CN110659924B

CN110659924B - Visual analysis method, device and equipment for product competition relationship

Info

Publication number: CN110659924B
Application number: CN201810712722.8A
Authority: CN
Inventors: 徐立鑫
Original assignee: Beijing Qihoo Technology Co Ltd
Current assignee: Beijing Qihoo Technology Co Ltd
Priority date: 2018-06-29
Filing date: 2018-06-29
Publication date: 2023-09-01
Anticipated expiration: 2038-06-29
Also published as: CN110659924A

Abstract

The invention discloses a visual analysis method, device and equipment for a product competition relationship, and relates to the technical field of computers. The method comprises the following steps: according to a preset audience user acquisition strategy of the product and preset audience user attributes, acquiring high-dimensional product coordinates and high-dimensional audience user attribute coordinates through a corresponding analysis method; based on the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates, calculating a distance matrix of each product and each attribute by adopting Euclidean distance; smoothing the distance matrix; adopting a multidimensional scale analysis method, and mapping the smoothed distance matrix into a coordinate matrix in a two-dimensional space in a dimension-reducing way; and drawing coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph, and displaying the competition relationship among products. The technical problem of the too concentrated or dispersion of coordinate point, lead to the graphic to demonstrate chaotic is solved. The method has the beneficial effects of accurately and effectively displaying the competition relationship among the products and facilitating the analysis and use of advertisers.

Description

Visual analysis method, device and equipment for product competition relationship

Technical Field

The invention relates to the technical field of computers, in particular to a visual analysis method, device and equipment for product competition relationship.

Background

Advertising the product through advertisements is a common means for promoting sales of the product, and an advertiser can analyze the product before advertising, analyze potential competing products of the product, analyze differences between the product and competing products and the like, so as to guide advertising behaviors of the advertiser.

When the product competition relationship is analyzed by adopting a corresponding analysis method at present, the relationship matrix of the product and the attributes of the audience users is obtained by determining the attributes such as the gender, the age and the like of the product audience users, the relationship between the product and the attributes is analyzed from the relationship matrix, the product coordinates and the attribute coordinates are determined, and then the graphical display is performed.

However, when the method is adopted for analysis, the product coordinates and the attribute coordinates output by the corresponding analysis method are found, and the scatter diagram displayed in the two-dimensional space directly appears in the condition that the coordinate points are too concentrated or scattered, wherein a plurality of coordinate points are far away from other points, so that the graphic display is disordered, the competition relationship among products cannot be accurately and effectively displayed, and the analysis and the use of advertisers are not facilitated.

Disclosure of Invention

The present invention has been made in view of the above problems, and it is an object of the present invention to provide a product competition relationship visualization analysis method, apparatus and device which overcome or at least partially solve the above problems.

According to one aspect of the present invention, there is provided a product competition relationship visual analysis method, comprising:

according to a preset audience user acquisition strategy of the product and preset audience user attributes, acquiring high-dimensional product coordinates and high-dimensional audience user attribute coordinates through a corresponding analysis method;

based on the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates, calculating a distance matrix of each product and each attribute by adopting Euclidean distance;

smoothing the distance matrix;

adopting a multidimensional scale analysis method, and mapping the smoothed distance matrix into a coordinate matrix in a two-dimensional space in a dimension-reducing way;

and drawing coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph, and displaying the competition relationship among products.

Optionally, the step of obtaining the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates through the correspondence analysis method according to the preset audience user obtaining strategy and the preset audience user attributes of the product includes:

Acquiring user data of an audience user of a product from back-end data according to a preset audience user acquisition strategy of the product;

determining a statistic value of the preset audience user attribute from the user data of the audience user;

establishing a relation matrix of the product and the audience user according to the statistic value;

and (3) calculating the relation between the product and the audience user attribute through corresponding analysis, and determining high-dimensional product coordinates and high-dimensional audience user attribute coordinates.

Optionally, the step of acquiring the user data of the audience user of the product from the back-end data according to a preset audience user acquisition policy of the product includes:

acquiring initial user data of an audience user of a product from back-end data according to one or more of preset searching behaviors, application downloading behaviors, application starting behaviors and website access behaviors;

and selecting user data meeting preset advertiser conditions from the initial user data of the audience users.

Optionally, the audience user attributes include system attributes and custom attributes;

the system attributes at least comprise at least one of gender, age, academic, occupation, purchasing power, region and interests; the custom attributes include one or more attributes that are custom by the advertiser.

Optionally, the step of smoothing the distance matrix includes:

taking the distance matrix as a distance matrix to be smoothed, and determining the variance of the distance matrix to be smoothed;

squaring the distance matrix to obtain a smoothed distance matrix after smoothing;

calculating a variance of the smoothed distance matrix;

if the variance of the distance matrix to be smoothed is greater than the variance of the smoothed distance matrix and the number of loops is less than a preset threshold, the smoothed distance matrix is continuously used as the distance matrix to be smoothed to execute smoothing until the variance of the distance matrix to be smoothed is less than or equal to the variance of the smoothed distance matrix or the number of loops is greater than or equal to the preset threshold.

