target user significantly.
Based on the above, further, in some embodiments, preset trigger condition described above includes below one
Item is multinomial:
The degree of association of at least one pre-set user is higher than the degree of association of other pre-set users, such as chooses the degree of association
The related content of highest several pre-set users pushes to target user;
The degree of association of at least one pre-set user is higher than a default degree of association threshold value, such as meets the requirements default
The related content of user is pushed to target user;
The speedup of the degree of association of at least one pre-set user be higher than a speedup threshold value, such as target user have " connect
Closely " the trend of certain pre-set user illustrates the transfer tendency of target user's attention, can be predicted simultaneously target user
The possible interested content of push, to improve customer experience.
Present invention also provides a kind of computer readable storage medium, the computer-readable recording medium storage has calculating
Machine code, when the computer code is performed, such as preceding described in any item methods are performed.
Present invention also provides a kind of computer program products, when the computer program product is executed by computer equipment
When, such as preceding described in any item methods are performed.
Present invention also provides a kind of computer equipment, the computer equipment includes:
One or more processors;
Memory, for storing one or more computer programs;
When one or more of computer programs are executed by one or more of processors so that it is one or
Multiple processors realize such as preceding described in any item methods.
Fig. 3 shows the exemplary system that can be used for implementing each embodiment described herein;
As shown in Figure 3 in some embodiments, system 100 can be true as any one information in each embodiment
Locking equipment.In some embodiments, system 100 may include one or more computer-readable mediums with instruction (for example, being
System memory or NVM/ store equipment 120) and coupled with the one or more computer-readable medium and be configured as executing
Instruction with realize module thereby executing movement described herein one or more processors (for example, (one or more)
Processor 105).
For one embodiment, system control module 110 may include any suitable interface controller, with to (one or
It is multiple) at least one of processor 105 and/or any suitable equipment or component that communicate with system control module 110 mentions
For any suitable interface.
System control module 110 may include Memory Controller module 130, to provide interface to system storage 115.It deposits
Memory controller module 130 can be hardware module, software module and/or firmware module.
System storage 115 can be used for for example, load of system 100 and storing data and/or instruction.For a reality
Example is applied, system storage 115 may include any suitable volatile memory, for example, DRAM appropriate.In some embodiments
In, system storage 115 may include four Synchronous Dynamic Random Access Memory of Double Data Rate type (DDR4SDRAM).
For one embodiment, system control module 110 may include one or more input/output (I/O) controller, with
Equipment 120 is stored to NVM/ and (one or more) communication interface 125 provides interface.
For example, NVM/ storage equipment 120 can be used for storing data and/or instruction.NVM/ storage equipment 120 may include appointing
It anticipates nonvolatile memory appropriate (for example, flash memory) and/or to may include that any suitable (one or more) is non-volatile deposit
Equipment is stored up (for example, one or more hard disk drives (HDD), one or more CD (CD) drivers and/or one or more
Digital versatile disc (DVD) driver).
NVM/ storage equipment 120 may include a part for the equipment being physically mounted on as system 100
Storage resource or its can by the equipment access without a part as the equipment.For example, NVM/ storage equipment 120 can
It is accessed by network via (one or more) communication interface 125.
(one or more) communication interface 125 can be provided for system 100 interface with by one or more networks and/or with
Other any equipment communications appropriate.System 100 can be according to any mark in one or more wireless network standards and/or agreement
Quasi- and/or agreement is carried out wireless communication with the one or more components of wireless network.
For one embodiment, at least one of (one or more) processor 105 can be with system control module 110
The logic of one or more controllers (for example, Memory Controller module 130) is packaged together.For one embodiment, (one
It is a or multiple) at least one of processor 105 can encapsulate with the logic of one or more controllers of system control module 110
Together to form system in package (SiP).For one embodiment, at least one of (one or more) processor 105
It can be integrated on same mold with the logic of one or more controllers of system control module 110.For one embodiment,
At least one of (one or more) processor 105 can be with the logic of one or more controllers of system control module 110
It is integrated on same mold to form system on chip (SoC).
In various embodiments, system 100 can be, but not limited to be: server, work station, desk-top calculating equipment or movement
It calculates equipment (for example, lap-top computing devices, handheld computing device, tablet computer, net book etc.).In various embodiments,
System 100 can have more or fewer components and/or different frameworks.For example, in some embodiments, system 100 includes
One or more video cameras, keyboard, liquid crystal display (LCD) screen (including touch screen displays), nonvolatile memory port,
Mutiple antennas, graphic chips, specific integrated circuit (ASIC) and loudspeaker.
