CN115860865A

CN115860865A - Commodity combination construction method and device, equipment, medium and product thereof

Info

Publication number: CN115860865A
Application number: CN202211515083.9A
Authority: CN
Inventors: 钟媛媛
Original assignee: Guangzhou Huanju Shidai Information Technology Co Ltd
Current assignee: Guangzhou Huanju Shidai Information Technology Co Ltd
Priority date: 2022-11-29
Filing date: 2022-11-29
Publication date: 2023-03-28

Abstract

The application relates to a commodity combined construction method and a device, equipment, medium and product thereof, wherein the method comprises the following steps: acquiring a commodity set of each user, carrying out random item number combination on commodity items in the commodity set, enumerating all commodity combinations to form a user combination set, and carrying out de-coincidence on all the user combination sets to form a total combination set; determining the occurrence frequency of each commodity combination in the full combination set; determining the degree of aggregation based on the mutual information entropy of each commodity combination in the full-quantity combination set, wherein the degree of aggregation indicates the collocation stability of a plurality of commodity items which are accessed by the same user in a related manner; determining the degree of freedom based on the adjacent entropy of each commodity combination in the total combination set relative to all the commodity sets, wherein the degree of freedom represents the collocation abundance degree of the corresponding commodity combination which is matched with other commodity items and is associated and accessed by the same user; and screening a plurality of commodity combinations based on the appearance frequency, the degree of agglomeration and the degree of freedom to form a high-quality combination list. The method and the device can determine the high-quality commodity combination for the commodity recommendation service.

Description

Commodity combination construction method and device, equipment, medium and product thereof

Technical Field

The application relates to the E-commerce information processing technology, in particular to a commodity combined construction method and a device, equipment, medium and product thereof.

Background

The commodity combination recommendation is a high-frequency popularization mode in an e-commerce platform, particularly in the e-commerce platform based on independent stations, new commodities or new users of each independent station can not provide proper new commodity recommendation for the new users due to the fact that the new commodities and the new users lack historical data and need to realize cold start of recommendation services, high-quality commodity combinations are determined according to priori knowledge, a plurality of commodity items with associated access, particularly associated purchase relation, form the high-quality commodity combinations, according to the collocation relation among the commodity items in the high-quality commodity combinations, when the users access one commodity item, other commodity items in the commodity combinations are recommended, obstacles faced by cold start can be overcome, and the purpose of effective marketing and popularization is achieved.

In the conventional technology, in addition to manual customization of commodity combinations, a common mode is to obtain user behavior data of each commodity item by using various commodity similarity algorithms, determine co-occurrence information between the commodity items according to the characteristics that the commodity items are purchased by the same user, and identify two commodity items with higher co-occurrence chances as a basic combination. The algorithm is characterized in that the algorithm cannot macroscopically examine extensive contact information among the commodity items due to the fact that the algorithm identifies based on two commodity items each time, combination results are not accurate, the number of recommended commodity items which can be provided each time is limited, and the result is that a mechanism for recommending based on commodity combination plays a very weak role.

Actually, a commodity combination can become a good-quality commodity combination, not only the usage amount of the commodity combination needs to be considered, but also the fixed collocation and collocation stability among a plurality of commodity items in the commodity combination can be considered, and the independence shown by the fact that the commodity combination and other commodity items can be widely collocated is considered, so that the conventional technology which needs to find the commodity combination to be obtained can find out that the commodity combination determined by the conventional technology can only actually find out the similarity of purchased behaviors among different commodity items, and can not systematically and comprehensively ensure that the commodity combination obtains the rationality corresponding to a factual sales scene.

In view of this, there is still room for improvement in the data mining technology for commodity combinations, and further research is needed.

Disclosure of Invention

The present application is directed to solving the above-mentioned problems and providing a method for building a combination of articles, and a corresponding apparatus, device, non-volatile readable storage medium, and computer program product.

According to one aspect of the present application, there is provided a merchandise composite construction method, including the steps of:

acquiring a commodity set of each user, wherein the commodity set comprises a plurality of commodity items visited by the user, carrying out random item number combination on the commodity items in the commodity set, enumerating all commodity combinations to form a user group set, and de-overlapping all the user group sets to form a total combination set;

determining the occurrence frequency of each commodity combination in the total combination set in each user combination set;

determining the degree of aggregation of the commodity combinations based on a mutual information entropy obtained by performing any two segmentations on each commodity combination in the full-amount combination set, wherein the degree of aggregation indicates the collocation stability degree of a plurality of commodity items included in the corresponding commodity combinations and accessed by the same user in a related manner;

determining the degree of freedom of each commodity combination relative to the adjacency entropy of the commodity sets of all users based on each commodity combination in the full-scale combination set, wherein the degree of freedom represents the collocation abundance degree of the corresponding commodity combination which is collocated with other commodity items and is accessed by the same user in a relevant manner;

and screening out a plurality of commodity combinations in the total combination set based on the appearance frequency, the condensation degree and the degree of freedom to form a high-quality combination list.

Optionally, obtaining a commodity set of each user, where the commodity set includes a plurality of commodity items visited by the user, performing any number of combinations on the commodity items in the commodity set, enumerating all commodity combinations to form a user group set, and de-overlapping all user group sets to form a full-volume combination set, including:

the method comprises the steps of obtaining access behavior data generated by each user in the same online shop in historical access events, determining a plurality of commodity items accessed by the user from the access behavior data to form a commodity set of the user, wherein the historical access events comprise an event of adding the commodity items to a shopping cart and/or an event of adding the commodity items to a purchase order;

aiming at the commodity set of each user, combining the commodity items in the commodity set according to a plurality of random commodity items, enumerating all possible commodity combinations in the commodity set to form a user combination set;

and (4) the commodity combinations in the user combination set of all the users are subjected to de-coincidence and constructed into a full-quantity combination set.

Optionally, determining the degree of aggregation of each commodity combination based on a mutual information entropy obtained by performing any two segmentations on each commodity combination in the full-amount combination set includes:

aiming at each commodity combination in the full-quantity combination set, carrying out any possible two segmentations on the commodity combination by taking a commodity item as a unit to obtain a left side set and a right side set corresponding to each segmentation;

calculating mutual information entropy corresponding to each segmentation based on the left side set and the right side set obtained by each segmentation;

and taking the lowest mutual information entropy obtained after each commodity combination is segmented for multiple times as the degree of aggregation of the commodity combination.