Optionally, the step of mapping the smoothed distance matrix into a coordinate matrix in a two-dimensional space by using a multidimensional scale analysis method specifically includes:

determining a corresponding square matrix according to the distance matrix after the smoothing treatment;

based on the square matrix, performing center correction, and determining a scalar product matrix and m corresponding feature vectors;

arranging the eigenvectors into an eigenvector matrix E _m Forming a diagonal matrix Λ by taking m eigenvalues as diagonal elements;

according to the feature vectorMatrix E _m And determining a coordinate matrix Ω=e after dimension reduction by the diagonal matrix Λ _m Λ ^1/2 _m 。

Optionally, the drawing the coordinate points corresponding to the coordinate matrix in the two-dimensional space coordinate graph, and displaying the competition relationship between the products specifically includes:

acquiring a two-dimensional coordinate point based on the coordinate matrix;

and drawing the two-dimensional coordinate points in a two-dimensional space coordinate graph.

According to another aspect of the present invention, there is provided a product competition relationship visual analysis device including:

the coordinate acquisition module is used for acquiring high-dimensional product coordinates and high-dimensional audience user attribute coordinates through a corresponding analysis method according to an audience user acquisition strategy of a preset product and preset audience user attributes;

the distance matrix calculation module is used for calculating the distance matrix of each product and each attribute by adopting Euclidean distance based on the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates;

the smoothing processing module is used for carrying out smoothing processing on the distance matrix;

the mapping module is used for adopting a multidimensional scale analysis method to map the smoothed distance matrix into a coordinate matrix in a two-dimensional space in a dimension reduction way;

And the display module is used for drawing coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph and displaying the competition relationship among products.

Optionally, the coordinate acquisition module includes:

the user data acquisition sub-module is used for acquiring user data of an audience user of a product from the back-end data according to a preset audience user acquisition strategy of the product;

a statistic value determining submodule, configured to determine a statistic value of the preset audience user attribute from user data of the audience user;

a relation matrix establishing sub-module for establishing a relation matrix between the product and the audience user according to the statistic value;

and the coordinate determination submodule is used for determining high-dimensional product coordinates and high-dimensional audience user attribute coordinates by correspondingly analyzing and calculating the relation between the product and the audience user attributes.

Optionally, the user data acquisition sub-module includes:

the initial user data acquisition unit is used for acquiring initial user data of audience users of the product from the back-end data according to one or more of preset searching behaviors, application downloading behaviors, application starting behaviors and website access behaviors;

And the user data selection unit is used for selecting user data meeting preset advertiser conditions from the initial user data of the audience users.

Optionally, the smoothing processing module includes:

the first variance determining submodule is used for taking the distance matrix as a distance matrix to be smoothed and determining variances of the distance matrix to be smoothed;

the square opening sub-module is used for square opening the distance matrix to obtain a smoothed distance matrix after smoothing;

a second variance determining sub-module for calculating a variance of the smoothed distance matrix;

and the circulation smoothing sub-module is used for circulating to continue to serve as the distance matrix to be smoothed to execute smoothing processing until the variance of the distance matrix to be smoothed is smaller than or equal to the variance of the smoothed distance matrix or the circulation times are larger than or equal to a preset threshold value if the variance of the distance matrix to be smoothed is larger than the variance of the smoothed distance matrix and the circulation times are smaller than a preset threshold value.

Optionally, the mapping module includes:

a square matrix determining sub-module, configured to determine a corresponding square matrix according to the distance matrix after the smoothing process;

the feature vector determining submodule is used for executing center correction based on the square matrix and determining a scalar product matrix and m feature vectors corresponding to the scalar product matrix;

a diagonal matrix construction sub-module for arranging the eigenvectors into an eigenvector matrix E _m Forming a diagonal matrix Λ by taking m eigenvalues as diagonal elements;

a coordinate matrix determining sub-module for determining a coordinate matrix according to the eigenvector matrix E _m And determining a coordinate matrix Ω=e after dimension reduction by the diagonal matrix Λ _m Λ ^1/2 _m 。

Optionally, the display module includes:

the two-dimensional coordinate point acquisition sub-module is used for acquiring two-dimensional coordinate points based on the coordinate matrix;

and the coordinate graph drawing sub-module is used for drawing the two-dimensional coordinate points in a two-dimensional space coordinate graph.

According to another aspect of the present invention, there is provided a product competition relationship visual analysis apparatus comprising:

a memory loaded with a plurality of executable instructions;

a processor executing the plurality of executable instructions; the plurality of executable instructions includes a method of performing the steps of:

smoothing the distance matrix;

According to the visual analysis method for the product competition relationship, the obtained distance matrix can be processed smoothly on the basis of the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates output by the corresponding analysis method, the distance relationship between visual display time points is controlled, and the visual display time points are displayed after the dimension reduction processing. Therefore, the technical problem of disordered graph display caused by too concentrated or scattered coordinate points is solved. The method has the beneficial effects of accurately and effectively displaying the competition relationship among the products and facilitating the analysis and use of advertisers.

The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent.

Drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:

FIG. 1 shows a flow chart of steps of a method for visual analysis of product competition relationships according to one embodiment of the present invention;

FIG. 2 shows a flow chart of steps of a method for visual analysis of product competition relationships according to one embodiment of the present invention;

FIG. 3 is a schematic diagram showing the structure of a visual analysis device for product competition relationship according to one embodiment of the present invention;

FIG. 4 is a schematic diagram showing the structure of a visual analysis device for product competition relationship according to one embodiment of the present invention;

Fig. 5 shows a schematic structural diagram of a product competition relationship visual analysis device according to an embodiment of the present invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Example 1

The visual analysis method for the product competition relationship provided by the embodiment of the invention is described in detail.

