CN101697162A - Method and system for intelligently recommending ordering dishes - Google Patents

Method and system for intelligently recommending ordering dishes Download PDF

Info

Publication number
CN101697162A
CN101697162A CN 200910193222 CN200910193222A CN101697162A CN 101697162 A CN101697162 A CN 101697162A CN 200910193222 CN200910193222 CN 200910193222 CN 200910193222 A CN200910193222 A CN 200910193222A CN 101697162 A CN101697162 A CN 101697162A
Authority
CN
Grant status
Application
Patent type
Prior art keywords
dishes
attribute
association
main component
rule
Prior art date
Application number
CN 200910193222
Other languages
Chinese (zh)
Other versions
CN101697162B (en )
Inventor
祁亨年
谢文修
马文科
黄美丽
Original Assignee
杭州因豪信息科技开发有限公司;
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date

Links

Abstract

The invention relates to a method and a system for intelligently recommending ordering dishes. The method for intelligently recommending ordering the dishes comprises the following steps of: A, excavating attributes of all dishes to generate an association rule set, and calculating degree of confidence of each association rule of the association rule set; B, receiving a first dish, and storing the first dish into a database of ordered dishes; C, searching a matched attribute association rule set taking the attributes of the ordered dishes as an association rule antecedent and taking attribute of a dish X as an association rule consequent from the association rule set, and adding degree of confidence of each association rule in the acquired attribute association rule set together to acquire recommendation value of the dish X; D, sorting recommendation values, selecting N dishes serving as recommended dishes, and outputting the dishes; and E, judging whether the dishes input by a user are received, if so, storing the input dishes into the database of the ordered dishes, and continuing step C, otherwise, ending the recommendation. The system for intelligently recommending ordering the dishes comprises an association rule generating module, a dish receiving module, a recommendation value generating module, an output module and a judgment module. The method and the system can make the recommended dishes more scientific and reasonable.

Description

一种智能化推荐点菜方法及系统 An intelligent recommend ordering method and system

技术领域 FIELD

[0001] 本发明涉及餐饮业的电子点菜技术领域,特别是涉及一种智能化推荐点菜方法及系统。 [0001] Technical Field The present invention relates to an electronic ordering food industry, particularly to a method and system for ordering intelligent recommendation. 背景技术 Background technique

[0002] 现有技术中,餐饮业的电子点菜方法是利用传统的购物篮分析方法对餐饮业的交易记录进行数据挖掘。 [0002] prior art electronic ordering method is the restaurant industry catering industry transaction data mining market basket analysis using traditional methods. 用于传统的购物篮分析的数据库的数据项都是商品的名称本身。 Data items used in traditional basket analysis is the product name of the database itself. [0003] 中国发明专利申请,其申请号为200710046499. X,公开了一种"具有智能化推荐功能的电子点菜系统",这种具有智能化推荐功能的电子点菜系统仅简单的把菜品的名称作为数据项,是一种传统的购物篮分析方法,虽然这种菜品名称间的依赖关系能够对市场决策提供一些支持,但是由于这种依赖关系并不涉及商品的内在属性之间的依赖关系,而往往决策者所关心的是顾客已点的菜品具有什么样的属性,喜欢包括这种属性的菜品的人,还可能喜欢具有什么属性的其它菜品,因此,传统的购物篮分析方法的作用是有限的。 [0003] Chinese invention patent application, the application No. 200710046499. X, discloses an "electronic ordering system with intelligent recommendation function", which has the function of intelligent electronic ordering system recommended simply put dishes as the name of the data item, it is a traditional market basket analysis methods, although dependent on the relationship between the name of this dish can provide some support to the market decisions, but because of this dependency does not involve trade between the intrinsic properties of the dependencies, and often policy makers are concerned about the dishes customers have been the point of what kind of property, like people, including dishes such attributes may also like other dishes have what attributes, therefore, the traditional shopping basket analysis the role is limited.