Optionally, determining the degree of freedom of each commodity combination relative to the commodity set of all users based on the adjacency entropy of each commodity combination in the full-amount combination set includes:

sorting the commodity items in the commodity set of each user according to the user access time;

calculating left adjacency entropy and right adjacency entropy of each commodity combination in the full-amount combination set relative to each commodity set after the commodity combinations are sorted by users, and determining the lowest value of the left adjacency entropy and the right adjacency entropy as the lowest adjacency entropy;

for each combination of items, its lowest adjacency entropy with respect to all users is summarized as the degree of freedom for that combination of items.

Optionally, screening out a plurality of commodity combinations in the full combination set based on the appearance frequency, the degree of aggregation, and the degree of freedom to form a high-quality combination list, including:

multiplying the occurrence frequency, the degree of agglomeration and the degree of freedom of each commodity combination in the full combination set to obtain a comprehensive score corresponding to the commodity combination;

and screening the commodity combinations in the full combination set based on the comprehensive scores to obtain part of commodity combinations with relatively high comprehensive scores to form a high-quality combination list.

Optionally, after forming the list of good quality combinations, the method includes:

responding to an access event triggered by any user, and acquiring a target commodity item corresponding to the access event;

inquiring all commodity combinations containing the target commodity item from the high-quality combination list to form a candidate list;

acquiring actual measurement scores of all commodity combinations in the candidate list, and determining the commodity combination with the highest actual measurement score as a target commodity combination;

and constructing commodity information of all other commodity items except the target commodity item contained in the target commodity combination and pushing the commodity information to the arbitrary user.

According to another aspect of the present application, there is provided a commodity composite construction apparatus including:

the data processing module is used for acquiring a commodity set of each user, wherein the commodity set comprises a plurality of commodity items visited by the user, randomly combining the commodity items in the commodity set, enumerating all commodity combinations to form a user combination set, and de-overlapping all the user combination sets to form a total combination set;

the frequency determining module is set to determine the occurrence frequency of each commodity combination in the full combination set in each user combination set;

the aggregation degree determining module is set to determine the aggregation degree of the commodity combination based on a mutual information entropy obtained by carrying out any two segmentations on each commodity combination in the full quantity combination set, and the aggregation degree characterizes the collocation stability degree of a plurality of commodity items included in the corresponding commodity combination which are accessed by the same user in a relevant manner;

the freedom degree determining module is used for determining the freedom degree of each commodity combination relative to the adjacent entropy of the commodity sets of all users based on the adjacent entropy of each commodity combination in the total combination set, and the freedom degree represents the collocation abundance degree of the corresponding commodity combination which is matched with other commodity items and is associated and accessed by the same user;

and the list construction module is used for screening out a plurality of commodity combinations in the total combination set based on the appearance frequency, the condensation degree and the freedom degree to form a high-quality combination list.

According to another aspect of the present application, there is provided a product portfolio construction apparatus comprising a central processing unit and a memory, the central processing unit being configured to invoke execution of a computer program stored in the memory to perform the steps of the product portfolio construction method described in the present application.

According to another aspect of the present application, a non-volatile readable storage medium is provided, which stores a computer program implemented according to the method for building a product portfolio in the form of computer readable instructions, and when the computer program is called by a computer, the steps included in the method are executed.

According to another aspect of the present application, there is provided a computer program product comprising computer programs/instructions which, when executed by a processor, implement the steps of the method described in any one of the embodiments of the present application.

The present application achieves various technical advantages over the prior art, including but not limited to:

firstly, the commodity set formed by commodity items accessed by the history of all users is subjected to data processing, all commodity combinations in the commodity set are exhausted according to any quantity to form a user combination set corresponding to each user, then all user combination sets are subjected to de-overlapping and merging to obtain a full quantity combination set, then the data characteristics of each commodity combination in the full quantity combination set relative to the user combination set and the commodity set of each user are considered to obtain the corresponding statistical characteristics of the appearance frequency, the cohesion degree, the freedom degree and the like of each commodity combination, and then part of commodity combinations are selected according to the statistical characteristics to form a high-quality combination list, the quantity of the commodity items of each commodity combination is not limited by the technology, the commodity combinations in the high-quality combination list are obtained in an optimized mode under the effect of common statistical characteristics in multiple aspects, the commodity combinations are more corresponding to the global data, the association access characteristics among the commodity items can be comprehensively reflected, and the commodity items have higher combination value.

Secondly, the popularity of each commodity combination is expressed by counting the occurrence frequency of the commodity combination relative to global data, the fixed collocation and stability among a plurality of commodity items is expressed by counting the degree of agglomeration, the collocation and richness expressed by the commodity combination and other commodity items which can be widely collocated are expressed by counting the degree of freedom, the dimensionality is rich, the information value quantization is accurate and effective, the overall information value of each commodity combination is more accurate, and the information quality of the commodity combination which is preferably selected according to the statistical characteristics is higher.

In addition, the method and the device can be applied to the independent station of the E-commerce platform, adapt to the characteristic that historical data of the independent station are relatively less, determine a high-quality combination list, provide commodity combination recommendation for an access user of the independent station, and realize cold start of the commodity recommendation, so that the method and the device have high practical value.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic network architecture diagram of an application environment according to the present invention;

FIG. 2 is a schematic flow chart diagram illustrating one embodiment of a method for building a product portfolio of the present application;

FIG. 3 is a schematic flow chart illustrating data processing performed on user access behavior data according to an embodiment of the present application;

FIG. 4 is a schematic flow chart illustrating the process of determining the degree of cohesion of a combination of articles according to an embodiment of the present application;

FIG. 5 is a schematic flow chart illustrating the process of determining the degrees of freedom of the combinations of commodities in the embodiment of the present application;

FIG. 6 is a schematic flow chart illustrating advertisement promotion by applying a high-quality combined list according to an embodiment of the present application;

FIG. 7 is a schematic block diagram of the merchandise assembly construction apparatus of the present application;

fig. 8 is a schematic structural diagram of a commercial product combined construction apparatus used in the present application.

Detailed Description

The models cited or possibly cited in the application comprise a traditional machine learning model or a deep learning model, and unless specified in clear text, the models can be deployed in a remote server and remotely called at a client, and can also be deployed in a client with qualified equipment capability to be directly called.