Referring to fig. 1, a flow chart of steps of a visual analysis method for product competition relationship in an embodiment of the invention is shown.

Step 110, according to the preset audience user acquisition strategy and the preset audience user attribute of the product, obtaining high-dimensional product coordinates and high-dimensional audience user attribute coordinates through a corresponding analysis method.

After any product is released to the market, the audience users of the product are rich and various and have a certain rule. For example, a certain professional group may purchase and use a lot, and a certain age group may purchase and use a lot. Moreover, the user itself also has diverse attributes such as: various attributes that can describe the user's preference for products, such as age, gender, occupation, age, wedding, etc. Whether the advertisement is an own product or a potential competitive product, the advertisement has audience users with a certain scale, and when the advertisement is analyzed before the advertisement is put on the product, the advertisement can be combined with the characteristics of the own product according to factors such as the actual market sales condition of the product, the characteristics of a user group and the like, and the own product, the potential competitive product, the audience users and the attributes thereof which need to be analyzed are determined from the actual demand. Therefore, the audience user acquisition strategy and the audience user attribute are preset, the audience user acquisition strategy is a representation method and an acquisition means of the audience user, taking search words as an example, an advertiser can determine the audience user of the product by defining the search words and search conditions, for example, the audience user can be the audience user by setting people searching for certain keywords for more than a certain times. The preset user attributes include physiological attributes such as age, gender, academic, occupation and the like and various attributes in terms of social life. The audience user who meets the set conditions can be obtained by presetting an audience user acquisition strategy and an audience user attribute, and the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates are obtained by a corresponding analysis method.

Correspondence analysis is a multivariate statistical analysis method aimed at revealing the correlation between a variable and a sample or between a variable and its class in a qualitative variable profile. One of the important output results of the correspondence analysis is that the variables and the samples are simultaneously reflected on one graph of the same coordinate axis (factor axis), and the relationship among the variables, the relationship among the samples and the correspondence among the variables and the samples can be intuitively observed on the drawn graph by combining the calculation results. The essence of the correspondence analysis is that the cross table of the variable and the sample is transformed into a scatter diagram, so that the related information of the variable and the sample is expressed in the form of the spatial position relation of each scatter diagram. Along with the application of computer software, the corresponding analysis method has wide application value in the fields of social science and natural science. Particularly, in the market research and study in recent years, fields related to market segmentation, product positioning, brand image, satisfaction study and the like are getting more and more attention and application. Therefore, the invention adopts a corresponding analysis method to analyze the product and the audience user attributes of the product. Typically, the output of the correspondence analysis is reduced product coordinates and audience user attribute coordinates. However, the method is different from the operation of dimension reduction in the traditional correspondence analysis method, and the correspondence between the product and the audience user attribute is realized through the correspondence analysis method, so that the subsequent processing of the data is facilitated.

Step 120, calculating a distance matrix of each product and each attribute using the Euclidean distance based on the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates.

The Euclidean distance can be used for calculating the distance between any two points, measuring the tightness degree between two variables, and reflecting the relevance between the two variables to a certain extent. After the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates are obtained through a correspondence analysis method, the Euclidean distance can be used for calculating the distance matrix of each product and each attribute.

One representation of the distance matrix is given in table 1 below.

TABLE 1

	Product 1	Product 2	Attribute 1	Attribute 2
					Product 1	0	Distance of	Distance of	Distance of
Product 2	Distance of	0	Distance of	Distance of
					Attribute 1	Distance of	Distance of	0	Distance of
Attribute 2	Distance of	Distance of	Distance of	0

The distance matrix of each product and each attribute is obtained, and the relationships among the products, the attributes and among the products and the attributes can be described by the distance between the points.

And 130, performing smoothing processing on the distance matrix.

The input matrix of the corresponding analysis method is an original data matrix of each product and each audience user attribute, and is only the original statistical data, and the smoothing processing cannot be performed. The distance matrix output based on the correspondence analysis method can be smoothed. By carrying out smoothing treatment on the distance matrix, the distance relation between the visual display time points can be controlled on the premise of not influencing the relation between the coordinate points, and the problem of confusion of graphic display is solved.

And 140, adopting a multidimensional scale analysis method to map the smoothed distance matrix into a coordinate matrix in a two-dimensional space in a dimension reduction manner.

The smoothed distance matrix is still a high-dimensional matrix that cannot exhibit data relationships in a low-dimensional space. The multidimensional scale analysis method provides a dimensionality reduction algorithm for the invention, provides searching and finding a mapping to map data of a high-dimensional space into a low-dimensional space, and the mapping is to keep the distance between data points as unchanged as possible. Therefore, the smoothed high-dimensional distance matrix can be dimension-reduced and mapped into a coordinate matrix in a two-dimensional space by a multi-dimensional dimension analysis method, and the data relationship can be conveniently displayed in the two-dimensional space.

And 150, drawing coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph, and displaying the competition relationship among products.

After the two-dimensional coordinate matrix is obtained through multidimensional dimension analysis and dimension reduction, coordinate points corresponding to the coordinate matrix can be drawn in a two-dimensional space coordinate graph, and the competition relationship among products is displayed. Thus, advertisers can intuitively see the product-to-product, the attributes-to-attributes, and the relationships between the products and the attributes.