It should be noted that the application can be carried out in the assembly of software and/or software and hardware, for example, can adopt
With specific integrated circuit (ASIC), general purpose computer or any other realized similar to hardware device.In one embodiment
In, the software program of the application can be executed to implement the above steps or functions by processor.Similarly, the application
Software program (including relevant data structure) can be stored in computer readable recording medium, for example, RAM memory,
Magnetic or optical driver or floppy disc and similar devices.In addition, hardware can be used to realize in some steps or function of the application, example
Such as, as the circuit cooperated with processor thereby executing each step or function.
In addition, a part of the application can be applied to computer program product, such as computer program instructions, when its quilt
When computer executes, by the operation of the computer, it can call or provide according to the present processes and/or technical solution.
Those skilled in the art will be understood that the existence form of computer program instructions in computer-readable medium includes but is not limited to
Source file, executable file, installation package file etc., correspondingly, the mode that computer program instructions are computer-executed include but
Be not limited to: the computer directly execute the instruction or the computer compile the instruction after execute program after corresponding compiling again,
Perhaps the computer reads and executes the instruction or after the computer reads and install and execute corresponding installation again after the instruction
Program.Here, computer-readable medium can be for computer access any available computer readable storage medium or
Communication media.
Communication media includes whereby including, for example, computer readable instructions, data structure, program module or other data
Signal of communication is transmitted to the medium of another system from a system.Communication media may include having the transmission medium led (such as electric
Cable and line (for example, optical fiber, coaxial etc.)) and can propagate wireless (not having the transmission the led) medium of energy wave, such as sound, electricity
Magnetic, RF, microwave and infrared.Computer readable instructions, data structure, program module or other data can be embodied as example wireless
Medium (such as carrier wave or be such as embodied as spread spectrum technique a part similar mechanism) in modulated message signal.
Term " modulated message signal " refers to that one or more feature is modified or is set in a manner of encoded information in the signal
Fixed signal.Modulation can be simulation, digital or Hybrid Modulation Technology.
As an example, not a limit, computer readable storage medium may include such as computer-readable finger for storage
Enable, the volatile and non-volatile that any method or technique of the information of data structure, program module or other data is realized, can
Mobile and immovable medium.For example, computer readable storage medium includes, but are not limited to volatile memory, such as with
Machine memory (RAM, DRAM, SRAM);And nonvolatile memory, such as flash memory, various read-only memory (ROM, PROM,
EPROM, EEPROM), magnetic and ferromagnetic/ferroelectric memory (MRAM, FeRAM);And magnetic and optical storage apparatus (hard disk,
Tape, CD, DVD);Or other currently known media or Future Development can store the computer used for computer system
Readable information/data.
Here, including a device according to one embodiment of the application, which includes for storing computer program
The memory of instruction and processor for executing program instructions, wherein when the computer program instructions are executed by the processor
When, trigger method and/or technology scheme of the device operation based on aforementioned multiple embodiments according to the application.
It is obvious to a person skilled in the art that the application is not limited to the details of above-mentioned exemplary embodiment, Er Qie
In the case where without departing substantially from spirit herein or essential characteristic, the application can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and scope of the present application is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included in the application.Any reference signs in the claims should not be construed as limiting the involved claims.This
Outside, it is clear that one word of " comprising " does not exclude other units or steps, and odd number is not excluded for plural number.That states in device claim is multiple
Unit or device can also be implemented through software or hardware by a unit or device.The first, the second equal words are used to table
Show title, and does not indicate any particular order.
Claims (10)
1. a kind of method for determining the customization financial information of target user, wherein method includes the following steps:
A pre-processes financial information to be processed, to obtain at least one first financial information;
B carries out matching operation to target user and first financial information;
C is determined as the target when the operating result of the matching operation includes successful match, by first financial information
The customization financial information of user;
Wherein, the matching operation includes implicit loopback matching operation, and the implicit loopback matching operation includes:
Corresponding implicit loopback matching user is determined based on the user category information of the target user, and according to described implicit time
The user preference information for sending the association preference information of matching user to determine the target user is then believed based on the user preference
Breath carries out matching operation;
The pretreatment includes following at least any one:
Extract keyword corresponding to the financial information to be processed;
Extract label corresponding to the financial information to be processed;
Duplicate removal is carried out to the extraction information of the financial information to be processed;
Determine the temperature information of the extraction information of the financial information to be processed.