Referring to fig. 1, a network architecture adopted by an exemplary application scenario in the present application includes a terminal device 80, an independent station 81, and an application server 82, where the application server 82 may be used to deploy the commodity recommendation service in the present application. The independent station 81 may be used to deploy and open an online store for e-commerce services. The user on the terminal device 80 can browse the commodity information of the commodity item of the independent station 81 in the page of the online shop, the independent station 81 can call the commodity recommendation service of the application server 82 based on the target commodity item currently accessed by any user, the commodity recommendation service determines other commodity items matched with the target commodity item according to the commodity combination in the pre-customized high-quality combination list, the other commodity items are returned to the independent station 81, and the independent station 81 recommends the commodity information of the other commodity items to the terminal device 80 for further browsing of the user, so that the purpose of commodity recommendation is achieved.

The commodity combinations in the high-quality combination list can be determined by instructions obtained after a computer program product programmed according to the commodity combination construction method of the present application is run, and related instructions execute the commodity combination construction method based on various information provided by the independent station 81, including user access behavior data, various commodity information of commodity items, and the like, so as to really generate high-quality commodity combinations, construct the high-quality combination list, and store the high-quality combination list corresponding to the independent station 81 for the commodity recommendation service to call.

It should be noted that the computer program product implemented according to the product portfolio construction method code of the present application can be run in any computer device, including but not limited to the independent station 81, the application server 82 or other servers, as long as it can call up various basic data required for constructing a good portfolio list to the independent station 81.

With reference to the above disclosed principles, referring to fig. 2, a method for constructing a commercial product combination according to the present application, in one embodiment, comprises the following steps:

step S1100, acquiring a commodity set of each user, wherein the commodity set comprises a plurality of commodity items visited by the user, combining the commodity items in the commodity set by any number, enumerating all commodity combinations to form a user group set, and de-overlapping all the user group sets to form a total combination set;

in an exemplary application scenario, a list of good combinations required by the product recommendation service is generated for an independent station, and a plurality of good product combinations are provided in the list. Accordingly, the high-quality combination list can be constructed using the user visiting behavior data generated in the on-line shop of the independent station as basic data.

The user access behavior data generally refers to various historical data generated corresponding to various access behaviors of the user, and the historical data can be obtained by setting a buried point code in a page accessed by the user, for example, for the order placing and purchasing behavior of the user, a corresponding order contains commodity items purchased by the user, and thus mapping relation data between the user and the commodity items purchased by the user is obtained. Similarly, for the behavior of adding the commodity item to the shopping cart by the user, the behavior of browsing the corresponding commodity items by the user during the same session, and the like, corresponding behavior data is generated in the independent station. For each type of behavior data, one or more types of behavior data may be selected as the base data required to construct a list of good portfolio. For example, in the following texts, behavior data corresponding to the ordering and purchasing behavior of the user is determined as the basic data for understanding.

In an embodiment, the time of the access behavior data that needs to be obtained may be restricted, for example, the time may be access behavior data obtained by obtaining a current time and tracing back to a preset time length, where the preset time length may be any value such as three days, one week, one month, one quarter, and half a year, and the access behavior data may reflect the latest dynamics of a commodity combination through the time restriction.

For the same access behavior, mapping relation data such as access time, access users, accessed commodity items and the like exist in behavior data accessed by each user in history, so that commodity items accessed by each user can be extracted to form a commodity set corresponding to the user, and the commodity set contains all commodity items accessed by the user in history. Considering that the same commodity item may appear in multiple historical behavior data at the same time, it may be deduplicated as needed.

It will be understood that each user can obtain its corresponding commodity set, which contains a plurality of commodity items that the user has visited, constituting the basic data.

Further, for each commodity item in the commodity set of each user, according to any exhaustive number of terms, wherein the number of terms is more than or equal to 2, selecting a plurality of commodity items with corresponding number of terms to combine to form a commodity combination. For example, when there are three merchandise items { a; b; c, four combinations of goods can be obtained, i.e. { AB; BC; AC; ABC, and so on, through enumeration, various possible commodity combinations corresponding to each user are obtained. After all commodity combinations are enumerated for each commodity set, a user combination set is obtained, wherein all commodity combinations corresponding to the user are contained.

The user combination sets of different users may have the same commodity combination, all the user combination sets can be combined into the same full combination set for facilitating subsequent data processing, and the duplicate removal processing can be performed on the same commodity combination in different user combination sets, so that the same commodity combination only appears once in the full combination set.

S1200, determining the occurrence frequency of each commodity combination in the full combination set in each user combination set;

the sum of the times of occurrence of each commodity combination in each user combination set, that is, the frequency of occurrence, may indicate the popularity of the commodity combination in the historical visiting behavior, and the higher the frequency of occurrence, the higher the frequency of historical use of the commodity combination by the user is theoretically. For this purpose, for each combination of products in the total number of combinations, the frequency of occurrence V can be determined first ₀ As its first statistical feature.

When the occurrence frequency of each commodity combination in the total combination set is calculated, the total occurrence frequency of the commodity combination in all the user combination sets of all the users is counted, and the corresponding occurrence frequency is determined. Accordingly, in the total combination set, each commodity combination can obtain the corresponding occurrence frequency.

Step S1300, determining the degree of cohesion of each commodity combination based on a mutual information entropy obtained by carrying out any two segmentations on each commodity combination in the full quantity combination set, wherein the degree of cohesion characterizes the degree of stability of collocation of a plurality of commodity items included in the corresponding commodity combination by the associated access of the same user;

in order to measure the collocation stability of a fixed collocation formed by multiple commodity items in each commodity combination and accessed by a user in an associated manner, the compactness among the multiple commodity items is actually characterized, and the compactness can be determined by considering the mutual information entropy of the commodity combinations.

The commodity combination is composed of commodity items, each commodity item is a fixed information object, therefore, when the mutual information entropy is calculated, for any commodity combination needing to calculate the mutual information entropy, the commodity combination can be subjected to multiple times of two segmentation, each time the commodity combination is segmented into two parts, and then the obtained mutual information entropy under each segmentation condition is calculated according to a mutual information entropy formula.

The formula for calculating the mutual information entropy is as follows:

wherein:

p () is the access behavior corresponding to the corresponding commodity combination, and the occurrence probability of the commodity combination appearing in all behavior data corresponding to the access behavior of the entire independent station can be determined by dividing the frequency of the commodity combination appearing in all behavior data corresponding to the access behavior by the total number of behavior data corresponding to the access behavior.

