CN110992153B - Commodity recommendation method, system and equipment based on user attribute and commodity type - Google Patents

Abstract

The invention provides a commodity recommending method based on user attribute and commodity type, firstly, the method forms a binary relation network of user attribute-commodity type by attribute information of the user and commodity type information purchased by the user, and then calculates association weight, frequency weight, value weight, triggering centrality and network core degree of commodity according to certain attribute information of the user; and then carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger center degree and the weighted network core degree to obtain the optimal node degree of the classified user commodity, and carrying out commodity recommendation on the corresponding classified user according to the optimal node degree of the classified user commodity. The method can obtain the commodity with the most spread and the most influence under a certain specific user attribute, and the commodity pushing is more refined and more accurate.

Description

Commodity recommendation method, system and equipment based on user attribute and commodity type

Technical Field

The present invention relates to the field of data mining, and in particular, to a method, a system, and an apparatus for recommending commodities based on user attributes and commodity types.

Background

The commodity recommendation method in the prior art generally adopts collaborative filtering recommendation technology or click degree centrality technology of a relational network.

Drawbacks of collaborative filtering recommendation are: (1) it depends on recommendations; that is, the user can only purchase a certain amount of commodities (in practical application, purchase can also represent clicking, like, selected, collected, added shopping cart and other like preference selection conditions), and then can have a recommendation target, which belongs to passive recommendation. (2) The recommended merchandise only considers a single purchase (i.e., based on only one exact correlation behavior), and does not consider a secondary derivative purchase, with subsequent derivative recommendations being poor.

The defect of the centrality of the point degree of the relation network is: (1) Only the transmissibility of the commodity (whether the commodity is coupled with many other commodities) is considered, but the influence of the commodity is not considered. Firstly, although a plurality of commodities are related to the commodity, the proportion of the number of times of related simultaneous purchase of the commodity to the total number of times of purchase of the commodity is very low, for example, commodity A is purchased for 10 times in total, 9 times are all purchased once, and only 1 time is used for simultaneously purchasing other commodities; and secondly, although the commodity is connected with a plurality of other commodities, the commodity is low in purchase frequency, is not a popular commodity, has poor influence, is likely to be not purchased at all by a spectator, and has poor overall effect. Third, the commodity is related to a plurality of commodities, but the total value of the commodity is low, and no money is earned for platform operators; (2) lack of refinement: the user attribute characteristics such as gender, income and the like of the users are not distinguished by the general recommendation, and the effect of the general recommendation can have different degrees of influence on the users with different attributes, so that the recommendation effect has larger difference.

Disclosure of Invention

The embodiment of the invention aims to provide a commodity recommending method, system and equipment based on user attributes and commodity types, and aims to solve the problems that in the commodity recommending method in the prior art, commodity factors are not comprehensive and attribute characteristics of users are not considered, so that recommending effects are poor and inaccurate.

A first object of an embodiment of the present invention is to provide a commodity recommendation method based on a user attribute and a commodity type, where the method includes:

creating a classified user commodity association information table according to the user commodity transaction data;

counting the total purchase frequency of each user class to each commodity; counting the associated purchase frequency of each user class to each commodity;

calculating the total value of each commodity according to the total purchase frequency of each commodity by each user class and the unit price of each commodity;

calculating the association weight, frequency weight and value weight of each commodity of each user class;

counting the frequency of the directed simultaneous purchase of commodities in each user category according to the commodity transaction data of the users; creating a directed commodity purchase frequency matrix of each user category;

establishing an undirected commodity purchase frequency matrix of each user category according to the directed commodity purchase frequency matrix of each user category;

Carrying out normalization processing on the directed commodity purchase frequency matrixes of all the user categories to obtain corresponding directed commodity purchase weighting frequency matrixes;

calculating the triggering centrality of each commodity of each user class;

calculating the network core degree of each commodity of each user class;

respectively carrying out normalization processing on the association weight, the frequency weight, the value weight, the trigger center degree and the network core degree of each commodity of each user category to obtain corresponding weighted association weight, weighted frequency weight, weighted value weight, weighted trigger center degree and weighted network core degree;

carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger centrality and the weighted network core degree to obtain the optimal node degree of the classified user commodity;

and recommending the commodities to the corresponding classified users according to the optimal node degree of the commodities of the classified users.

A second object of an embodiment of the present invention is to provide a commodity recommendation system based on user attributes and commodity types, the system including:

the user commodity associated information table creation module is used for creating a classified user commodity associated information table according to the user commodity transaction data;

The commodity purchase frequency statistics module is used for counting the total purchase frequency of each user class on each commodity; counting the associated purchase frequency of each user class to each commodity;

the commodity total value calculation module is used for calculating the total value of each commodity according to the total purchase frequency of each commodity and the unit price of each commodity;

the weight calculation module is used for calculating the association weight, the frequency weight and the value weight of each commodity in each user category;

the directed commodity purchase frequency matrix creation module is used for counting the frequency of directed simultaneous purchase of commodities of each user category according to commodity transaction data of the user; creating a directed commodity purchase frequency matrix of each user category;

the undirected commodity purchase frequency matrix creation module is used for creating undirected commodity purchase frequency matrixes of all user categories according to the directed commodity purchase frequency matrixes of all user categories;

the first normalization processing module is used for performing normalization processing on the directed commodity purchase frequency matrixes of all user categories to obtain corresponding directed commodity purchase weighting frequency matrixes;

the trigger center calculating module is used for calculating the trigger center of each commodity in each user category;

The network core degree calculating device is used for calculating the network core degree of each commodity of each user class;

the second normalization processing module is used for respectively carrying out normalization processing on the association weight, the frequency weight, the value weight, the trigger centrality and the network core degree of each commodity of each user category to obtain corresponding weighted association weight, weighted frequency weight, weighted value weight, weighted trigger centrality and weighted network core degree;

the optimal node degree calculation module is used for carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger center degree and the weighted network core degree to obtain the optimal node degree of the classified user commodity;

and the commodity recommendation module is used for recommending commodities to the corresponding classified users according to the optimal node degree of the commodities of the classified users.

A third object of an embodiment of the present invention is to provide an apparatus, including a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the commodity recommendation method based on user attributes and commodity types when the computer program is executed.

The beneficial effects of the invention are that

The invention provides a commodity recommending method based on user attribute and commodity type, firstly, the method forms a binary relation network of user attribute-commodity type by attribute information of the user and commodity type information purchased by the user, and then calculates association weight, frequency weight, value weight, triggering centrality and network core degree of commodity according to certain attribute information of the user; and then carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger center degree and the weighted network core degree to obtain the optimal node degree of the classified user commodity, and carrying out commodity recommendation on the corresponding classified user according to the optimal node degree of the classified user commodity. The method can obtain the commodity with the most spread and the most influence under a certain specific user attribute, and the commodity pushing is more refined and more accurate.

Drawings

FIG. 1 is a flow chart of a commodity recommendation method based on user attributes and commodity types provided by an embodiment of the present invention;

FIG. 2 is a flowchart of a method for calculating network core of each commodity of each user category according to an embodiment of the present invention;

FIG. 3 is a commodity node G according to an embodiment of the present invention ₁ 、G ₂ 、G ₃ 、G ₄ 、G ₅ A connection relation network diagram of (1);

FIG. 4 is a block diagram of a commodity recommendation system based on user attributes and commodity types according to an embodiment of the present invention;

fig. 5 is a block diagram of a network core computing device according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be further described in detail with reference to the accompanying drawings and examples, and only the portions related to the examples of the present invention are shown for convenience of description. It is to be understood that the specific embodiments described herein are for illustrative purposes only and are not limiting, as other equivalent embodiments according to the present invention may be made by those of ordinary skill in the relevant art without departing from the inventive concepts herein.