发明内容 SUMMARY

[0004] 基于现有技术的不足,本发明需要解决的问题是:提供一种能够使推荐菜品更具科学化和合理化的智能化推荐点菜方法及系统。 [0004] Based on the deficiencies of the prior art, the present invention needs to solve is: to provide a more intelligent recommend the Recommended dishes a la carte method and system for scientific and rationalization.

[0005] 为解决上述问题,本发明提供了一种智能化推荐点菜方法,其包括以下步骤:[0006] A、根据获取的历史点菜数据库中所有菜品的属性数据,对所述所有菜品的属性数据之间进行数据挖掘生成关联规则集,所述关联规则集由多条关联规则组成,并计算所述关联规则集中每条关联规则的置信度; [0005] In order to solve the above problems, the present invention provides an intelligent recommendation ordering method comprising the steps of: [0006] A, ordering history database according to the acquired attribute data for all the dishes, the dishes all generating a data mining association rules between sets of attribute data, the association rule set consisting of a plurality of association rules, and calculating the degree of confidence associated rule set for each association rules;

[0007] 其中,所述关联规则由关联规则前件和关联规则后件组成,所述关联规则后件由所述关联规则前件推导得出,定义所述历史点菜数据库中所有菜品的属性数据的数据集为P,在数据集P中有Nl次出现所述关联规则前件的同时,又有N2次出现所述关联规则后件,则所述关联规则的置信度为N2/N1W00X,用于表示所述关联规则在置信度为N2/肌*100%的概率上是可信的; [0007] wherein the association rule by the association rules front member and rear member composed of association rules, after the association rule member by the first member of the association rules deduced, ordering history database defining the attributes of all the dishes dataset is P, there occurs while the association rule Nl front views of the member, there are N2 times after the association rules in the data set member P, then the association rule confidence level N2 / N1W00X, It is used to indicate the confidence of the association rule to be authentic in the N2 / * muscle probability of 100%;

[0008] B、接收用户输入的第一菜品,并将所述输入的第一菜品存储至已点菜品数据库;[0009] C、从所述关联规则集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种属性数据或者一种以上组合的属性数据为关联规则前件,以所述菜品X的一种属性数据或者一种以上组合的属性数据为关联规则后件的属性关联规则集,并将获得的所述属性关联规则集中的每条关联规则的置信度相加计算得到所述菜品X的推荐价值; [0008] B, a first dishes receive user input, and the input of the first point has been stored to the dishes dishes database; [0009] C, from the associated rule set, all of the dishes for X in dishes , to find a match in the database has been dishes point dot data an attribute of dishes or in combination one or more rules associated with the attribute data of the front member, the attribute data of the dishes in an X or a combination of one or more after the attributes associated rule set attribute data pieces of association rules, and adding the confidence associated attributes of each set of rules obtained in association rules recommended value of the calculated X, dishes;

[0010] 其中,X为所述所有菜品中的其中的一个菜品; [0010] wherein, X is a dish in which all the dishes in;

[0011] D、将所述所有菜品中的每一个菜品的推荐价值进行排序,选取推荐价值排名靠前的N个菜品作为推荐菜品,并输出所述N个推荐菜品,其中,所述N为大于0的整数;[0012] E、判断是否接收到用户输入的菜品,若是,则将所述输入的菜品存储至所述已点 [0011] D, the recommended value of all the dishes in a dish of each sort, recommend selecting the N highest-ranked value dishes as recommended dishes, and outputs the N Recommended dishes, wherein said N is integer greater than 0; [0012] E, determines whether a user input is received dishes, the dishes if stored, then the point has been input to the

5菜品数据库并继续步骤C ;否则,结束推荐点菜。 5 dishes Database and continue to step C; otherwise, the end of the recommended order.