P (com) is the probability of occurrence of a combination of items in the total number of combinations set, P (com) ₁ ) Is the probability of occurrence of the commodity combination of the left part obtained by the two-division of the commodity combination, P (com) ₂ ) Is the occurrence probability of the commodity combination of the right part obtained by the two-point segmentation of the commodity combination.

It should be noted that after a commodity combination is cut twice, the left part and/or the right part of the commodity combination may be a single commodity item, which is allowed as well, and the calculation of the mutual information entropy is not affected.

In one embodiment, for convenience of measurement, the mutual information entropy may be logarithmized to map its range to a smoother range of values. It is understood that when P (com) and P (com) are used ₁ )*P(com ₂ ) At the same order of magnitude, V ₁ Close to 1, and the logarithm is taken to be 0, which corresponds to a very low mutual information entropy. Taking the product combination of two product items as an example, if the mutual information entropy of the product combination is very low, which indicates that the correlation between the first product item and the second product item is very weak, even irrelevant, then P (com) and P (com) are provided ₁ )*P(com ₂ ) Will be in the same order of magnitude. On the contrary, if the mutual information entropy of the commodity combination is very high and the probabilities of the commodity items "A", "B" and "AB" are very close, V is ₁ Will be a value much greater than 1.

According to the above process, after each commodity combination is subjected to any two divisions for multiple times, the mutual information entropy corresponding to each two divisions can be obtained, so as to obtain multiple mutual information entropies. In another embodiment, the mutual information entropies corresponding to the multiple second partitions may be averaged, and the average value of the mutual information entropies is determined as the degree of aggregation. Since the degree of aggregation is essentially mutual information entropy, it is not difficult to understand that the degree of aggregation represents the stability of the fixed collocation accessed by the association of the fixed collocation constituted by each commodity item in the corresponding commodity combination, i.e., the higher the degree of aggregation, the higher the stability of the collocation constituted by each commodity item in the commodity combination; the lower the degree of aggregation, the lower the degree of stability of the arrangement.

According to the above principle, each commodity combination in the total combination set can obtain its corresponding degree of aggregation, so as to obtain the second statistical characteristic corresponding to each commodity combination.

Step S1400, determining the degree of freedom of each commodity combination relative to the commodity sets of all users based on the adjacent entropy of each commodity combination in the full combination set, wherein the degree of freedom represents the collocation abundance degree of the corresponding commodity combination which is matched with other commodity items and is accessed by the same user in a relevant manner;

as described above, the commodity set of each user is formed by extracting commodity items from the behavior data corresponding to the access behavior of the corresponding user, and each behavior data carries corresponding time information, so that the commodity sets of the users can be sorted by using the time information, so that the sorted commodity set is used as reference data for determining the degree of freedom of each commodity combination in the full-quantity combination set.

In one embodiment, the method is used to refer to a commodity set for determining degrees of freedom, wherein even if the same commodity item exists in the behavior data at different times, the same commodity item can be retained without deduplication, so that when next entropy is calculated for determining degrees of freedom, the fact data characteristics reflected by the more original behavior data can be directly used as the basis. Of course, it is also feasible to deduplicate the same merchandise items in the set at different times, keeping the time up-to-date therein.

In order to determine the degree of freedom of each commodity combination in the total combination set, the commodity sets of all users sorted according to time information may be referred to first, the adjacent entropy of each commodity combination in the commodity sets is calculated first, and then the degree of freedom corresponding to the commodity combination is determined according to the summary of the adjacent entropy of the commodity combination relative to all users.

Specifically, for each commodity combination in the total combination set, according to the commodity set ordered by time information of each user, all commodity items appearing on the left side of the commodity combination and all commodity items appearing on the right side of the commodity combination are determined, and the positions of the commodity combinations appearing in the commodity set are only treated by dividing left and right sides, then the left adjacent entropy of the left side set formed by all commodity items on the left side is calculated, and similarly, the right adjacent entropy of the right side set formed by all commodity items on the right side is calculated. The formula for the adjacency entropy is as follows:

V _cal ＝-∑pi*log(pi)

where pi is the probability of occurrence of each item in its respective side set, e.g., the left side set or the right side set, and may be the number of occurrences divided by the total number of items in the set in which it is located.

It can be seen that whether the left adjacent entropy or the right adjacent entropy is smaller, the lower the degree of freedom of the combination is, the more frequently and fixedly the combination is matched with some other commodity items; when the obtained numerical value is larger, the higher the degree of freedom is, the richer, diversified and even disordered matching relationship between the commodity combination and other commodity items is represented, so that the left adjacent entropy and the right adjacent entropy can both represent the degree of freedom, and the matching richness degree of the corresponding commodity combination matched with other commodity items and accessed by the same user in a related manner is represented.

It is understood that the left adjacent entropy and the right adjacent entropy determined by the same commodity combination relative to the commodity set of the same user are not necessarily consistent, and for this reason, in one embodiment, the lowest of the left adjacent entropy and the right adjacent entropy, i.e. the lowest adjacent entropy, can be selected as the basic data for determining the degree of freedom of the commodity combination, which makes the measure of the degree of freedom more conservative. In another embodiment, the average value of the left adjacent entropy and the right adjacent entropy can be determined as the lowest adjacent entropy to be used as basic data for determining the degree of freedom of the commodity combination.

It can be seen from the above that, for each product combination in the total combination set, the lowest adjacent entropy corresponding to each user can be determined with respect to the product set of each user, and the lowest adjacent entropies corresponding to all users are summed and averaged to obtain a corresponding numerical value, which can be used as the degree of freedom V for the final use of the product combination ₂ . The degree of freedom of each commodity combination is determined according to the process, and the third statistical characteristic of each commodity combination is determined.

In another alternative to the above embodiment of determining the degree of freedom, the commodity sets of all users may be merged into a full commodity set, then the commodity items in the full commodity set are sorted according to the time information of the behavior data, then the left adjacent entropy and the right adjacent entropy of each commodity combination in the full commodity set are calculated based on the full commodity set, and the corresponding lowest adjacent entropy is determined and is directly used as the degree of freedom, which is the same as the above embodiment.

And S1500, screening out a plurality of commodity combinations in the total combination set based on the appearance frequency, the cohesion and the degree of freedom to form a high-quality combination list.