The invention provides a commodity recommending method based on user attribute and commodity type, firstly, the method forms a binary relation network of user attribute-commodity type by attribute information of the user and commodity type information purchased by the user, and then calculates association weight, frequency weight, value weight, triggering centrality and network core degree of commodity according to certain attribute information of the user; and then carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger center degree and the weighted network core degree to obtain the optimal node degree of the classified user commodity, and carrying out commodity recommendation on the corresponding classified user according to the optimal node degree of the classified user commodity. The method can obtain the commodity with the most spread and the most influence under a certain specific user attribute, and the commodity pushing is more refined and more accurate.

FIG. 1 is a flow chart of a commodity recommendation method based on user attributes and commodity types provided by an embodiment of the present invention; the method comprises the following steps:

step1, creating a classified user commodity association information table according to user commodity transaction data;

the user commodity transaction data includes fields: user number, user category, merchandise category, transaction time;

the classified user commodity association information table comprises the following fields: user number, user category and specific commodity categories; when the value in the user commodity correlation information table indicates that a purchasing behavior occurs, whether a corresponding line of users has purchased a corresponding commodity, 1 indicates that the corresponding commodity is purchased, and 0 indicates that the corresponding commodity is not purchased; each row in the table represents a purchase record; the same user can have multiple purchase records;

in a specific embodiment, the user number may be a number, a character, or a combination thereof, which may uniquely mark the user identity, such as a mobile phone number, a card number, a number, etc.; the user category is classified according to user attributes or characteristics, such as gender, age, household income, region, etc.; the commodity transaction data is acquired from a commodity super or an electronic commodity platform and the like;

the user commodity transaction data fragments according to the embodiment of the invention are shown in the table 1; the user number is a card number, and the user category is gender; the commodity is pay video, including love pay video, variety pay video and the like.

Card number	Sex (sex)	Goods commodity	Transaction time
				26630	Female	Love of love	11-01 19:00
26630	Female	Variety of arts	11-01 19:20
				26630	Female	Youth	11-01 19:40
62995	Female	Variety of arts	11-02 17:50
				62995	Female	Cartoon	11-02 19:00
38765	Man's body	Action	11-03 18:00
				…	...	...	...

TABLE 1

In particular embodiments, purchased may also represent favorite, watched, selected, collected, shopping cart, and the like; non-purchase may also mean dislike, not viewed, or not selected, and the like, as will be appreciated by those skilled in the art, without limiting the scope of the present invention;

the table 2 shows the classification user commodity associated information table segments according to the embodiment of the present invention;

TABLE 2

Step2, counting the total purchase frequency of each user class for each commodity; counting the associated purchase frequency of each user class to each commodity;

in the embodiment of the invention, the commodity platform is provided with n commodities in total, and the set G= { G is used ₁ ，G ₂ ，…G _n -representation; n represents the number of commodity types, also called commodity node number; there are k user categories in total, with the set c= { C ₁ ，c ₂ ，…c _k I= {1,2, 3..n }, x= {1,2, 3..k };

user class c _x For commodity G _i The total purchase frequency is represented by a kxn matrix cf3= { (R3 (c) _x )) _i (R3 (c) _x )) _i The value of (c) represents the user category c _x For commodity G _i Is a total purchase frequency of (2);

user class c _x For commodity G _i The associated purchase frequency of (a) is represented by a kxn matrix cf2= { (R2(c _x )) _i Represented by (R2 (c) _x )) _i The value of (c) represents the user category c _x For commodity G _i Associated purchase frequency of (a);

(R2(c _x )) _i ＝(R3(c _x )) _i -(R1(c _x )) _i the method comprises the steps of carrying out a first treatment on the surface of the Wherein (R1 (c) _x )) _i Representing user class c _x For commodity G _i Is a separate purchase frequency of (1); commodity G _i The individual purchase frequency indicates that only a single commodity G is purchased _i Frequency of (3); commodity G _i The related purchase frequency indicates that the purchased commodity contains the commodity G _i Frequency of (i.e. commodity G) _i Frequency of simultaneous purchase with all other merchandise;

as shown in Table 3, the total purchase frequency data table of the commodities of the male and female users in the embodiment of the invention is shown;

	love of love	Suspense doubt	Literature and art	Variety of arts	Martial arts	Action	History of	Youth	Crime	War	Cartoon
												Man's body	0	2	0	0	4	6	2	3	6	4	6
Female	6	1	4	18	2	2	1	10	2	0	2

TABLE 3 Table 3

As shown in Table 4, the data table of the commodity association purchasing frequency of the male and female users according to the embodiment of the invention;

	love of love	Suspense doubt	Literature and art	Variety of arts	Martial arts	Action	History of	Youth	Crime	War	Cartoon
												Man's body	0	1	0	0	2	2	1	2	2	2	3
Female	5	0	4	10	1	1	0	6	1	0	1

TABLE 4 Table 4

Step3, calculating the total value of each commodity according to the total purchase frequency of each commodity and the unit price of each commodity;

user class c _x Purchased goods G _i The total value of the matrix was represented by kxn cf4= { (Val 3 (c) _x )) _i (Val 3 (c) _x )) _i The value of (c) represents the user category c _x Purchased goods G _i Is a total value of (2);

(Val3(c _x )) _i ＝(R3(c _x )) _i ×(Pr) _i ；

table 5 shows a table of total value data for each commodity for a male user category in accordance with an embodiment of the present invention;

	love of love	Suspense doubt	Literature and art	Variety of arts	Martial arts	Action	History of	Youth	Crime	War	Cartoon
												Man's body	0	2	0	0	4	6	2	3	6	4	6
Commodity unit price	3	5	3	2	5	5	3	3	5	6	4
												Total value of commodity	0	10	0	0	20	30	6	9	30	24	24

TABLE 5

Table 7 shows a total value data table for each commodity of the female user class according to the embodiment of the present invention;

	love of love	Suspense doubt	Literature and art	Variety of arts	Martial arts	Action	History of	Youth	Crime	War	Cartoon
												Female	6	1	4	18	2	2	1	10	2	0	2
Commodity unit price	3	5	3	2	5	5	3	3	5	6	4
												Total value of commodity	18	5	12	36	10	10	3	30	10	6	8

TABLE 6

Step4, calculating the association weight, the frequency weight and the value weight of each commodity of each user class;

wherein ,(W₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Associated weight of (c), i.e. user category c _x For commodity G _i Is a function of the frequency of purchase (R2 (c) _x )) _i Occupy user class c _x For commodity G _i Is (R3 (c) _x )) _i Is a ratio of (3);

in the embodiment of the invention, according to the tables 3 and 4, the association weight of the commodity 'love' purchased by the female user is as follows

wherein ,(V₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Frequency weight of (c), i.e. user class c _x For commodity G _i The total purchase frequency of (c) was (R3 (c) _x )) _i Occupy user class c _x Total frequency of purchase for all goods

Is a ratio of (3);

in the embodiment of the invention, according to Table 3, the frequency weight of the commodity "love" purchased by the female user is

(U ₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Value weight of (c), i.e. user category c _x Purchased commodity G _i Is (Val 3 (c) _x )) _i Occupy user class c _x Total value of all items purchased