2. according to the method described in claim 1, wherein, the implicit loopback matching operation includes:
Corresponding implicit loopback matching user is determined based on the user category information of the target user, and according to described implicit time
The user preference information for sending the association preference information of matching user to determine the target user is then believed based on the user preference
Breath carries out matching operation.
3. according to the method described in claim 1, wherein, the step b includes:
According to financial information Matching Model, matching operation is carried out to target user and first financial information, wherein the wealth
It is developed through information matches model based on deep neural network;
Wherein, the financial information Matching Model includes at least two layers full articulamentum, and every layer of full articulamentum is through batch normalization
Processing.
4. according to the method described in claim 3, wherein, the variance of the output of at least two layers full articulamentum is equal.
5. according to the method described in claim 1, wherein, the step b includes:
R operating time according to corresponding to the user's operation of target user and the user's operation, determine the target user
User preference information at least one of preferences purpose preference weight;
S is based on the preference information and the preference weight, determines the feature vector of the target user;
T is based on described eigenvector and carries out matching operation to the target user and first financial information.
6. according to the method described in claim 5, wherein, the pretreatment includes determining the extraction of the financial information to be processed
The temperature information of information;
The step r includes:
Obtain the operating time corresponding to the user's operation and the user's operation of target user;
If the user's operation is associated with the extraction information, according to the user's operation of target user, user's operation institute
Corresponding operating time and the temperature information determine at least one preference in the user preference information of the target user
The preference weight of project, wherein the preference weight and the temperature information are negatively correlated.
7. according to the method described in claim 1, wherein, the method also includes:
Determine the degree of association of the target user Yu at least one pre-set user;
When the degree of association meets preset trigger condition, the second financial information associated with the pre-set user is provided to institute
State target user.
8. according to the method described in claim 7, wherein, the preset trigger condition includes following at least any one:
The degree of association of at least one pre-set user is higher than the degree of association of other pre-set users;
The degree of association of at least one pre-set user is higher than a default degree of association threshold value;
The speedup of the degree of association of at least one pre-set user is higher than a speedup threshold value.
9. a kind of equipment for determining the customization financial information of target user, wherein the equipment includes:
Processor;And
It is arranged to the memory of storage computer executable instructions, the executable instruction makes the processor when executed
It executes according to claim 1 to the operation of any one of 8 the methods.
10. a kind of computer-readable medium of store instruction, described instruction wants system execution according to right
Ask the operation of any one of 1 to 8 the method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811611247.1A CN109697258A (en) | 2018-12-27 | 2018-12-27 | It is a kind of for determining the method and apparatus of the customization financial information of target user |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811611247.1A CN109697258A (en) | 2018-12-27 | 2018-12-27 | It is a kind of for determining the method and apparatus of the customization financial information of target user |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109697258A true CN109697258A (en) | 2019-04-30 |
Family
ID=66232191
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811611247.1A Pending CN109697258A (en) | 2018-12-27 | 2018-12-27 | It is a kind of for determining the method and apparatus of the customization financial information of target user |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109697258A (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102236710A (en) * | 2011-06-30 | 2011-11-09 | 百度在线网络技术(北京)有限公司 | Method and equipment for displaying news information in query result |
CN104933201A (en) * | 2015-07-15 | 2015-09-23 | 蔡宏铭 | Content recommendation method and system based on peer information |