In the embodiment of the invention, on the basis of the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates output by the corresponding analysis method, the obtained distance matrix is subjected to smoothing processing, the distance relation between visual display time points is controlled, and the distance relation is displayed after the dimension reduction processing. The technical problem that the graph display is disordered due to the fact that the coordinate points are too concentrated or scattered is solved, the associated information of the scattered points can be accurately and reasonably reflected, the competition relationship among products is accurately and effectively displayed, and an advertiser can conveniently analyze the products before advertising.

Example two

Referring to fig. 2, a flow chart of steps of a visual analysis method for product competition relationship in an embodiment of the invention is shown.

Step 210, according to a preset audience user acquisition strategy of the product, acquiring user data of the audience user of the product from the back-end data.

Typically, various access records of the user in the network are stored in a background server, such as: a history of web browsing, a keyword record of searching, a record of online shopping, and the like. Therefore, once the audience user acquisition strategy is determined, the user data of the audience user meeting the preset conditions can be automatically screened and captured from the back-end data according to the strategy.

Optionally, in the embodiment of the present invention, the step 210 may specifically include the following sub-steps:

the method comprises the following substeps: and acquiring initial user data of audience users of the product from the back-end data according to one or more of preset searching behaviors, application downloading behaviors, application starting behaviors and website access behaviors.

Sub-step two: and selecting user data meeting preset advertiser conditions from the initial user data of the audience users.

The network using behaviors of the users are various, namely, products are searched by a search engine, corresponding programs are downloaded, a certain application program is opened, a certain product official network is directly browsed, and the like, accordingly, different data acquisition methods can be preset for the various network using behaviors, and initial user data of audience users of the products are acquired from the back-end data of the server according to one or more of the preset search behaviors, application downloading behaviors, application starting behaviors and website access behaviors, and the initial user data, namely, raw statistical data which is not processed, can provide huge data support for subsequent analysis, and can ensure the number of samples. From the initial user data, the data are further screened in combination with various attribute conditions preset by an advertiser, so that the user data which meet the preset advertiser conditions are selected, and the quality of the data for product analysis can be improved.

Step 211, determining a statistic value of the preset audience user attribute from the user data of the audience user.

The obtained user data of the audience users, including the audience user attributes of the products, can be used for counting various different attributes, and calculating the statistic value of each attribute, namely the occurrence times of the corresponding attribute in the audience user data. Such as: for the product A, the number of female users, the number of student users, the number of elderly users and the like can be counted respectively.

Optionally, in an embodiment of the present invention, the audience user attribute includes a system attribute and a custom attribute;

In order to maximize the coverage and breadth of user data, the audience user attributes may be customized by the advertiser, such as by a statement describing the audience user attributes, in addition to default system attributes (e.g., gender, age, academic, profession, purchasing power, territory, interests, etc.).

And 212, establishing a relation matrix of the product and the audience user according to the statistic value.

For each product, the attributes of the audience users can be counted, so that a statistic table of each attribute corresponding to each product is obtained, and a relationship matrix of the product and the audience users is established, wherein the relationship matrix represents the relationship between the product and the attributes of the audience users, and the specific dimension of the relationship matrix is shown by referring to table 2.

TABLE 2

	Attribute 1	Attribute 2	Attribute 3	……
					Product 1	Statistics value	Statistics value	Statistics value	Statistics value
Product 2	Statistics value	Statistics value	Statistics value	Statistics value
					……	Statistics value	Statistics value	Statistics value	Statistics value

In step 213, the relationship between the product and the audience user attribute is calculated by the correspondence analysis, and the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates are determined.

Typically, the output of the correspondence analysis is reduced product coordinates and audience user attribute coordinates. However, since the dimensions and orders of magnitude of the analyzed product and the audience user attribute are different and may be quite different, in order to make the two have the equality, the application adopts the thought of data transformation in the corresponding analysis method to normalize the original data, and does not adopt the dimension reduction method provided by the corresponding analysis method to reduce the dimension, so as to directly output the high-dimension product coordinate and the high-dimension audience user attribute coordinate. And then, the inherent connection of the product and the audience user attribute is found out by combining other methods, and the product and the audience user attribute can be simultaneously reflected on one graph with the same coordinate axis, so that analysis of the product competition relationship is facilitated.

For example, according to the preset condition, an original data matrix x= (X) containing n products with p audience user attributes is obtained _ij ) Transform it into a transition matrix z= (Z) _ij ) I.e. x _ij Conversion to z _ij Thereafter, z _ij Should be satisfied that the variables and samples are equivalent and can pass through z _ij The association of the product with the attributes of the audience user is established.