Is a ratio of (3);

in the embodiment of the invention, according to Table 6, the value weight of the commodity 'love' purchased by the female user is

Step5, counting the frequency of the directed simultaneous purchase of the commodities of each user category according to the commodity transaction data of the user; and creates a directed commodity purchase frequency matrix GG1 (c) for each user category _x )；

Extracting user commodityUser category c in transaction data _x Based on user category c _x Creating user category c _x A directed commodity purchase frequency matrix;

commodity G _i And G _j N×n matrix GG1 (c) for directed simultaneous purchase trigger relationship _x )＝{(R4(c _x )) _ij -representation; (R4 (c) _x )) _ij The value of (c) represents the user category c _x Purchasing commodity G in transaction data _i Whether or not to trigger the purchase of commodity G _j ；(R4(c _x )) _ij =1 indicates purchase of commodity G _i Time triggered purchase of commodity G _j ；(R4(c _x )) _ij =0 indicates purchase of commodity G _i When not triggering the purchase of commodity G _j ；i＝{1，2，3...n}，j＝{1，2，3...n}；

In the embodiment of the invention, in transaction data of which the user category is female, a matrix of the purchase frequency of the directed commodity is shown in table 7;

love of love

Suspense doubt

Literature and art

Variety of arts

Martial arts

Action

History of

Youth

Crime

War

Cartoon

Love of love

/

1

1

2

Suspense doubt

/

Literature and art

/

1

Variety of arts

1

2

/

3

1

Martial arts

/

1

Action

/

History of

/

Youth

1

/

Crime

1

/

War

/

Cartoon

/

TABLE 7

Step6, based on the directed commodity purchase frequency matrix GG1 (c) _x ) Creating an undirected commodity purchase frequency matrix GG2 (c) for each user category _x )；

User class c _x Is used for the undirected commodity purchase frequency matrix of (c) and is represented by an n x n matrix GG2 (c _x )＝{(R5(c _x )) _ij -representation; (R5 (c) _x )) _ij The value of (c) represents the user category c _x Commodity G in transaction data _i And G _j Frequency of simultaneous purchase;

commodity G _i And G _j The existence of the simultaneous purchase trigger relationship includes: purchasing commodity G _i Time triggered purchase of commodity G _j Relationship and purchase of goods G _j Time triggered purchase of commodity G _i Is a relationship of (2);

in the embodiment of the invention, in transaction data of which the user category is female, a matrix of undirected commodity purchase frequency is shown in table 8;

love of love

Suspense doubt

Literature and art

Variety of arts

Martial arts

Action

History of

Youth

Crime

War

Cartoon

Love of love

/

1

2

2

Suspense doubt

/

Literature and art

1

/

3

Variety of arts

2

3

/

4

1

Martial arts

/

1

1

Action

1

/

History of

/

Youth

2

4

/

Crime

1

/

War

/

Cartoon

1

/

TABLE 8

Step7, a directed commodity purchase frequency matrix GG1 (c) _x ) Normalization processing is performed to obtain a corresponding directed commodity purchase weighted frequency matrix GG11 (c _x )；

For user category c _x Directed commodity purchase frequency matrix GG1 (c) _x ) The normalized calculation formula is:

wherein ,(F2_x ) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Is also called a weighted frequency value; (R4 (c) _x )) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Frequency of (3); min (GG 1 (c) _x ) Matrix GG1 (c) representing frequency of purchase of directed goods _x ) A minimum frequency value of (a); max (GG 1 (c) _x ) A directed commodity purchase frequency matrix GG1 (c) _x ) The maximum frequency value of (a);

in the embodiment of the invention, the data of the weighted frequency matrix of the purchase of the directed commodities with the user category of women is shown in a table 9;

love of love

Suspense doubt

Literature and art

Variety of arts

Martial arts

Action

History of

Youth

Crime

War

Cartoon

Love of love

/

0.1

0.1

0.2

Suspense doubt

/

Literature and art

/

0.1

Variety of arts

0.1

0.2

/

0.3

0.1

Martial arts

/

0.1

Action

/

History of

/

Youth

0.1

/

Crime

0.1

/

War

/

Cartoon

/

Table 9Step8, the trigger centrality (D ₁ (c _x )) _i ；

wherein ,(D₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Is triggered by the trigger centrality of (2); (CF (c) _x )) _i Representing user class c _x Purchased commodity G _i Trigger degree of (a), i.e. user category c _x Purchasing commodity G _i Triggering normalized frequency of purchasing other commodity types;

wherein i≠j；

step9, calculating the network core degree (H ₁ (c _x )) _i ；

Fig. 2 is a flowchart of a method for calculating network core degree of each commodity of each user category according to an embodiment of the present invention, including the following steps:

s91, according to the undirected commodity purchase frequency matrix GG22 (c _x ) Constructing a commodity connection relation network;

the method comprises the following steps: the commodity is taken as a node, the connection relation between commodities is taken as an edge, the frequency of simultaneous purchase between commodities is taken as an edge weight, and a commodity connection relation network is constructed;

If commodity node G _i and G_j The commodity node G is in a simultaneous purchase triggering relationship (direct connection relationship) _i and G_j A connecting edge is added between the two connecting edges; commodity node G _i and G_j There may also be an indirect connection between the two, connected by a path comprising at least 1 intermediate node and at least 2 connecting edges, at least one of the number of paths;

s92, for commodity node G with direct or indirect connection _i and G_j Obtaining commodity node G _j Opposite node G _i Node level lev of (a);

set G _i As the original node, if the commodity node G _i and G_j If there is a direct connection relationship, node G _j Is G _i Is a level 1 node of (2);

if commodity node G _i and G_j There is indirect connection relation and connected by m node paths, m is more than or equal to 1, and p= { path is used for m node path collection ₁ ，path ₂ ，…G _m -representation; the number of connecting edges correspondingly contained in each node path in P is S= { side ₁ ，side ₂ ，…side _m Indicated, node G _j Opposite node G _i Node level lev=min (S), min (S) > 2, min () represents taking the minimum value;

as shown in FIG. 3, the commodity node G according to the embodiment of the present invention ₁ 、G ₂ 、G ₃ 、G ₄ 、G ₅ A connection relation network diagram of (1);

wherein G₁ and G₂ If there is a direct connection relationship, G ₂ Is G ₁ Is a level 1 node of (2);

G ₁ and G₃ An indirect connection relationship exists, and the first node path and the second node path can be indirectly connected; the first node path includes 2 connection edges: g ₁ -G ₂ and G₂ -G ₃ The method comprises the steps of carrying out a first treatment on the surface of the The second node path includes 4 connection edges: g ₁ -G ₂ 、G ₂ -G ₄ 、G ₄ -G ₅ and G₅ -G ₃ The method comprises the steps of carrying out a first treatment on the surface of the Then G ₃ Is G ₁ Is a level 2 node of (2);

the value on the connecting edge represents the edge weight (i.e., the frequency of simultaneous purchases) between two directly connected commodity nodes, such as G ₁ and G₂ The edge weight of (2) is 10, G ₂ and G₃ The edge weight of (2) is 5;

s93, calculating the network position F of each commodity node of each user class _b (i)；