[0013] 另,步骤C是针对所有菜品中的每一菜品都计算推荐价值。 [0013] Also, Step C is the recommended value are calculated for each dish in all the dishes. 步骤D中的输出N个推荐菜品的数量可以根据用户的要求进行更改,可以选取N的数量为5,若没有生成5个符合条件且推荐价值大于0的推荐菜品,则可以选取菜品推荐价值排序表中排名最高的菜品或者也可以是选取餐厅特别推荐的其它菜品,凑足5个菜品作为推荐菜品输出。 Number output N Recommended dishes step D may be changed according to the user's requirements, may be selected number N is 5, if not generated 5 Matched and the recommended value is greater than 0 recommended dishes, it is possible to select dishes recommendable ordering the highest ranking table dishes or may also be selected restaurant especially recommended other dishes, gather five dishes recommended dishes as output. 步骤E中的用户输入的菜品可以是从推荐菜品中选择的菜品,也可以是用户自选的菜品,还可以是餐厅推荐的常点菜品等。 Step E dishes user input may be selected from the recommended dishes dishes, may be user selectable dishes, may also often points Dining dishes and the like.

[0014] 优选的,所述所有菜品的属性数据包括以下属性类别:主成分属性和特征属性。 [0014] Preferably, all of the dishes attribute data includes attribute categories: a main component attributes and characteristics of attributes. 其中特征属性包括以下属性子类别:口味、烹饪方法、辅料、调料、菜系和产地等,还可以是根据用户的需求设定的其它属性子类别。 Wherein the attributes include the following attributes characteristic subcategories: taste, cooking methods, materials, seasoning, origin and the like cuisine, it may also be based on other properties of the needs of the sub-set of user categories. [0015] 更优选的,所述步骤A具体包括: [0015] More preferably, the step A comprises:

[0016] 根据获取的历史点菜数据库中所有菜品的主成分属性数据,对所述所有菜品的主成分属性数据之间进行数据挖掘生成主成分关联规则集,所述主成分关联规则集由多条主成分关联规则组成,并计算所述主成分关联规则集中每条主成分关联规则的置信度; [0017] 根据获取的历史点菜数据库中所有菜品的特征属性数据,对所述所有菜品的特征属性数据之间进行数据挖掘生成特征关联规则集,所述特征关联规则集由多条特征关联规则组成,并计算所述特征关联规则集中每条特征关联规则的置信度; [0018] 所述步骤C具体包括: [0016] The ordering history database acquired attribute data for all the main component of the dishes, a main component to generate data mining association rules between sets attribute data of the main component for all the dishes, the main component by a multi-association rule set Article of rules associated with the main component, and calculating the principal components associated with each ruleset main component confidence association rules; [0017] the attribute data acquired in the history database ordering all features dishes, the dishes all between feature attribute data for generating feature data mining association rules set, wherein the association rule set of a plurality of feature association of rules, and calculating said confidence feature association rule set of association rules for each feature; [0018] the step C comprises:

[0019] Cl、对所述主成分属性和所述特征属性分别设立主成分属性权值函数和特征属性权值函数; [0019] Cl, the main component of the characteristic attributes and attribute set up a main component attribute weight function and the weight function characteristic attribute;

[0020] C2、从所述主成分关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点 [0020] C2, the main component from the associated set, for all the dishes in the dish X, has been to find a match to the point

菜品数据库中的已点菜品的一种主成分属性数据或者一种以上组合的主成分属性数据为 Dishes main component attribute data in the database have a master-point component attribute data dishes or one or more combinations of

关联规则前件,以所述菜品X的一种主成分属性数据或者一种以上组合的主成分属性数据 Association rule antecedent attribute data in a main component of the dish X or one or more combinations of a main component attribute data

为关联规则后件的主成分属性关联规则集,并将获得的所述主成分属性关联规则集中的每 As a main component attributes associated rule set of association rules rear member and to focus the primary component attribute association rules obtained per