The frequency V of occurrence of each commodity combination in the full combination set has been determined through the above process ₀ Degree of coagulation V ₁ And a degree of freedom V ₂ The appearance frequency represents the popularity of the commodity combination, the cohesion represents the stable matching degree of a plurality of commodity items in the commodity combination, the freedom represents the matching abundance corresponding to the matching of the commodity combination and other commodity items, and the quantitative standards of the commodity combination are provided from different dimensions respectivelyIn one embodiment, the quantified values of the different dimensions may be fused, for example, the occurrence frequency, the aggregation degree, and the degree of freedom are directly multiplied or weighted and summed, so as to determine the comprehensive Score corresponding to each commodity combination, and then, according to the comprehensive Score, a part of the commodity combinations with relatively high comprehensive scores may be selected as good-quality commodity combinations to form a good-quality combination list for the commodity recommendation service to call.

In another embodiment, a grid search mode may be adopted, the initial threshold values corresponding to the occurrence frequency, the degree of aggregation, and the degree of freedom are first set, then parameter-by-parameter tuning is performed, the screening threshold values corresponding to the occurrence frequency, the degree of aggregation, and the degree of freedom are sequentially determined, then the screening threshold values are used to screen the commodity combinations in the total combination set, so as to determine the high-quality commodity combinations therein, and the high-quality combination list is constructed.

When the high-quality combination list is determined from the full-quantity combination set, the selection of the full-quantity commodity combination is realized, and the selected commodity combination has high popularity, high collocation stability and abundant collocation degree, and is obviously a high-quality commodity combination. When the method is used as the basis of the collocation scheme of the commodity recommendation service, the dependence on historical data can be avoided, effective associated popularization information is provided for the commodity recommendation strategy, the promotion of advertisement success indexes is facilitated, and good commodity recommendation success is obtained.

From the above embodiments, it can be seen that the present application achieves various technical advantages, including but not limited to:

On the basis of any embodiment of the present application, please refer to fig. 3, in which a commodity set of each user is obtained, where the commodity set includes a plurality of commodity items visited by the user, the commodity items in the commodity set are combined by any number of items, all commodity combinations are enumerated to form a user combination set, and all user combination sets are de-overlapped to form a full combination set, including:

step S1110, obtaining access behavior data generated by each user in the same online shop in historical access events, and determining a plurality of commodity items accessed by the user from the access behavior data to form a commodity set of the user, wherein the historical access events comprise an event of adding the commodity items to a shopping cart and/or an event of adding the commodity items to a purchase order;

in this embodiment, it is mainly considered that the visit behavior data of the user is collected based on the same online store, so that the obtained high-quality combination list is more suitable for the commodity recommendation service serving the independent station where the online store is located. In this way, the online stores based on the independent stations are considered, when the independent stations are crossed, the commodity items sold by the online stores are likely to be greatly different, and the necessary association is lacked among the behavior data associated with the commodity items, so that the acquisition of the visiting behavior data is limited to the same online store, and the obtained high-quality combination list is more targeted.

Therefore, access behavior data corresponding to a certain access behavior event generated by the online shop within a specific time range can be acquired from a user behavior database of an independent station where the online shop is located, then, the access behavior data of each user is analyzed, access time and commodity items in the access behavior data are extracted, then, all commodity items historically accessed by each user are constructed into the same commodity set by taking the user as a unit, and preferably, the commodity items in the commodity set are sorted according to the access time.

In the embodiment of the present application, the event of adding the commodity item to the shopping cart and/or the event of adding the commodity item to the purchase order are recommended, the former is mainly because the strong desire of the user to purchase the corresponding commodity item is expressed, and the latter is mainly because the fact action of the user to implement the purchase is expressed, so that both can express the requirement of the user to purchase the corresponding commodity item, and thus the access behavior event has the information reference value required as basic data.

Step S1120, for each commodity set of each user, combining the commodity items in the commodity set according to any plurality of commodity items, enumerating all possible commodity combinations in the commodity set, and forming a user combination set;

the commodity set of each user may include a large number of commodity items, and a commodity combination can be constructed by combining any two or more commodity items with any number.

And S1130, performing superposition elimination on commodity combinations in the user combination set of all the users to form a full-quantity combination set.

Although no duplicate commodity combination exists in each user combination set, duplicate commodity combinations may exist in different user combination sets, and in consideration of the requirement of evaluating each unique commodity combination, a full-weight combination set can be constructed, all the commodity combinations in the user combination sets of all the users are added into the full-weight combination set, and only one of the commodity combinations is reserved for the identical commodity combination, so that the duplicate removal is realized, and each commodity combination in the full-weight combination set has uniqueness.

The embodiment realizes the deep data processing of the access behavior data of all users in the same online shop, constructs the user combination set and the full combination set, wherein the user combination set comprises enumeration of all possible commodity combinations in the historical behavior data, and the number of commodity items in the commodity combinations is not additionally limited, so that the subsequent commodity combination evaluation in the full combination set can be orderly and efficiently operated, and the high-quality combination list can be quickly and efficiently constructed by the optimal operation amount.

On the basis of any embodiment of the present application, please refer to fig. 4, determining the aggregation level of each product combination based on the mutual information entropy obtained by performing any two segmentations on each product combination in the full-amount combination set includes:

step 1310, aiming at each commodity combination in the total quantity combination set, randomly dividing the commodity combination into two parts by taking a commodity item as a unit to obtain a left side set and a right side set corresponding to each division;

each specific step of the present embodiment may be executed by performing a separate process for each combination of commodities in the total combination set, so as to determine the corresponding degree of aggregation.

For a commodity combination, the commodity combination is composed of a plurality of commodity items, so that the commodity items can be represented as PI ds corresponding to the commodity items in the commodity combination, and accordingly, the commodity combination can be subjected to two segmentation according to any one commodity item as a boundary, that is, the commodity combination is segmented into a left part and a right part, and a left side set and a right side set are respectively obtained, wherein the left side set and the right side set both at least contain more than one commodity item, that is, are not empty sets. That is, for the combination of commodities, in this step, all possible bipartition schemes are exhausted, and a left side set and a right side set corresponding to each segmentation are obtained.

Step S1320, calculating mutual information entropy corresponding to each segmentation based on the left side set and the right side set obtained by each segmentation;

for the left side set and the right side set corresponding to each segmentation of each commodity combination, the mutual information entropy of the commodity combination can be calculated based on the occurrence probability of the commodity combination itself and the occurrence probabilities of the left side set and the right side set, and a specific calculation manner has been disclosed in the foregoing embodiment, which is not repeated here. Therefore, each segmentation of one commodity combination can obtain the corresponding mutual information entropy.