Wherein, (FB (c) _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network location of (2); lev (Lev) _ij Representation and commodity node G _i Directly or indirectly connected commodity node G _j Is a node level of (2); (R3 (c) _x )) _j Representation and commodity node G _i Directly or indirectly connected commodity node G _j The total purchase frequency of the corresponding commodity, namely the user class c _x For commodity G _j Is a total purchase frequency of (2); n represents the total number of commodity nodes;

s94, calculating the network core degree (H) of each commodity node of each user class ₁ (c _x )) _i ；

wherein ,(H₁ (c _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network core degree of (1);

step10, the association weights (W ₁ (c _x )) _i Frequency weight (V) ₁ (c _x )) _i Value weight (U) ₁ (c _x )) _i Degree of trigger center (D) ₁ (c _x )) _i Network core (H) ₁ (c _x )) _i Normalization processing is performed to obtain a corresponding weighted association weight (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i ；

The normalized calculation formula is:

wherein F represents the normalized numerical value; r represents the user category c respectively _x Purchasing commodity G _i Is (W) ₁ (c _x )) _i Frequency weight (V) ₁ (c _x )) _i Value weight (U) ₁ (c _x )) _i Degree of trigger center (D) ₁ (c _x )) _i Network core (H) ₁ (c _x )) _i The method comprises the steps of carrying out a first treatment on the surface of the min (T) respectively represents the minimum value in the association weight, the frequency weight, the value weight, the trigger centrality and the network centrality data of all commodities; max (T) respectively represents the maximum numerical value in the association weight, the frequency weight, the value weight, the trigger centrality and the network centrality data of all commodities;

for example, if F represents the normalized network core, then R represents the user class c _x Commodity node G of (1) _i Network core (H) ₁ (c _x )) _i The method comprises the steps of carrying out a first treatment on the surface of the min represents user category c _x Network core set of all commodity nodes { (H) ₁ (c _x )) _i Minimum of i=1, 2,3 … … n }, max denotes user category c _x Network core set of all commodity nodes { (H) ₁ (c _x )) _i I=maximum of 1,2,3 … … n; others and so on;

step11, associate the weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i The weighted summation is performed to obtain the optimal node degree (S (c) _x )) _i ；

(S(c _x )) _i ＝((W ₂ (c _x )) _i +(V ₂ (c _x )) _i +(U ₂ (c _x )) _i +(D ₂ (c _x )) _i +(H ₂ (c _x )) _i )/5；

Wherein (S (c) _x )) _i Representing user class c _x Commodity node G of (1) _i Is defined as the optimal node degree;

alternatively, the process may be carried out in a single-stage,

the weighted associated weights (W) are calculated according to the analytic hierarchy process ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i Weights in commodity recommendation decision targets are weighted to correlate weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i Weighted summation is carried out to obtain the classificationOptimal node degree of household commodity (S (c) _x )) _i ；

(S(c _x )) _i ＝α ₁ (W ₂ (c _x )) _i +α ₂ (V ₂ (c _x )) _i +α ₃ (U ₂ (c _x )) _i +α ₄ (D ₂ (c _x )) _i +α ₅ (H ₂ (c _x )) _i ；

wherein ,α₁ 、α ₂ 、α ₃ 、α ₄ 、α ₅ Respectively represent (W) ₂ (c _x )) _i 、(V ₂ (c _x )) _i 、(U ₂ (c _x )) _i 、(D ₂ (c _x )) _i 、(H ₂ (c _x )) _i Weights in the commodity recommendation decision target;

specifically, commodity recommendation is used as a decision target, and a weighted association weight (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i As a judgment matrix element in the analytic hierarchy process, carrying out pairwise comparison evaluation on the importance degree of the element; the analytic hierarchy process is a conventional process and is further described herein;

in the application scene, judging according to specific service, wherein the specific evaluation filling of a judging model is shown in a table 10;

Table 10

The weighted associated weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i Weight alpha of (2) ₁ 、α ₂ 、α ₃ 、α ₄ 、α ₅ The method comprises the following steps of: 0.55 023,0.08,0.08,0.08;

the optimal node degree of the user commodity is classified (S (c) _x )) _i The method comprises the following steps:

(S(c _x )) _i ＝0.55(W ₂ (c _x )) _i +0.23(V ₂ (c _x )) _i +0.08(U ₂ (c _x )) _i +0.08(D ₂ (c _x )) _i +0.08(H ₂ (c _x )) _i ；

step12, recommending the commodities to the corresponding classified users according to the optimal node degree of the commodities of the classified users.

In the embodiment, sorting the commodities according to the optimal node degree of the commodities, and selecting the commodities with the large optimal node degree value to push the corresponding classified users preferentially;

corresponding to the commodity recommendation method based on the user attribute and the commodity type described in the above embodiments, fig. 4 is a diagram of a commodity recommendation system based on the user attribute and the commodity type according to the embodiment of the present invention; the system comprises:

the user commodity associated information table creation module is used for creating a classified user commodity associated information table according to the user commodity transaction data;

the commodity purchase frequency statistics module is used for counting the total purchase frequency of each user class on each commodity; counting the associated purchase frequency of each user class to each commodity;

The commodity total value calculation module is used for calculating the total value of each commodity according to the total purchase frequency of each commodity and the unit price of each commodity;

the weight calculation module is used for calculating the association weight, the frequency weight and the value weight of each commodity in each user category;

the directed commodity purchase frequency matrix creation module is used for counting the frequency of directed simultaneous purchase of commodities of each user category according to commodity transaction data of the user; creating a directed commodity purchase frequency matrix of each user category;

the undirected commodity purchase frequency matrix creation module is used for creating undirected commodity purchase frequency matrixes of all user categories according to the directed commodity purchase frequency matrixes of all user categories;

the first normalization processing module is used for performing normalization processing on the directed commodity purchase frequency matrixes of all user categories to obtain corresponding directed commodity purchase weighting frequency matrixes;

the trigger center calculating module is used for calculating the trigger center of each commodity in each user category;

the network core degree calculating device is used for calculating the network core degree of each commodity of each user class;

the second normalization processing module is used for respectively carrying out normalization processing on the association weight, the frequency weight, the value weight, the trigger centrality and the network core degree of each commodity of each user category to obtain corresponding weighted association weight, weighted frequency weight, weighted value weight, weighted trigger centrality and weighted network core degree;

The optimal node degree calculation module is used for carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger center degree and the weighted network core degree to obtain the optimal node degree of the classified user commodity;

and the commodity recommendation module is used for recommending commodities to the corresponding classified users according to the optimal node degree of the commodities of the classified users.

Further, the user commodity transaction data includes fields: user number, user category, merchandise category, transaction time;

the classified user commodity association information table comprises the following fields: user number, user category and specific commodity categories; when the value in the user commodity correlation information table indicates that a purchasing behavior occurs, whether a corresponding line of users has purchased a corresponding commodity, 1 indicates that the corresponding commodity is purchased, and 0 indicates that the corresponding commodity is not purchased; each row in the table represents a purchase record; the same user can have multiple purchase records;

further, in the embodiment of the present invention, the commodity platform is assumed to have n commodities in total, and the set g= { G is used ₁ ，G ₂ ，…G _n -representation; n represents the number of commodity types, also called commodity node number; there are k user categories in total, with the set c= { C ₁ ，c ₂ ，…c _k I= {1,2, 3..n }, x= {1,2, 3..k };

User class c _x For commodity G _i The total purchase frequency is represented by a kxn matrix cf3= { (R3 (c) _x )) _i (R3 (c) _x )) _i The value of (c) represents the user category c _x For commodity G _i Is a total purchase frequency of (2);

user class c _x For commodity G _i The associated purchase frequency is represented by a kxn matrix cf2= { (R2 (c) _x )) _i Represented by (R2 (c) _x )) _i The value of (c) represents the user category c _x For commodity G _i Associated purchase frequency of (a);