The specific transition matrix Z is obtained by transformation according to the following method, namely:

wherein, the liquid crystal display device comprises a liquid crystal display device,

this data transformation is based on the fact that χ is calculated when the independence check is performed on the list ² The statistics are obtained by inspiring the method. X-shaped articles ² The calculation formula of the statistics is:

to facilitate an understanding of the above data transformation, a further explanation is given below. The method comprises the steps of setting n products, wherein each product has p audience user attributes, and an original data matrix X is as follows:

assuming element X of the original data matrix X _ij If the sum is greater than 0, the proper number is added to all data, so that the requirement can be met, and then the row sum, column sum and sum of X are written out, and are respectively marked as X _i. ，x _.j And x _.. 。

here, x is _.. Denoted as T, dividing each element of the original data matrix X by the corresponding measure scale is changed to provide the audience user attributes with the same scale as the product, i.e. Obviously 0 < p _ij < 1, and->Thus P _ij Can be interpreted as a "probability" such that a normalized "probability" matrix p= (P) _ij ) _n×p . Similarly, the row sums, column sums of the P-array, respectively denoted as P, can be written out _i. ,p _.j 。

Wherein the method comprises the steps of

Based on the above p= (P _ij ) _n×p Both type R and type Q factor analysis may be performed:

1. type R factor analysis

For a product, if n products are considered points in p-dimensional space (i.e., p audience user attributes), then their coordinates are used for n productsThe representation is called n product points.

For audience user attributes, p audience user attributes can similarly be considered points of an n-dimensional space (i.e., n products), usingCoordinates representing p audience user attributes are referred to as p audience user attribute points.

It is a common method to express the relative proportions of the variables in the sample so that a study of the relationship between n samples can be translated into a study of the relationship between n sample points. If the sample is to be classified, the distance between the sample points can be used to characterize. If the Euclidean distance is introduced, the Euclidean distance between any two sample points k and l is as follows:

of course, in the process of calculating the Euclidean distance, in order to further eliminate the magnitude difference of each product point, for example, the kth product point has a larger magnitude, the influence of the difference of the action scale of the product point is raised when the distance is calculated. So that the coefficient can be reused Multiplying the distance formula to obtain a weighted distance formula, which is:

the above can also be regarded as coordinatesThe distance between two product points k and l of the n product points of (a). Further, the coordinates of each product point are written out, and the distance matrix of the product point after probability weighting can be obtained is as follows:

product points or attribute points can be classified by calculating the distance between the two points, but this cannot be graphically represented. In order to more intuitively represent the relationship between points, when a processing method of correspondence analysis is adopted, definition of the attribute point cooperative difference matrix needs to be given according to the weighted distance matrix. For this purpose, the average value of the ith attribute in the distance matrix is given as follows:

here, not arithmetic mean, but by probability p _i. Weighting can verify that the result of the above formula is not only the product average point coordinates, but also just the average value of each attribute point. Therefore, a synergistic difference array of attribute points in the product space can be written, namely, the synergistic difference array of the ith attribute and the jth attribute is as follows: a= (a) _ij ) Wherein, the method comprises the steps of, wherein,

wherein, the liquid crystal display device comprises a liquid crystal display device,(z _αi to alpha, i are peer-to-peer

α＝1,…,n；i＝1,…,p

Let z= (Z) _ij ) Then there is a=z 'Z, i.e. the covariance matrix of attribute points can be expressed in the form of Z' Z.

Therefore, the corresponding analysis is only needed to be carried out on the attribute from A=Z' Z, and the R-type factor load matrix corresponding to the attribute point is as follows:

2. q-factor analysis

Similarly, from the above-described obtained audience user attribute coordinates, the weighted distance between two attributes i and j can be calculated as:

similar to the above method, the product space can be writtenThe product point is a synergistic difference matrix, namely the synergistic difference matrix of the kth product and the ith product is B= (B) _kl ) Wherein, the method comprises the steps of, wherein,

thus there is b=zz ', i.e. the co-ordinated matrix of product points can be expressed in form of ZZ'.

For this reason, only need from B=ZZ' to carry out corresponding analysis to the product can, the Q type factor load matrix that the product point corresponds is:

in summary, when the original data array X is transformed into Z, the covariance arrays of the attribute points and the product points are a=z 'Z and b=zz', respectively. The matrix A and the matrix B obviously have simple corresponding relation and the original data x _ij Conversion to z _ij Thereafter, for i, j is peer-to-peer, i.e. z _ij Has a peer to peer property and product.

After the F matrix and the G matrix are obtained, coordinate points of p attributes and coordinate points of n products, namely, coordinate points of corresponding attributes of each behavior in the F matrix and coordinate points of products of each behavior in the G matrix can be obtained respectively.

Step 214, calculating a distance matrix for each product and each attribute using Euclidean distance based on the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates.

After the coordinates of the p attribute points and the n products in the m dimension are obtained, wherein m is greater than 2, the distances between the attributes and the products, between the products and between the attributes and the attributes can be calculated.

For example, the coordinate point of the product 1 isThe coordinate point of attribute 1 isThe euclidean distance between product 1 and attribute 1 is then:

thus, the distance matrix of table 1 can be obtained.

Step 215, taking the distance matrix as a distance matrix to be smoothed, and determining the variance of the distance matrix to be smoothed;

step 216, squaring the distance matrix to obtain a smoothed distance matrix after smoothing;

step 217, calculating the variance of the smoothed distance matrix;

step 218, if the variance of the distance matrix to be smoothed is greater than the variance of the smoothed distance matrix and the number of loops is less than the preset threshold, then looping the smoothed distance matrix to continue as the distance matrix to be smoothed to execute smoothing until the variance of the distance matrix to be smoothed is less than or equal to the variance of the smoothed distance matrix or the number of loops is greater than or equal to the preset threshold.