条主成分关联规则的置信度与主成分权值相乘,即得到所述所有菜品中的菜品X的主成分 Confidence that the main component by multiplying the weight of association rules article main component, i.e., all of the dishes to give a main component in the X dishes

属性推荐置信度,其中,所述主成分权值是根据所述主成分属性权值函数计算得到; Recommended confidence attribute, wherein the main component is calculated according to the weight of the main component attribute weight function;

[0021] C3、从所述特征关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜 [0021] C3, from the feature correlation set, for all the dishes in the dish X, has been to find a match to the a la carte

品数据库中的已点菜品的一种特征属性数据或者一种以上组合的特征属性数据为关联规则前件,以所述菜品X的一种特征属性数据或者一种以上组合的特征属性数据为关联规则后件的特征属性关联规则集,并将获得的所述特征属性关联规则集中的每条特征关联规则的置信度与特征权值相乘,即得到所述所有菜品中的菜品X的特征属性推荐置信度,其中, 所述特征权值是根据所述特征属性权值函数计算得到; A characteristic feature attribute data has attribute data points dishes or product database and one or more combinations of feature attribute data for association rule antecedent to a characteristic of the attribute data X dishes or one or more combinations of related and wherein the confidence weights each associated feature attributes associated rule after rule member wherein the rule set, and the focus characteristic attribute association rule obtained by multiplying, i.e., all the dishes to obtain the characteristic properties of the dishes X recommended confidence, wherein said characteristic weights are calculated according to the weight function characteristic attribute;

[0022] C4、将所述菜品X的主成分属性推荐置信度与特征属性推荐置信度相加,即得到所述菜品X的推荐价值。 [0022] C4, the main component of the attribute X dishes recommended recommended confidence confidence characteristic attribute added to obtain the recommended value of X dishes.

[0023] 另,所述主成分权值与所述特征权值还可以是预先设置。 [0023] Also, the feature value of the weight main component may also be pre-set weight.

[0024] 进一步的,所述步骤C1中的主成分属性权值函数和特征属性权值函数均为递增函数。 [0024] Further, the main component of step C1 and the attribute weight function characteristic attribute weights are a function of an increasing function.

[0025] 更进一步的,所述步骤Cl中的主成分属性权值函数为Y二Fl(tl) 二tl衬l,其中, 所述主成分属性权值函数的自变量tl的计算方法为,针对每条主成分关联规则,在关联规 [0025] Further, the main component of the attribute weights are a function of the step is Cl Y = Fl (tl) tl two lining L, wherein, calculated from the variables tl main component attribute is weight function, for each of the main components of association rules, the association rules

6则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值; A main component attribute member 6 in front of the point number of dishes have occurred and the properties of the main component and the number of dishes point has occurred ratio;

[0026] 所述特征属性权值函数Y二F2(t2) = t2衬2,其中,所述特征属性权值函数的自变 [0026] The characteristic attribute weight function Y = F2 (t2) = t2 liner 2, wherein the characteristic attribute of the independent variable weight function

量t2的计算方法为,针对每条特征关联规则,在关联规则前件中的特征属性在所述已点菜 T2 is an amount calculated for each feature association rule, wherein the rule association property antecedent has the ordering

品中出现的序号之和与所述特征属性在所述已点菜品中出现的次数之和的比值。 And the number of characteristic attributes of the article appearing in the number of dishes and the ratio has occurred points.

[0027] 另,所述主成分属性权值函数和所述特征属性权值函数为相同的权值函数都为Y [0027] Also, the main component attribute weight function and the weight function is a characteristic property of the weight functions are the same as Y

=F(t) = t*t。 = F (t) = t * t. 设立主成分属性权值函数和特征属性权值函数是为了标识某条关联规则在 Establish a main component attribute weight function and the weight function is a characteristic attribute for identifying association rules in a strip

推荐菜品过程中的重要程度,而选取主成分属性权值函数与特征属性权值函数相同,是考虑到主成分属性与特征属性的重要性相当。 Recommended dishes importance of the process, and wherein the same value of the function selection, taking into account the properties of the main component attributes and characteristics of a main component attribute considerable importance weights attribute weight function.