(R2(c _x )) _i ＝(R3(c _x )) _i -(R1(c _x )) _i the method comprises the steps of carrying out a first treatment on the surface of the Wherein (R1 (c) _x )) _i Representing user class c _x For commodity G _i Is a separate purchase frequency of (1); commodity G _i The individual purchase frequency indicates that only a single commodity G is purchased _i Frequency of (3); commodity G _i The related purchase frequency indicates that the purchased commodity contains the commodity G _i Frequency of (i.e. commodity G) _i Frequency of simultaneous purchase with all other merchandise;

further, user category c _x Purchased goods G _i The total value of the matrix was represented by kxn cf4= { (Val 3 (c) _x )) _i (Val 3 (c) _x )) _i The value of (c) represents the user category c _x Purchased goods G _i Is a total value of (2); (Val 3 (c) _x )) _i ＝(R3(c _x )) _i ×(Pr) _i ；

Further, the calculating the association weight, the frequency weight and the value weight of each commodity in each user category specifically comprises the following steps:

wherein ,(W₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Associated weight of (c), i.e. user category c _x For commodity G _i Is a function of the frequency of purchase (R2 (c) _x )) _i Occupy user class c _x For commodity G _i Is (R3 (c) _x )) _i Is a ratio of (3);

wherein ,(V₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Frequency weight of (c), i.e. user class c _x For commodity G _i The total purchase frequency of (c) was (R3 (c) _x )) _i Occupy user class c _x Total frequency of purchase for all goods

Is a ratio of (3); />

(U ₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Value weight of (c), i.e. user category c _x Purchased commodity G _i Is (Val 3 (c) _x )) _i Occupy user class c _x Total value of all items purchased

Is a ratio of (3);

further, according to the commodity transaction data of the users, the frequency of the directed simultaneous purchase of commodities in each user category is counted; and creates a directed commodity purchase frequency matrix GG1 (c) for each user category _x ) The method comprises the following steps:

extracting user category c in user commodity transaction data _x Based on user category c _x Creating user category c _x A directed commodity purchase frequency matrix;

commodity G _i And G _j N×n matrix GG1 (c) for directed simultaneous purchase trigger relationship _x )＝{(R4(c _x )) _ij -representation; (R4 (c) _x )) _ij The value of (c) represents the user category c _x Purchasing commodity G in transaction data _i Whether or not to trigger the purchase of commodity G _j ；(R4(c _x )) _ij =1 indicates purchase of commodity G _i Time triggered purchase of commodity G _j ；(R4(c _x )) _ij =0 indicates purchase of commodity G _i When not triggering the purchase of commodity G _j ；i＝{1，2，3...n}，j＝{1，2，3...n}；

Further, the directed commodity purchase frequency matrix GG1 (c) _x ) Creating an undirected commodity purchase frequency matrix GG2 (c) for each user category _x ) The method comprises the following steps:

user class c _x Is used for the undirected commodity purchase frequency matrix of (c) and is represented by an n x n matrix GG2 (c _x )＝{(R5(c _x )) _ij -representation; (R5 (c) _x )) _ij The value of (c) represents the user category c _x Commodity G in transaction data _i And G _j Frequency of simultaneous purchase;

commodity G _i And G _j The existence of the simultaneous purchase trigger relationship includes: purchasing commodity G _i Time triggered purchase of commodity G _j Relationship and purchase of goods G _j Time triggered purchase of commodity G _i Is a relationship of (2);

further, the directed commodity purchase frequency matrix GG1 (c) _x ) Normalization processing is performed to obtain a corresponding directed commodity purchase weighted frequency matrix GG11 (c _x ) The method comprises the following steps:

for user category c _x Directed commodity purchase frequency matrix GG1 (c) _x ) The normalized calculation formula is:

wherein ,(F2_x ) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Is also called a weighted frequency value; (R4 (c) _x )) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Frequency of (3); min (GG 1 (c) _x ) Matrix GG1 (c) representing frequency of purchase of directed goods _x ) A minimum frequency value of (a); max (GG 1 (c) _x ) A directed commodity purchase frequency matrix GG1 (c) _x ) The maximum frequency value of (a);

further, the trigger centrality (D ₁ (c _x )) _i The method comprises the following steps:

wherein ,(D₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Is triggered by the trigger centrality of (2); (CF (c) _x )) _i Representing user class c _x Purchased commodity G _i Trigger degree of (a), i.e. user category c _x Purchasing commodity G _i Triggering normalized frequency of purchasing other commodity types;

wherein i≠j；

further, fig. 5 is a diagram of a network core computing device according to an embodiment of the present invention; the network core computing device comprises:

the commodity connection relation network construction module is used for constructing a matrix GG22 (c) according to the undirected commodity purchase frequency _x ) Constructing a commodity connection relation network;

the method comprises the following steps: the commodity is taken as a node, the connection relation between commodities is taken as an edge, the frequency of simultaneous purchase between commodities is taken as an edge weight, and a commodity connection relation network is constructed;

if commodity node G _i and G_j The commodity node G is in a simultaneous purchase triggering relationship (direct connection relationship) _i and G_j A connecting edge is added between the two connecting edges; commodity node G _i and G_j There may also be an indirect connection between the two, connected by a path comprising at least 1 intermediate node and at least 2 connecting edges, at least one of the number of paths;

a node grade acquisition module for acquiring commodity node G with direct or indirect connection _i and G_j Obtaining commodity node G _j Opposite node G _i Node level lev of (a);

set G _i As the original node, if the commodity node G _i and G_j If there is a direct connection relationship, node G _j Is G _i Is a level 1 node of (2);

if commodity node G _i and G_j There is indirect connection relation and connected by m node paths, m is more than or equal to 1, and p= { path is used for m node path collection ₁ ，path ₂ ，…G _m -representation; the number of connecting edges correspondingly contained in each node path in P is S= { side ₁ ，side ₂ ，…side _m Indicated, node G _j Opposite node G _i Node level lev=min (S), min (S) > 2, min () represents taking the minimum value;

the network position calculating module is used for calculating the network position F of each commodity node of each user class _b (i)；

Wherein, (FB (c) _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network location of (2); lev (Lev) _ij Representation and commodity node G _i Directly or indirectly connected commodity node G _j Is a node level of (2); (R3 (c) _x )) _j Representation and commodity node G _i Directly or indirectly connected toCommodity node G _j The total purchase frequency of the corresponding commodity, namely the user class c _x For commodity G _j Is a total purchase frequency of (2); n represents the total number of commodity nodes;

a network core calculation module for calculating the network core (H) of each commodity node of each user class ₁ (c _x )) _i ；

wherein ,(H₁ (c _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network core degree of (1).