Steps 215 to 218 give a specific procedure for the distance matrix smoothing process, and steps 215 to 218 loop the process. Assuming that a distance matrix M is obtained according to step 214, the distance matrix M is taken as a distance matrix to be smoothed, and the variance of the distance matrix to be smoothed is calculated: m_variance, while the distance matrix M can be squared, results in a smoothed distance matrix N, i.e. n=sqrt (M).

To ensure the effect of the smoothing process, the variance of the smoothed distance matrix N is further calculated: n_variance. If M_variance is greater than N_variance and the number of loops is less than a preset threshold, the following steps are performed in a loop: 1. m=n; 2. n=sqrt (N); 3. calculating M_variance; 4. calculating N_variance; 5. the number of cycles is increased by 1. Until the variance of the distance matrix M to be smoothed is smaller than or equal to the variance of the smoothed distance matrix N, or the circulation times are larger than or equal to a preset threshold. When the preset threshold value can be set according to the size of the actual data amount, the embodiment of the present invention does not restrict this.

Step 219, determining a corresponding square matrix according to the distance matrix after the smoothing process;

in the process of performing multidimensional dimension analysis method dimension reduction mapping processing, assuming that the distance matrix obtained through the smoothing processing is D, further determining a square matrix D of the distance matrix D ⁽²⁾ 。

Step 220, based on the square matrix, performing center correction to determine a scalar product matrix and m feature vectors corresponding to the scalar product matrix;

based on the square matrix D obtained ⁽²⁾ Center correction is performed on it to obtain a scalar product matrix b= (-1/2) JD ⁽²⁾ J, where j=i _n -n ^-1 ee ^T ，e＝[1,1……,1] ^T (n 1 column vectors) the center correction uses a double center (double center) method. The process of performing the center correction corresponds to a translation of the coordinate point, and does not change the distance between points, i.e., does not change the relative relationship between points.

Based on the obtained scalar product matrix B, singular value decomposition is carried out on the scalar product matrix B to obtain the first m largest eigenvalues [ lambda ] of the scalar product matrix B ₁ ，λ ₂ ，……，λ _m ]And m feature vectors [ e ] corresponding thereto ₁ ，e ₂ ，……，e _m ]。

Step 221, arranging the feature vectors into a feature vector matrix E _m Forming a diagonal matrix Λ by taking m eigenvalues as diagonal elements;

arranging the obtained eigenvectors into an eigenvector matrix E _m And the diagonal matrix Λ is formed by taking m eigenvalues as diagonal elements.

Step 222, according to the feature vector matrix E _m And determining a coordinate matrix Ω=e after dimension reduction by the diagonal matrix Λ _m Λ ^1/2 _m 。

Based on the obtained eigenvector matrix E _m And a diagonal matrix Λ, the coordinate matrix Ω=e after dimension reduction can be calculated _m Λ ^1/2 _m

Therefore, the multidimensional scale analysis method tries to find a new set of substitute points in the space on the low dimension, such as 2 dimensions, according to the known distance between every two points in the high-dimensional space, so that the distance between every two points after dimension reduction is equal to the distance between every two points on the high dimension, and the display analysis of the data relationship in the low-dimensional space can be realized.

Step 223, acquiring a two-dimensional coordinate point based on the coordinate matrix;

and obtaining two-dimensional coordinate points of the products 1 to n, the attributes 1 to p for the coordinate matrix omega after the dimension reduction.

And 224, drawing the two-dimensional coordinate points in a two-dimensional space coordinate graph.

By using the two-dimensional coordinate points of the product and the attribute obtained by the process, the advertisement delivery system can draw the two-dimensional coordinate points in a two-dimensional space coordinate graph, and the data relationship is intuitively displayed.

In the embodiment of the invention, on the basis of the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates output by the corresponding analysis method, the obtained distance matrix is subjected to smoothing treatment, the distance relation between the visual display time points is controlled, the dimension reduction conversion from the high-dimensional coordinate points to the low-dimensional coordinate points is realized on the premise that the distance relation between the coordinate points is unchanged, the coordinate points subjected to optimization treatment are not too concentrated or scattered, and the graph can be clearly displayed. The advertiser can analyze the competition relationship between the product and the potential competition product, analyze the audience user attribute condition of the competition product, find the difference between the product and the competition product, and guide the advertiser to deliver the advertisement.

Example III

The embodiment of the invention provides a visual analysis device for product competition relationship.

Referring to fig. 3, a schematic structural diagram of a visual analysis device for product competition relationship according to an embodiment of the present invention is shown, including:

the coordinate acquisition module 310 is configured to acquire a high-dimensional product coordinate and a high-dimensional audience user attribute coordinate through a correspondence analysis method according to a preset audience user acquisition strategy and a preset audience user attribute of a product;

a distance matrix calculation module 320, configured to calculate a distance matrix of each product and each attribute by using the euclidean distance based on the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates;

a smoothing module 330, configured to smooth the distance matrix;

the mapping module 340 is configured to dimension-reduce the distance matrix after the smoothing processing into a coordinate matrix in a two-dimensional space by adopting a multidimensional scale analysis method;

and the display module 350 is configured to draw and display coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph, and display a competition relationship between products.