Step S1330, the lowest mutual information entropy obtained after each commodity combination is segmented for multiple times is used as the aggregation level of the commodity combination.

In this embodiment, the lowest mutual information entropy is used as the aggregation of the commodity combination, and the collocation stability of the fixed collocation of each commodity item in the commodity combination is expressed in a conservative manner.

According to the embodiment, the mutual information entropy corresponding to each segmentation is determined by performing multiple secondary segmentation on each commodity combination in the full-scale combination set, the lowest information entropy is used for representing the degree of aggregation of the commodity combination, the collocation stability degree of each commodity item in the whole commodity combination for forming fixed collocation can be effectively represented, it is easy to understand that the higher the degree of aggregation of the commodity combination is, the more the commodity items in the commodity combination are used as stable commodity combinations, the more the combination is stable, and the reference information value of each commodity combination in the aspect of collocation stability degree is effectively represented according to historical behavior data.

Based on any embodiment of the present application, please refer to fig. 5, determining the degree of freedom of each commodity combination in the full amount combination set based on the adjacency entropy of the commodity combination relative to the commodity sets of all users includes:

step S1410, sorting the commodity items in the commodity set of each user according to the user access time;

because the commodity items in the commodity set of each user are extracted with the corresponding access behavior data, and the access behavior data has the time information corresponding to the user access behavior, the commodity set can be sorted according to the user access time in the access behavior data, so that the commodity set can show the access sequence of each commodity item more orderly. It will be understood that the item set may have a plurality of items that make up the item combination arranged together, and thus, in fact, there will still be a representation of the plurality of item combinations by their positional adjacency.

Step S1420, aiming at each commodity combination in the full quantity combination set, calculating the left adjacent entropy and the right adjacent entropy of the commodity combination which is sequenced relative to each user, and determining the lowest value as the lowest adjacent entropy;

for each commodity combination in the full combination set, the collocation richness degree corresponding to the commodity combination collocated with other commodity items and accessed by the user in an associated manner can be represented by calculating the adjacency entropy thereof, that is, when the adjacency entropy of one commodity combination is lower, the commodity combination is collocated with a smaller amount of other commodity items and accessed by the user in an associated manner, and when the adjacency entropy is higher, the commodity combination is more collocated with other commodity items and accessed by the user in an associated manner. The calculation methods of the left adjacent entropy and the right adjacent entropy have been given in the foregoing, and are not repeated here. For a commodity combination, after calculating the left adjacent entropy and the right adjacent entropy, in this embodiment, the lowest value of the left adjacent entropy and the right adjacent entropy is selected as the lowest adjacent entropy for representing the collocation abundance degree of the commodity combination relative to the corresponding user.

And step S1430, summarizing the lowest adjacent entropy of each commodity combination relative to all users as the freedom degree of the commodity combination.

According to the above process, for all users, each commodity combination can obtain the lowest adjacent entropy corresponding to each user, in order to convert the lowest adjacent entropy into the representation of the degree of freedom, the lowest adjacent entropy of each user can be added and averaged to realize summary, and the numerical value obtained by the summary is used as the degree of freedom of the commodity combination. Accordingly, each commodity combination in the total combination set can obtain the corresponding degree of freedom.

The embodiment provides a specific calculation mode of the degree of freedom, in the calculation process, the granularity of the commodity set of each user is refined, the degree of freedom is summarized after the lowest adjacent entropy is determined based on each user, and the method can be refined to the level of each user, so that the obtained degree of freedom can more accurately represent the collocation abundance degree of each commodity combination, and the high-quality commodity combination can be more accurately and preferably selected subsequently.

On the basis of any embodiment of the application, screening out a plurality of commodity combinations in the total combination set based on the appearance frequency, the condensation degree and the degree of freedom to form a high-quality combination list, which comprises the following steps:

step S1510, multiplying the appearance frequency, the cohesion degree and the freedom degree of each commodity combination in the total combination set as a comprehensive score corresponding to the commodity combination;

in this embodiment, the frequency of occurrence V is used for each product combination in the total number of combinations ₀ Degree of coagulation V ₁ And a degree of freedom V ₂ Summarizing, determining a comprehensive Score corresponding to each commodity combination, and realizing comprehensive quantification, wherein the adopted formula is as follows:

Score＝V ₀ *V ₁ *V ₂

according to the formula, the appearance frequency, the condensation degree and the freedom degree are strongly correlated to determine the comprehensive score of the commodity combination, so that the comprehensive score can effectively represent the overall quality of the commodity combination, the comparison effect based on the comprehensive score among the commodity combinations is more prominent, and the screening is convenient.

In a further embodiment, when the above formula is applied, the occurrence frequency, the degree of aggregation, and the degree of freedom may be matched with corresponding preset weights as needed, and those skilled in the art may flexibly set the weights as needed.

And step S1520, screening the commodity combinations in the total combination set based on the comprehensive scores to obtain part of commodity combinations with relatively higher comprehensive scores to form a high-quality combination list.

After the comprehensive score corresponding to each commodity combination in the total combination set is determined, part of commodity combinations with higher comprehensive scores can be screened out from the total combination set based on the comprehensive score to serve as high-quality commodity combinations, and a high-quality combination list is formed.

In one embodiment, a preset screening threshold may be adopted, each composite score in the total combination set is compared with the total combination set, and the commodity combination with the composite score higher than the screening threshold is used as a high-quality commodity combination.

In another embodiment, a preset screening number may be adopted, all commodity combinations in the total combination set are sorted from large to small according to the comprehensive scores, and then the commodity combinations with the corresponding number in the top sorting are intercepted according to the screening number to serve as the high-quality commodity combinations.

According to the embodiments, the occurrence frequency, the degree of agglomeration and the degree of freedom are strongly correlated to determine the comprehensive score of each commodity combination, and then the comprehensive score is used for screening the high-quality commodity combinations.