Further, the association weights (W ₁ (c _x )) _i Frequency weight (V) ₁ (c _x )) _i Value weight (U) ₁ (c _x )) _i Degree of trigger center (D) ₁ (c _x )) _i Network core (H) ₁ (c _x )) _i Normalization processing is performed to obtain a corresponding weighted association weight (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i ；

The normalized calculation formula is:

wherein F represents the normalized numerical value; r represents the user category c respectively _x Purchasing commodity G _i Is (W) ₁ (c _x )) _i Frequency weight (V) ₁ (c _x )) _i Value weight (U) ₁ (c _x )) _i Degree of trigger center (D) ₁ (c _x )) _i Network core (H) ₁ (c _x )) _i The method comprises the steps of carrying out a first treatment on the surface of the min (T) represents all the commodities respectivelyAssociating minimum values in the weight, frequency weight, value weight, trigger centrality and network centrality data; max (T) respectively represents the maximum numerical value in the association weight, the frequency weight, the value weight, the trigger centrality and the network centrality data of all commodities;

for example, if F represents the normalized network core, then R represents the user class c _x Commodity node G of (1) _i Network core (H) ₁ (c _x )) _i The method comprises the steps of carrying out a first treatment on the surface of the min represents user category c _x Network core set of all commodity nodes { (H) ₁ (c _x )) _i Minimum of i=1, 2,3 … … n }, max denotes user category c _x Network core set of all commodity nodes { (H) ₁ (c _x )) _i I=maximum of 1,2,3 … … n; others and so on;

further, the weighted association weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i The weighted summation is performed to obtain the optimal node degree (S (c) _x )) _i The method comprises the steps of carrying out a first treatment on the surface of the The method comprises the following steps:

(S(c _x )) _i ＝((W ₂ (c _x )) _i +(V ₂ (c _x )) _i +(U ₂ (c _x )) _i +(D ₂ (c _x )) _i +(H ₂ (c _x )) _i )/5；

further, the "associate weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i The weighted summation is performed to obtain the optimal node degree (S (c) _x )) _i "Ke TieThe method comprises the following steps:

"calculate weighted association weights according to analytic hierarchy process (W) ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i Weights in commodity recommendation decision targets are weighted to correlate weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i The weighted summation is performed to obtain the optimal node degree (S (c) _x )) _i ；”；

(S(c _x )) _i ＝α ₁ (W ₂ (c _x )) _i +α ₂ (V ₂ (c _x )) _i +α ₃ (U ₂ (c _x )) _i +α ₄ (D ₂ (c _x )) _i +α ₅ (H ₂ (c _x )) _i wherein ,α₁ 、α ₂ 、α ₃ 、α ₄ 、α ₅ Respectively represent (W) ₂ (c _x )) _i 、(V ₂ (c _x )) _i 、(U ₂ (c _x )) _i 、(D ₂ (c _x )) _i 、(H ₂ (c _x )) _i Weights in the commodity recommendation decision target;

specifically, commodity recommendation is used as a decision target, and a weighted association weight (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i As a judgment matrix element in the analytic hierarchy process, the importance of the elementPerforming pairwise comparison evaluation on the sex degree; the analytic hierarchy process is a conventional process and is further described herein;

in the application scene, judging according to specific service, wherein the specific evaluation filling of a judging model is shown in a table 10;

table 10

The weighted associated weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i Weight alpha of (2) ₁ 、α ₂ 、α ₃ 、α ₄ 、α ₅ The method comprises the following steps of: 0.55 023,0.08,0.08,0.08;

the optimal node degree of the user commodity is classified (S (c) _x )) _i The method comprises the following steps:

(S(c _x )) _i ＝0.55(W ₂ (c _x )) _i +0.23(V ₂ (c _x )) _i +0.08(U ₂ (c _x )) _i +0.08(D ₂ (c _x )) _i +0.08(H ₂ (c _x )) _i ；

further, the goods are recommended to the corresponding classified users according to the optimal node degree of the goods of the classified users, specifically in the embodiment, the goods are ordered according to the optimal node degree of the goods, and the goods with the large optimal node degree value are selected to be pushed to the corresponding classified users preferentially;

The embodiment of the invention also provides a terminal device, which comprises: a processor, a memory, and a computer program stored in the memory and executable on the processor. The processor, when executing the computer program, implements the steps of an embodiment of the commodity recommendation method based on the user attribute and the commodity type, for example, step1 to Step12 shown in fig. 1. Alternatively, the processor may implement the functions of the modules in the embodiment of the commodity recommendation system based on the user attribute and the commodity type when executing the computer program.

It will be appreciated by those of ordinary skill in the art that implementing all or part of the steps of the methods of the embodiments described above may be accomplished by program instruction related hardware, and the program may be stored on a computer readable storage medium, which may be a ROM, a RAM, a magnetic disk, an optical disk, etc.

The sequence numbers of the steps in the above embodiments do not mean the execution sequence, and the execution sequence of the processes should be determined according to the functions and internal logic, and should not limit the implementation process of the embodiments of the present invention.

The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.

Claims (11)

1. A commodity recommendation method based on user attributes and commodity types, the method comprising:

creating a classified user commodity association information table according to the user commodity transaction data;

counting the total purchase frequency of each user class to each commodity; counting the associated purchase frequency of each user class to each commodity;

wherein, the commodity platform is provided with n commodities in total, and the collection G= { G is used ₁ ，G ₂ ，…G _n -representation; n represents the number of commodity types, also called commodity node number; there are k user categories in total, with the set c= { C ₁ ，c ₂ ，…c _k -representation;

user class c _x For commodity G _i The total purchase frequency is represented by a kxn matrix cf3= { (R3 (c) _x )) _i -representation;

user class c _x For commodityG _i The associated purchase frequency is represented by a kxn matrix cf2= { (R2 (c) _x )) _i -representation;

(R2(c _x )) _i ＝(R3(c _x )) _i -(R1(c _x )) _i the method comprises the steps of carrying out a first treatment on the surface of the Wherein (R1 (c) _x )) _i Representing user class c _x For commodity G _i Is a separate purchase frequency of (1); commodity G _i The individual purchase frequency indicates that only a single commodity G is purchased _i Frequency of (3); commodity G _i The related purchase frequency indicates that the purchased commodity contains the commodity G _i Frequency of (i.e. commodity G) _i Frequency of simultaneous purchase with all other merchandise;

calculating the total value of each commodity according to the total purchase frequency of each commodity by each user class and the unit price of each commodity;

calculating the association weight, frequency weight and value weight of each commodity of each user class;

wherein ,

(W ₁ (c _x )) _i representing user class c _x Purchased commodity G _i Is a weight of the association of (a);

counting the frequency of the directed simultaneous purchase of commodities in each user category according to the commodity transaction data of the users; creating a directed commodity purchase frequency matrix of each user category;

wherein, commodity G _i And G _j N×n matrix GG1 (c) for directed simultaneous purchase trigger relationship _x )＝{(R4(c _x )) _ij [ Z ] represents GG1 (c) _x ) Representing user class c _x A directed commodity purchase frequency matrix; (R4 (c) _x )) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Frequency of (3);

establishing an undirected commodity purchase frequency matrix of each user category according to the directed commodity purchase frequency matrix of each user category;

wherein user category c _x N×n matrix GG2 for undirected commodity purchase frequency matrix(c _x )＝{(R5(c _x )) _ij -representation; GG2 (c) _x ) Representing user class c _x Matrix of unoccupied commodity purchase frequency, (R5 (c) _x )) _ij The value of (c) represents the user category c _x Commodity G in transaction data _i And G _j Frequency of simultaneous purchase;

carrying out normalization processing on the directed commodity purchase frequency matrixes of all the user categories to obtain corresponding directed commodity purchase weighting frequency matrixes;

calculating the triggering centrality of each commodity of each user class;

the calculating of the triggering centrality of each commodity of each user category comprises the following specific steps:

wherein ,(D₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Is triggered by the trigger centrality of (2); (CF (c) _x )) _i Representing user class c _x Purchased commodity G _i Trigger degree of (a), i.e. user category c _x Purchasing commodity G _i Triggering normalized frequency of purchasing other commodity types;

where i+.j, (F2) _x ) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Is a normalized frequency value of (2);