In the embodiment of the invention, on the basis of the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates output by the corresponding analysis method, the obtained distance matrix is subjected to smoothing processing, the distance relation between visual display time points is controlled, and the distance relation is displayed after the dimension reduction processing. Therefore, the technical problem of disordered graph display caused by too concentrated or scattered coordinate points is solved. The method has the beneficial effects of accurately and effectively displaying the competition relationship among the products and facilitating the analysis and use of advertisers.

Example IV

Referring to fig. 4, a schematic structural diagram of a visual analysis device for product competition relationship according to an embodiment of the present invention is shown, including:

the coordinate obtaining module 410 is configured to obtain a high-dimensional product coordinate and a high-dimensional audience user attribute coordinate according to a preset audience user obtaining policy and a preset audience user attribute of the product by a correspondence analysis method.

Optionally, in an embodiment of the present invention, the coordinate acquiring module 410 may include:

the user data acquisition sub-module 4101 is configured to acquire user data of an audience user of a product from back-end data according to a preset audience user acquisition policy of the product;

optionally, in an embodiment of the present invention, the user data obtaining submodule 4101 may include:

the initial user data acquisition unit is used for acquiring initial user data of audience users of the product from the back-end data according to one or more of preset searching behaviors, application downloading behaviors, application starting behaviors and website access behaviors; and the user data selection unit is used for selecting user data meeting preset advertiser conditions from the initial user data of the audience users.

A statistics determining sub-module 4102, configured to determine statistics of the preset audience user attribute from user data of the audience user;

A relationship matrix building sub-module 4103 for building a relationship matrix of products and the audience users according to the statistics;

the coordinate determination submodule 4104 is configured to determine high-dimensional product coordinates and high-dimensional audience user attribute coordinates by calculating a relationship between the product and the audience user attributes through correspondence analysis.

The distance matrix calculating module 411 is configured to calculate a distance matrix of each product and each attribute using the euclidean distance based on the high-dimensional product coordinates and the high-dimensional audience user attribute coordinates.

And a smoothing module 412, configured to smooth the distance matrix.

Optionally, in an embodiment of the present invention, the smoothing processing module 412 may include:

A first variance determining submodule 4121 configured to take the distance matrix as a distance matrix to be smoothed, and determine a variance of the distance matrix to be smoothed;

an open square sub-module 4122 for open square the distance matrix to obtain a smoothed distance matrix;

a second variance determining submodule 4123 for calculating variances of the smoothed distance matrix;

and a cyclic smoothing submodule 4124, configured to, if the variance of the distance matrix to be smoothed is greater than the variance of the smoothed distance matrix and the number of cycles is less than a preset threshold, cycle to continue the smoothed distance matrix as the distance matrix to be smoothed to perform smoothing until the variance of the distance matrix to be smoothed is less than or equal to the variance of the smoothed distance matrix, or the number of cycles is greater than or equal to the preset threshold.

The mapping module 413 is configured to dimension-reduce the distance matrix after the smoothing processing to a coordinate matrix in a two-dimensional space by adopting a multidimensional scale analysis method;

optionally, in an embodiment of the present invention, the mapping module 413 may include:

a square matrix determining submodule 4131 for determining a corresponding square matrix according to the distance matrix after the smoothing process;

A eigenvector determination submodule 4132 for performing center correction based on the square matrix to determine a scalar product matrix and m eigenvectors corresponding thereto;

diagonal matrix construction submodule 4133 for arranging the eigenvectors into an eigenvector matrix E _m Forming a diagonal matrix Λ by taking m eigenvalues as diagonal elements;

coordinate matrix determination submodule4134, a matrix E of feature vectors _m And determining a coordinate matrix Ω=e after dimension reduction by the diagonal matrix Λ _m Λ ^1/2 _m 。

And the display module 414 is used for drawing and displaying coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph, and displaying the competition relationship among products.

Optionally, in an embodiment of the present invention, the display module 414 may include:

a two-dimensional coordinate point acquisition submodule 4141 for acquiring a two-dimensional coordinate point based on the coordinate matrix;

a graph drawing submodule 4142 for drawing the two-dimensional coordinate point in a two-dimensional space graph.

Example five

The embodiment of the invention provides visual analysis equipment for product competition relationship.

Referring to fig. 5, a schematic structural diagram of a product competition relationship visualization analysis device according to an embodiment of the present invention is shown, including:

memory 501 loaded with a plurality of executable instructions;

a processor 502 executing the plurality of executable instructions; the plurality of executable instructions includes a method of performing the steps of:

smoothing the distance matrix;

and drawing and displaying the coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph, and displaying the competition relationship among products.

For the device embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments for relevant points.

The algorithms and displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general-purpose systems may also be used with the teachings herein. The required structure for a construction of such a system is apparent from the description above. In addition, the present invention is not directed to any particular programming language. It will be appreciated that the teachings of the present invention described herein may be implemented in a variety of programming languages, and the above description of specific languages is provided for disclosure of enablement and best mode of the present invention.

In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Similarly, it should be appreciated that in the foregoing description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. However, the disclosed method should not be construed as reflecting the intention that: i.e., the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. Thus, the claims following the detailed description are hereby expressly incorporated into this detailed description, with each claim standing on its own as a separate embodiment of this invention.