On the basis of any embodiment of the present application, please refer to fig. 6, which illustrates the following steps after forming the list of good combinations:

step S1600, responding to an access event triggered by any user, and acquiring a target commodity item corresponding to the access event;

and after the high-quality combination list is determined, the commodity recommendation service can be called. The merchandise recommendation service may respond to an access event triggered by any user accessing an online store, where the access event may be, for example, a user adding a target merchandise item to a shopping cart, entering a merchandise detail page of the target merchandise item, adding the target merchandise item to an order, completing a payment operation for the order in which the target merchandise item is located, and the like, and thus, in response to the access event, a target merchandise item corresponding to the access event may be obtained.

Step S1700, all commodity combinations containing the target commodity item are inquired from the high-quality combination list to form a candidate list;

after obtaining the target commodity item, the commodity recommendation service can query all commodity combinations including the target commodity item from the high-quality combination list, and the commodity combinations not only include the target commodity item but also include other commodity items forming habitual collocation relations with the target commodity item without difficulty in understanding. For the convenience of subsequent operation, all the commodity combinations matched from the high-quality combination list based on the target commodity item, namely all the high-quality commodity combinations are constructed into a candidate list.

Step S1800, actual measurement scores of all commodity combinations in the candidate list are obtained, and the commodity combination with the highest actual measurement score is determined to be used as a target commodity combination;

in the candidate list, a plurality of commodity combinations may exist, and when pushing is performed to the user each time, a plurality of strategies may be adopted to select other matched commodity items, and in this case, each commodity combination needs to be further preferentially determined to determine a target commodity combination, so that the other commodity items are accurately pushed according to the target commodity combination. One exemplary strategy is to select the most effective commodity combination as the target commodity combination as the basis for determining other commodity items that can be matched with the target commodity item. Therefore, the actual measurement scores corresponding to the commodity combinations in the candidate list can be obtained according to a preset mechanism, and then the commodity combination with the highest actual measurement score is used as the target commodity combination.

Regarding the preset mechanism, in an embodiment, each commodity combination in the high-quality combination list may be randomly applied to a user of the full platform in advance to implement an advertisement behavior, and then a conversion rate obtained after the advertisements are combined with the commodities is used as a corresponding actual measurement score, in this case, each commodity combination in the candidate list may be screened according to the corresponding actual measurement score, and a commodity combination with the highest actual measurement score among the commodity combinations is determined as a target commodity combination.

Step S1900, constructing commodity information of all other commodity items except the target commodity item included in the target commodity combination, and pushing the commodity information to the arbitrary user.

The target commodity combination determined through the above process includes the target commodity item and other commodity items which are habitually collocated with the target commodity item, so that commodity information of all the other commodity items, including but not limited to commodity titles, commodity prices, commodity pictures, commodity links and the like of the corresponding commodity items, can be acquired, the commodity information of the other commodity items is packaged into a commodity recommendation list, and the commodity recommendation list is pushed to the arbitrary user which triggers the access event. After the terminal device of the arbitrary user receives the commodity recommendation list, the information control corresponding to each commodity item is analyzed and packaged by the browser of the arbitrary user to be displayed, wherein for each other commodity item, the commodity link is associated, and information such as a commodity image, a commodity price and a commodity title of the arbitrary user is displayed according to needs.

According to the embodiments, the high-quality combination list obtained by the method can provide high-quality experience data for the commodity recommendation service, and the high-quality experience data can be used as the basis of advertisement decision, so that the commodity combination recommendation strategy is more effective, and cold-start advertisement promotion can be realized for new users.

Referring to fig. 7, a merchandise combination building apparatus according to an aspect of the present application includes a data processing module 1100, a frequency determining module 1200, an aggregation determining module 1300, a degree of freedom determining module 1400, and a list building module 1500, wherein: the data processing module 1100 is configured to acquire a commodity set of each user, where the commodity set includes a plurality of commodity items visited by the user, combine the commodity items in the commodity set by any number, enumerate all commodity combinations to form a user combination set, and de-overlap all user combination sets to form a full-volume combination set; the frequency determining module 1200 is configured to determine the occurrence frequency of each commodity combination in the total combination set appearing in each user combination set; the aggregation degree determining module 1300 is configured to determine the aggregation degree of each commodity combination based on a mutual information entropy obtained by performing any two segmentations on each commodity combination in the full-scale combination set, wherein the aggregation degree characterizes the collocation stability degree of multiple commodity items included in the corresponding commodity combination and accessed by the same user in an associated manner; the degree-of-freedom determination module 1400 is configured to determine the degree of freedom of each commodity combination in the full-scale combination set with respect to the adjacent entropy of the commodity sets of all users, where the degree of freedom represents the collocation abundance degree that the corresponding commodity combination is collocated with other commodity items and is accessed by the same user in an associated manner; the list construction module 1500 is configured to screen out a plurality of combinations of the commodities in the total number of combinations based on the appearance frequency, the degree of aggregation, and the degree of freedom to construct a high-quality combination list.

On the basis of any embodiment of the present application, the data processing module 1100 includes: the behavior data calling unit is used for acquiring access behavior data generated by each user in the same online shop in historical access events, determining a plurality of commodity items accessed by the user from the access behavior data to form a commodity set of the user, wherein the historical access events comprise an event of adding the commodity items to a shopping cart and/or an event of adding the commodity items to a purchase order; the combination enumeration unit is set to combine the commodity items in the commodity set of each user according to a plurality of random commodity items, enumerate all possible commodity combinations in the commodity set and form a user combination set; and the user combination unit is used for de-overlapping the commodity combinations in the user combination set of all the users and constructing the commodity combinations into a full-quantity combination set.

On the basis of any embodiment of the present application, the cohesion determination module 1300 includes: the combined segmentation unit is set to divide each commodity combination in the full-scale combined set into any possible two segments by taking a commodity item as a unit, and a left side set and a right side set corresponding to each segment are obtained; the information quantization unit is used for calculating mutual information entropy corresponding to each segmentation based on the left side set and the right side set obtained by each segmentation; and the condensation degree selecting unit is set to use the lowest mutual information entropy obtained after each commodity combination is segmented for multiple times as the condensation degree of the commodity combination.

On the basis of any embodiment of the present application, the degree of freedom determining module 1400 includes: the time sequencing unit is used for sequencing the commodity items in the commodity set of each user according to the user access time; the adjacency entropy calculation unit is used for calculating left adjacency entropy and right adjacency entropy of each commodity combination in the full combination set relative to each commodity set ordered by the user, and determining the lowest value of the left adjacency entropy and the right adjacency entropy as the lowest adjacency entropy; and the freedom degree summarizing unit is arranged to summarize the lowest adjacent entropy of each commodity combination relative to all the users as the freedom degree of the commodity combination.