CN105447038A (en) * | 2014-08-29 | 2016-03-30 | 国际商业机器公司 | Method and system for acquiring user characteristics |
CN105718464A (en) * | 2014-12-02 | 2016-06-29 | 阿里巴巴集团控股有限公司 | Information retrieval method and device |
US9529919B2 (en) * | 2010-03-29 | 2016-12-27 | Paypal, Inc. | Traffic driver for suggesting stores |
CN107391509A (en) * | 2016-05-16 | 2017-11-24 | 中兴通讯股份有限公司 | Label recommendation method and device |
CN108304441A (en) * | 2017-11-14 | 2018-07-20 | 腾讯科技(深圳)有限公司 | Network resource recommended method, device, electronic equipment, server and storage medium |
-
2018
- 2018-12-27 CN CN201811611247.1A patent/CN109697258A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9529919B2 (en) * | 2010-03-29 | 2016-12-27 | Paypal, Inc. | Traffic driver for suggesting stores |
CN102236710A (en) * | 2011-06-30 | 2011-11-09 | 百度在线网络技术(北京)有限公司 | Method and equipment for displaying news information in query result |
CN105447038A (en) * | 2014-08-29 | 2016-03-30 | 国际商业机器公司 | Method and system for acquiring user characteristics |
CN105718464A (en) * | 2014-12-02 | 2016-06-29 | 阿里巴巴集团控股有限公司 | Information retrieval method and device |
CN104933201A (en) * | 2015-07-15 | 2015-09-23 | 蔡宏铭 | Content recommendation method and system based on peer information |
CN107391509A (en) * | 2016-05-16 | 2017-11-24 | 中兴通讯股份有限公司 | Label recommendation method and device |
CN108304441A (en) * | 2017-11-14 | 2018-07-20 | 腾讯科技(深圳)有限公司 | Network resource recommended method, device, electronic equipment, server and storage medium |
Non-Patent Citations (2)
Title |
---|
JNINGWEI: "深度学习_ Batch Normalization (归一化)", 《HTTPS://CLOUD.TENCENT.COM/DEVELOPER/ARTICLE/1347847腾讯云》 * |
SERGEY LOFFE等: "Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift", 《ARXIV》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108446374B (en) | User's Intention Anticipation method, apparatus, electronic equipment, storage medium | |
US10366160B2 (en) | Automatic generation and display of context, missing attributes and suggestions for context dependent questions in response to a mouse hover on a displayed term | |
US11188720B2 (en) | Computing system including virtual agent bot providing semantic topic model-based response | |
US20210225380A1 (en) | Voiceprint recognition method and apparatus | |
US10430465B2 (en) | Dynamic faceting for personalized search and discovery | |
US10572929B2 (en) | Decision factors analyzing device and decision factors analyzing method | |
WO2017184204A1 (en) | Techniques for user-centric document summarization | |
US20210264517A1 (en) | Internet meme economy | |
US20200279289A1 (en) | Prompting web-based user interaction | |
CN108153848A (en) | Light searching method, equipment and the electronic equipment for applying data | |
US20190392077A1 (en) | Facet-based query refinement based on multiple query interpretations | |
US10255249B1 (en) | Previewing electronic book content within third-party websites | |
CN116821475B (en) | Video recommendation method and device based on client data and computer equipment | |
CN116108267A (en) | Recommendation method and related equipment | |
CN116933362A (en) | Automatic generation method and device for indoor design scheme, computer equipment and medium | |
CN115619448A (en) | User loss prediction method and device, computer equipment and storage medium | |
US10776411B2 (en) | Systematic browsing of automated conversation exchange program knowledge bases | |
JP7204903B2 (en) | INFORMATION PUSH METHOD, DEVICE, DEVICE AND STORAGE MEDIUM | |
CN109325227A (en) | Method and apparatus for generating amendment sentence | |
US20190188274A1 (en) | Recommendation prediction based on preference elicitation | |
CN110727853A (en) | Presenting controlled heterogeneous digital content to a user | |
CN109657153A (en) | It is a kind of for determining the method and apparatus of the association financial information of user | |
CN109740056A (en) | It is a kind of for provide a user customization financial information method and apparatus | |
CN109697258A (en) | It is a kind of for determining the method and apparatus of the customization financial information of target user | |
CN108376149A (en) | Method, apparatus, equipment and the computer storage media that information shows |
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 | ||
CB02 | Change of applicant information |
Address after: Room 6013, building 14, 785 Hutai Road, Jing'an District, Shanghai Applicant after: DANHAN INTELLIGENT TECHNOLOGY (SHANGHAI) Co.,Ltd. Address before: Room 118, Building 20, No. 83 Lane 1-42, Hongxiang North Road, Wanxiang Town, Pudong New Area, Shanghai Applicant before: DANHAN INTELLIGENT TECHNOLOGY (SHANGHAI) Co.,Ltd. |
|
CB02 | Change of applicant information | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190430 |
|
RJ01 | Rejection of invention patent application after publication |