Those skilled in the art will appreciate that the modules in the apparatus of the embodiments may be adaptively changed and disposed in one or more apparatuses different from the embodiments. The modules or units or components of the embodiments may be combined into one module or unit or component and, furthermore, they may be divided into a plurality of sub-modules or sub-units or sub-components. Any combination of all features disclosed in this specification (including any accompanying claims, abstract and drawings), and all of the processes or units of any method or apparatus so disclosed, may be used in combination, except insofar as at least some of such features and/or processes or units are mutually exclusive. Each feature disclosed in this specification (including any accompanying claims, abstract and drawings), may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise.

Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

Various component embodiments of the invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. Those skilled in the art will appreciate that some or all of the functions of some or all of the components in a product competition relationship visual analysis device according to embodiments of the present invention may be implemented in practice using a microprocessor or Digital Signal Processor (DSP). The present invention can also be implemented as an apparatus or device program (e.g., a computer program and a computer program product) for performing a portion or all of the methods described herein. Such a program embodying the present invention may be stored on a computer readable medium, or may have the form of one or more signals. Such signals may be downloaded from an internet website, provided on a carrier signal, or provided in any other form.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be able to design alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word "comprising" does not exclude the presence of elements or steps not listed in a claim. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by one and the same item of hardware. The use of the words first, second, third, etc. do not denote any order. These words may be interpreted as names.

The application discloses a visual analysis method for a product competition relationship, which comprises the following steps:

smoothing the distance matrix;

A2, the method of A1, according to the preset audience user acquisition strategy and preset audience user attributes of the product, the step of acquiring high-dimensional product coordinates and high-dimensional audience user attribute coordinates through a corresponding analysis method comprises the following steps:

A3, the method of A2, according to the preset audience user obtaining strategy of the product, the step of obtaining the user data of the audience user of the product from the back-end data comprises the following steps:

A4, the method of A1, wherein the audience user attributes comprise system attributes and custom attributes;

A5, the method of A1, the step of smoothing the distance matrix, includes:

calculating a variance of the smoothed distance matrix;

A6, the method as described in A1, wherein the step of performing dimension reduction mapping on the smoothed distance matrix into a coordinate matrix in a two-dimensional space by using a multidimensional dimension analysis method specifically comprises the following steps:

According to the eigenvector matrix E _m And determining a coordinate matrix omega = after dimension reduction by the diagonal matrix lambdaE _m Λ ^1/2 _m 。

A7, the method according to A1, wherein the drawing of the coordinate points corresponding to the coordinate matrix in the two-dimensional space coordinate graph shows the competition relationship between the products, specifically includes:

acquiring a two-dimensional coordinate point based on the coordinate matrix;

The application also discloses a B8 visual analysis device for the product competition relationship, which comprises:

B9, the apparatus of B8, the coordinate acquisition module comprising:

B10, the apparatus of B9, the user data acquisition sub-module comprising:

B11, the device of B8, the audience user attributes including system attributes and custom attributes;

B12, the apparatus of B8, the smoothing module comprising:

B13, the apparatus of B8, the mapping module comprising:

B14, the device of B8, the display module comprising:

The application also discloses C15, a visual analysis device for the product competition relationship, which comprises:

a memory loaded with a plurality of executable instructions;

smoothing the distance matrix;

Claims

1. A method for visual analysis of product competition relationships, comprising:

smoothing the distance matrix;

Drawing coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph, and displaying the competition relationship among products;

the step of adopting a multidimensional scale analysis method to map the smoothed distance matrix into a coordinate matrix in a two-dimensional space in a dimension reduction way specifically comprises the following steps:

according to the eigenvector matrix E _m And determining a coordinate matrix Ω=e after dimension reduction by the diagonal matrix Λ _m Λ ^1/2 _m ；

The step of smoothing the distance matrix includes:

calculating a variance of the smoothed distance matrix;

2. The method of claim 1, wherein the step of obtaining high-dimensional product coordinates and high-dimensional audience user attribute coordinates by correspondence analysis based on the preset audience user acquisition policy and the preset audience user attributes comprises:

3. The method of claim 2, wherein the step of obtaining user data of the audience user of the product from back-end data according to a preset audience user obtaining policy of the product comprises:

4. The method of claim 1, wherein the audience user attributes comprise system attributes and custom attributes;

5. The method of claim 1, wherein the drawing the coordinate points corresponding to the coordinate matrix in the two-dimensional space coordinate graph shows the competition relationship between the products, and specifically includes:

acquiring a two-dimensional coordinate point based on the coordinate matrix;

6. A product competition relationship visual analysis device, characterized by comprising:

the display module is used for drawing coordinate points corresponding to the coordinate matrix in a two-dimensional space coordinate graph and displaying the competition relationship among products;

the mapping module comprises:

a coordinate matrix determining sub-module for determining a coordinate matrix according to the eigenvector matrix E _m And determining a coordinate matrix Ω=e after dimension reduction by the diagonal matrix Λ _m Λ ^1/2 _m ；

The smoothing module comprises:

7. The apparatus of claim 6, wherein the coordinate acquisition module comprises:

8. The apparatus of claim 7, wherein the user data acquisition sub-module comprises:

9. The apparatus of claim 6, wherein the audience user attributes comprise system attributes and custom attributes;

10. The apparatus of claim 6, wherein the display module comprises:

11. A product competition relationship visual analysis device, comprising:

a memory loaded with a plurality of executable instructions;

smoothing the distance matrix;

The step of smoothing the distance matrix includes:

calculating a variance of the smoothed distance matrix;