On the basis of any embodiment of the present application, the list construction module 1500 includes: the fusion scoring unit is set to multiply the occurrence frequency, the agglomeration degree and the freedom degree of each commodity combination in the total combination set as the comprehensive score corresponding to the commodity combination; and the screening construction unit is arranged for screening the commodity combinations in the full combination set based on the comprehensive scores to obtain part of commodity combinations with relatively high comprehensive scores to form a high-quality combination list.

On the basis of any embodiment of the present application, the commodity combined construction device of the present application further includes: the access response module is set to respond to an access event triggered by any user and acquire a target commodity item corresponding to the access event; the combination query module is used for querying all commodity combinations containing the target commodity item from the high-quality combination list to form a candidate list; the reference preference module is used for acquiring the actual measurement scores of all commodity combinations in the candidate list and determining the commodity combination with the highest actual measurement score as a target commodity combination; and the result pushing module is arranged to construct commodity information of all other commodity items except the target commodity item contained in the target commodity combination to be pushed to the arbitrary user.

Another embodiment of the present application also provides a merchandise assembly construction apparatus. As shown in fig. 8, the internal structure of the commodity combination construction apparatus is schematically illustrated. The commodity building apparatus includes a processor, a computer-readable storage medium, a memory, and a network interface connected by a system bus. The computer readable non-volatile readable storage medium of the commodity combined construction device stores an operating system, a database and computer readable instructions, the database can store information sequences, and the computer readable instructions can cause a processor to realize a commodity combined construction method when being executed by the processor.

The processor of the commodity combined construction equipment is used for providing calculation and control capability and supporting the operation of the whole commodity combined construction equipment. The memory of the article portfolio construction device may have stored therein computer readable instructions that, when executed by the processor, may cause the processor to perform the article portfolio construction method of the present application. The network interface of the commodity combined construction equipment is used for connecting and communicating with the terminal.

It will be understood by those skilled in the art that the structure shown in fig. 8 is a block diagram of only a portion of the structure relevant to the present application, and does not constitute a limitation on the article composite construction apparatus to which the present application is applied, and that a particular article composite construction apparatus may include more or fewer components than shown in the drawings, or may combine certain components, or have a different arrangement of components.

In this embodiment, the processor is configured to execute specific functions of each module in fig. 7, and the memory stores program codes and various data required for executing the modules or sub-modules. The network interface is used for realizing data transmission between user terminals or servers. The non-volatile readable storage medium in the present embodiment stores program codes and data necessary for executing all modules in the product assembly structure device of the present application, and the server can call the program codes and data of the server to execute the functions of all modules.

The present application also provides a non-transitory readable storage medium storing computer-readable instructions which, when executed by one or more processors, cause the one or more processors to perform the steps of the method of building a combination of articles of any of the embodiments of the present application.

The present application also provides a computer program product comprising computer programs/instructions which, when executed by one or more processors, implement the steps of the method as described in any of the embodiments of the present application.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments of the present application may be implemented by hardware related to instructions of a computer program, which may be stored in a non-volatile readable storage medium, and when executed, may include the processes of the embodiments of the methods as described above. The storage medium may be a computer-readable storage medium such as a magnetic disk, an optical disk, a Read-Only Memory (ROM), or a Random Access Memory (RAM).

To sum up, the method and the device for recommending the commodity combinations screen out the high-quality commodity combinations according to the common constraints of the appearance frequency, the condensation degree and the freedom degree of the commodity combinations, ensure that the determined commodity combinations can more accurately reflect the incidence relations among a plurality of commodity items, and the commodity items in the same commodity combination are not limited by the number, provide accurate commodity combination information for the commodity recommendation service of the independent station, and enable the success rate obtained by commodity recommendation to be higher.

Claims

1. A method of building a combination of articles, comprising:

determining the degree of aggregation of the commodity combinations based on mutual information entropy obtained by performing any two segmentations on each commodity combination in the full-scale combination set, wherein the degree of aggregation characterizes the collocation stability of a plurality of commodity items included in the corresponding commodity combinations accessed by the same user in a related manner;

and screening out a plurality of commodity combinations in the total combination set based on the appearance frequency, the degree of agglomeration and the degree of freedom to form a high-quality combination list.

2. The merchandise combination construction method according to claim 1, wherein acquiring a merchandise set for each user including a plurality of merchandise items accessed by the user, combining the merchandise items in the merchandise set by an arbitrary number, enumerating all the merchandise combinations to form a user combination set, and forming a total combination set by de-superimposing all the user combination sets, comprises:

and the commodity combinations in the user combination set of all the users are subjected to de-coincidence and are constructed into a full-quantity combination set.

3. The commodity combination construction method according to claim 1, wherein determining the degree of aggregation of the commodity combination based on a mutual information entropy obtained by performing any two-segmentation on each commodity combination in a full-amount combination set comprises:

4. The commodity combination construction method according to claim 1, wherein determining the degree of freedom of each commodity combination with respect to commodity sets of all users based on the adjacency entropy of the commodity combination in the total quantity combination set comprises:

calculating left adjacent entropy and right adjacent entropy of each commodity combination in the full combination set relative to each commodity set ordered by the user, and determining the lowest value of the left adjacent entropy and the right adjacent entropy as the lowest adjacent entropy;

and summarizing the lowest adjacent entropy of each commodity combination relative to all users as the freedom degree of the commodity combination.

5. The commodity combination construction method according to any one of claims 1 to 4, wherein a plurality of commodity combinations in the total combination set are selected based on the appearance frequency, the degree of aggregation, and the degree of freedom to form a high-quality combination list, and the method includes:

6. The merchandise portfolio configuration method of any one of claims 1 through 4, after constructing the list of good quality portfolios, comprising:

7. An apparatus for assembling and constructing an article, comprising:

8. A commodity building apparatus comprising a central processor and a memory, wherein the central processor is arranged to invoke execution of a computer program stored in the memory to perform the steps of the method of any one of claims 1 to 6.

9. A non-transitory readable storage medium storing a computer program in the form of computer readable instructions, the computer program when invoked by a computer performing the steps comprised by the method according to any one of claims 1 to 6.

10. A computer program product comprising computer programs/instructions for performing the steps of the method according to any one of claims 1 to 6 when executed by a processor.