calculating the network core degree of each commodity of each user class;

the calculating the network core degree of each commodity of each user category comprises the following steps:

constructing a commodity connection relation network according to the undirected commodity purchase frequency matrix;

for commodity nodes G with direct or indirect connection _i and G_j Obtaining commodity node G _j Opposite node G _i Node level lev of (a);

calculating the network position of each commodity node of each user class;

calculating the network core degree of each commodity node of each user class;

the matrix GG2 (c) _x ) The commodity connection relation network is constructed, specifically:

the commodity is taken as a node, the connection relation between commodities is taken as an edge, the frequency of simultaneous purchase between commodities is taken as an edge weight, and a commodity connection relation network is constructed;

if commodity node G _i and G_j If the simultaneous purchase triggering relationship exists, the commodity node G _i and G_j A connecting edge is added between the two connecting edges; commodity node G _i and G_j There may also be an indirect connection between the two, connected by a path comprising at least 1 intermediate node and at least 2 connecting edges, at least one of the number of paths;

for commodity nodes G with direct or indirect connection _i and G_j Obtaining commodity node G _j Opposite node G _i Node level lev of (a);

set G _i As the original node, if the commodity node G _i and G_j If there is a direct connection relationship, node G _j Is G _i Is a level 1 node of (2);

if commodity node G _i and G_j There is indirect connection relation and connected by m node paths, m is more than or equal to 1, and p= { path is used for m node path collection ₁ ，path ₂ ，…G _m -representation; the number of connecting edges correspondingly contained in each node path in P is S= { side ₁ ，side ₂ ，…side _m Indicated, node G _j Opposite node G _i Node level lev=min (S), min (S) > 2, min () represents taking the minimum value;

the calculating of the network positions of commodity nodes of each user category specifically comprises the following steps:

Wherein, (FB (c) _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network location of (2); lev (Lev) _ij Representation and commodity node G _i Directly or indirectly connected commodity node G _j Is a node level of (2); (R3 (c) _x )) _j Representation and commodity node G _i Directly or indirectly connected commodity node G _j The total purchase frequency of the corresponding commodity, namely the user class c _x For commodity G _j Is a total purchase frequency of (2); n represents the total number of commodity nodes;

the network core degree of each commodity node of each user category is calculated specifically as follows:

wherein ,(H₁ (c _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network core degree of (1);

respectively carrying out normalization processing on the association weight, the frequency weight, the value weight, the trigger center degree and the network core degree of each commodity of each user category to obtain corresponding weighted association weight, weighted frequency weight, weighted value weight, weighted trigger center degree and weighted network core degree;

carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger centrality and the weighted network core degree to obtain the optimal node degree of the classified user commodity;

and recommending the commodities to the corresponding classified users according to the optimal node degree of the commodities of the classified users.

2. The commodity recommendation method according to claim 1, wherein the step of performing weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger center degree and the weighted network core degree to obtain the optimal node degree of the classified user commodity is replaced by:

And calculating weights of the weighted association weights, the weighted frequency weights, the weighted value weights, the weighted trigger centrality and the weighted network centrality in the commodity recommendation decision target according to the analytic hierarchy process, and carrying out weighted summation on the weighted association weights, the weighted frequency weights, the weighted value weights, the weighted trigger centrality and the weighted network centrality to obtain the optimal node degree of the classified user commodity.

3. The merchandise recommendation method based on user attributes and merchandise types of claim 2, wherein said user merchandise transaction data comprises the fields of: user number, user category, merchandise category, transaction time; the classified user commodity association information table comprises the following fields: user number, user category, and a number of specific merchandise categories.

4. The commodity recommending method according to the present invention as claimed in claim 3, wherein said calculating the total value of each commodity according to the total purchase frequency of each commodity and each commodity price of each user category comprises:

user class c _x Purchased goods G _i The total value of the matrix was represented by kxn cf4= { (Val 3 (c) _x )) _i (Val 3 (c) _x )) _i The value of (c) represents the user category c _x Purchased goods G _i Is a total value of (2); (Pr) _i Representing the unit price of each commodity;

(Val3(c _x )) _i ＝(R3(c _x )) _i ×(Pr) _i 。

5. the commodity recommending method according to claim 4, wherein said calculating the associated weight, the frequency weight, and the value weight of each commodity in each user category comprises:

wherein ,(W₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Associated weight of (c), i.e. user category c _x For commodity G _i Is a function of the frequency of purchase (R2 (c) _x )) _i Occupy user class c _x For commodity G _i Is (R3 (c) _x )) _i Is a ratio of (3);

wherein ,(V₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Frequency weight of (c), i.e. user class c _x For commodity G _i The total purchase frequency of (c) was (R3 (c) _x )) _i Occupy user class c _x Total frequency of purchase for all goods +.>

Is a ratio of (3);

(U ₁ (c _x )) _i representing user class c _x Purchased commodity G _i Value weight of (c), i.e. user category c _x Purchased commodity G _i Is (Val 3 (c) _x )) _i Occupy user class c _x Total value of all items purchased

Is a ratio of (2).

6. The commodity recommendation method according to claim 5, wherein the normalizing the directional commodity purchase frequency matrix of each user category to obtain a corresponding directional commodity purchase weighted frequency matrix specifically comprises:

for user category c _x Directed commodity purchase frequency matrix GG1 (c) _x ) The normalized calculation formula is:

wherein ,(F2_x ) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Is also called a weighted frequency value; (R4 (c) _x )) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Frequency of (3); min (GG 1 (c) _x ) Matrix GG1 (c) representing frequency of purchase of directed goods _x ) A minimum frequency value of (a); max (GG 1 (c) _x ) A directed commodity purchase frequency matrix GG1 (c) _x ) Is the maximum frequency value of (a).

7. The commodity recommending method according to claim 1, wherein the correlation weight, the frequency weight, the value weight, the trigger center and the network core of each commodity in each user category are normalized,

the normalized calculation formula is:

wherein F represents the normalized numerical value; r represents the user category c respectively _x Purchasing commodity G _i Is (W) ₁ (c _x )) _i Frequency weight (V) ₁ (c _x )) _i Value weight (U) ₁ (c _x )) _i Degree of trigger center (D) ₁ (c _x )) _i Network core (H) ₁ (c _x )) _i The method comprises the steps of carrying out a first treatment on the surface of the min (T) respectively represents the minimum value in the association weight, the frequency weight, the value weight, the trigger centrality and the network centrality data of all commodities; max (T) respectively represents the maximum numerical value in the association weight, the frequency weight, the value weight, the trigger center degree and the network core degree data of all commodities.

8. A method for recommending goods based on user attributes and type of goods as claimed in any of claims 1 or 3-7, characterized in that the weighted association weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i The weighted summation is performed to obtain the optimal node degree (S (c) _x )) _i ；

(S(c _x )) _i ＝((W ₂ (c _x )) _i +(V ₂ (c _x )) _i +(U ₂ (c _x )) _i +(D ₂ (c _x )) _i +(H ₂ (c _x )) _i )/5；

Wherein (S (c) _x )) _i Representing user class c _x Commodity node G of (1) _i Is defined as the optimal node degree of the node(s).

9. Commodity recommendation method according to one of the claims 1-7, characterized in that the weighted association weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i Weights in commodity recommendation decision targets are weighted to correlate weights (W ₂ (c _x )) _i Weighting frequency weight (V ₂ (c _x )) _i Weighted value weight (U) ₂ (c _x )) _i Weighted trigger centrality (D) ₂ (c _x )) _i Weighted network core (H) ₂ (c _x )) _i Carrying out weighted summation to obtain the optimal node degree of the classified user commodity;

(S(c _x )) _i ＝α ₁ (W ₂ (c _x )) _i +α ₂ (V ₂ (c _x )) _i +α ₃ (U ₂ (c _x )) _i +α ₄ (D ₂ (c _x )) _i +α ₅ (H ₂ (c _x )) _i wherein ,α₁ 、α ₂ 、α ₃ 、α ₄ 、α ₅ Respectively represent (W) ₂ (c _x )) _i 、(V ₂ (c _x )) _i 、(U ₂ (c _x )) _i 、(D ₂ (c _x )) _i 、(H ₂ (c _x )) _i Weights in commodity recommendation decision targets.

10. A merchandise recommendation system based on user attributes and merchandise types, the system comprising:

The user commodity associated information table creation module is used for creating a classified user commodity associated information table according to the user commodity transaction data;

the commodity purchase frequency statistics module is used for counting the total purchase frequency of each user class on each commodity; counting the associated purchase frequency of each user class to each commodity;

wherein, the commodity platform is provided with n commodities in total, and the collection G= { G is used ₁ ，G ₂ ，…G _n -representation; n represents the number of commodity types, also called commodity node number; there are k user categories in total, with the set c= { C ₁ ，c ₂ ，…c _k -representation;

user class c _x For commodity G _i The total purchase frequency is represented by a kxn matrix cf3= { (R3 (c) _x )) _i -representation;

user class c _x For commodity G _i The associated purchase frequency is represented by a kxn matrix cf2= { (R2 (c) _x )) _i -representation;

(R2(c _x )) _i ＝(R3(c _x )) _i -(R1(c _x )) _i the method comprises the steps of carrying out a first treatment on the surface of the Wherein (R1 (c) _x )) _i Representing user class c _x For commodity G _i Is a separate purchase frequency of (1); commodity G _i The individual purchase frequency indicates that only a single commodity G is purchased _i Frequency of (3); commodity G _i The related purchase frequency indicates that the purchased commodity contains the commodity G _i Frequency of (i.e. commodity G) _i Frequency of simultaneous purchase with all other merchandise;

the commodity total value calculation module is used for calculating the total value of each commodity according to the total purchase frequency of each commodity and the unit price of each commodity;

The weight calculation module is used for calculating the association weight, the frequency weight and the value weight of each commodity in each user category;

wherein ,

(W ₁ (c _x )) _i representing user class c _x Purchased commodity G _i Is a weight of the association of (a);

the directed commodity purchase frequency matrix creation module is used for counting the frequency of directed simultaneous purchase of commodities of each user category according to commodity transaction data of the user; creating a directed commodity purchase frequency matrix of each user category;

wherein, commodity G _i And G _j N×n matrix GG1 (c) for directed simultaneous purchase trigger relationship _x )＝{(R4(c _x )) _ij [ Z ] represents GG1 (c) _x ) Representing user class c _x A directed commodity purchase frequency matrix; (R4 (c) _x )) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Frequency of (3);

the undirected commodity purchase frequency matrix creation module is used for creating undirected commodity purchase frequency matrixes of all user categories according to the directed commodity purchase frequency matrixes of all user categories;

wherein user category c _x Is used for the undirected commodity purchase frequency matrix of (c) and is represented by an n x n matrix GG2 (c _x )＝{(R5(c _x )) _ij -representation; GG2 (c) _x ) Representing user class c _x Matrix of unoccupied commodity purchase frequency, (R5 (c) _x )) _ij Value table of (2)Showing user class c _x Commodity G in transaction data _i And G _j Frequency of simultaneous purchase;

The first normalization processing module is used for performing normalization processing on the directed commodity purchase frequency matrixes of all user categories to obtain corresponding directed commodity purchase weighting frequency matrixes;

the trigger center calculating module is used for calculating the trigger center of each commodity in each user category;

the calculating of the triggering centrality of each commodity of each user category comprises the following specific steps:

wherein ,(D₁ (c _x )) _i Representing user class c _x Purchased commodity G _i Is triggered by the trigger centrality of (2);

(CF(c _x )) _i representing user class c _x Purchased commodity G _i Trigger degree of (a), i.e. user category c _x Purchasing commodity G _i Triggering normalized frequency of purchasing other commodity types;

where i+.j, (F2) _x ) _ij Representing user class c _x Purchasing commodity G _i Time triggered purchase of commodity G _j Is a normalized frequency value of (2);

the network core degree calculating device is used for calculating the network core degree of each commodity of each user class;

the network core computing device comprises:

the commodity connection relation network construction module is used for constructing a commodity connection relation network according to the undirected commodity purchase frequency matrix;

the commodity is taken as a node, the connection relation between commodities is taken as an edge, the frequency of simultaneous purchase between commodities is taken as an edge weight, and a commodity connection relation network is constructed;

if commodity node G _i and G_j If the simultaneous purchase triggering relationship exists, the commodity node G _i and G_j A connecting edge is added between the two connecting edges; commodity node G _i and G_j There may also be an indirect connection between the two, connected by a path comprising at least 1 intermediate node and at least 2 connecting edges, at least one of the number of paths;

a node grade acquisition module for acquiring commodity node G with direct or indirect connection _i and G_j Obtaining commodity node G _j Opposite node G _i Node level lev of (a);

for commodity nodes G with direct or indirect connection _i and G_j Obtaining commodity node G _j Opposite node G _i Node level lev of (a);

set G _i As the original node, if the commodity node G _i and G_j If there is a direct connection relationship, node G _j Is G _i Is a level 1 node of (2);

if commodity node G _i and G_j There is indirect connection relation and connected by m node paths, m is more than or equal to 1, and p= { path is used for m node path collection ₁ ，path ₂ ，…G _m -representation; the number of connecting edges correspondingly contained in each node path in P is S= { side ₁ ，side ₂ ，…side _m Indicated, node G _j Opposite node G _i Node level lev=min (S), min (S) > 2, min () represents taking the minimum value;

the network position calculating module is used for calculating the network position of each commodity node of each user class;

Wherein, (FB (c) _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network location of (2); lev (Lev) _ij Representation and commodity node G _i Directly or indirectly connected commodity node G _j Is a node level of (2); (R3 (c) _x )) _j Representation ofWith commodity node G _i Directly or indirectly connected commodity node G _j The total purchase frequency of the corresponding commodity, namely the user class c _x For commodity G _j Is a total purchase frequency of (2); n represents the total number of commodity nodes;

the network core degree calculation module is used for calculating the network core degree of each commodity node of each user class;

wherein ,(H₁ (c _x )) _i Representing user class c _x Commodity node G of (1) _i Is a network core degree of (1);

the second normalization processing module is used for respectively carrying out normalization processing on the association weight, the frequency weight, the value weight, the trigger centrality and the network core degree of each commodity of each user category to obtain corresponding weighted association weight, weighted frequency weight, weighted value weight, weighted trigger centrality and weighted network core degree;

the optimal node degree calculation module is used for carrying out weighted summation on the weighted association weight, the weighted frequency weight, the weighted value weight, the weighted trigger center degree and the weighted network core degree to obtain the optimal node degree of the classified user commodity;

and the commodity recommendation module is used for recommending commodities to the corresponding classified users according to the optimal node degree of the commodities of the classified users.

11. An apparatus comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the commodity recommendation method according to any one of claims 1 to 9 based on user attributes and commodity types.

Images