[0028] 另一更进一步的,所述步骤C1中的主成分属性权值函数为Y二Fl(tl)= tl衬l衬l,其中,所述主成分属性权值函数的自变量tl的计算方法为,针对每条主成分关联规则,在关联规则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值; [0028] Another still further, the main component of step C1, the attribute weight function is Y = Fl (tl) = tl l liner lining l, wherein the main component attribute argument weight function of tl calculation method, the association rules for each principal component, a main component of the rule association property of the antecedent point number has appeared in the dishes and the dishes primary frequency component attribute has occurred in the point the ratio of the sum;

[0029] 所述特征属性权值函数Y二F2(t2) = t2衬2,其中,所述特征属性权值函数的自变 [0029] The characteristic attribute weight function Y = F2 (t2) = t2 liner 2, wherein the characteristic attribute of the independent variable weight function

量t2的计算方法为,针对每条特征关联规则,在关联规则前件中的特征属性在所述已点菜 T2 is an amount calculated for each feature association rule, wherein the rule association property antecedent has the ordering

品中出现的序号之和与所述特征属性在所述已点菜品中出现的次数之和的比值。 And the number of characteristic attributes of the article appearing in the number of dishes and the ratio has occurred points.

[0030] 另,设立主成分属性权值函数和特征属性权值函数是为了标识某条关联规则在推 [0030] Also, the establishment of a main component attribute weight function and the weight function is a characteristic attribute for identifying association rules in a push bar

荐菜品过程中的重要程度,而选取主成分属性权值函数与特征属性权值函数不相同,是考 Recommended dishes importance of the process, and the value of the function select component main weight function characteristic attribute property rights are not the same, a test

虑到主成分属性与特征属性的重要性不相当,当认为主成分属性的重要性大于特征属性的 Into account the importance of a main component attribute does not correspond to the characteristic property, when the main component of the importance of that attribute is greater than the characteristic attribute

重要性时,主成分属性权值函数Fl (t),以及特征属性权值函数F2 (t),并同时使得这两个 When the importance of, the main component attribute weight function Fl (t), and the feature attribute weight function F2 (t), and at the same time so that the two

函数满足如下要求: Function satisfies the following requirements:

[0031] 对于任意的t, Fl (t) > F2(t); [0031] for any t, Fl (t)> F2 (t);

[0032] 当认为主成分属性的重要性小于特征属性的重要性时,主成分属性权值函数Fl (t),以及特征属性权值函数F2 (t),并同时使得这两个函数满足如下要求: [0033] 对于任意的t, Fl (t) < F2 (t); [0032] When the main component of the importance of that attribute is less than the importance of the characteristic properties of the main component attribute weight function Fl (t), and the feature attribute weight function F2 (t), and such that the two functions simultaneously satisfy the following requirements: [0033] for any t, Fl (t) <F2 (t);

[0034] 另,以上对于主成分属性权值函数中的自变量的计算方法,可以是简单的针对每条主成分关联规则,在关联规则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值这种简单的计算,还可以是对已点菜品出现的序号分别赋菜品权值,采用在关联规则前件中的主成分属性在所述已点菜品中出现的序号与之相对应的菜品权值的乘积之和再与所述主成分属性在所述已点菜品中出现的次数之和的比值这种加权求和再求平均的算法,还可以是预先设定的一个递增数列的固定值,还可以是其他的计算方法。 [0034] Also, the above method for calculating the main component of the attribute weight argument function, may be a simple association rules for each principal component, a main component of the rule association property of said front member has points of dishes and a number of the main component attribute appearing in this simple calculation of the number and the ratio of the dishes has occurred point, it may also have a number of points appearing dishes were assigned weights dishes, using the associated main component attribute rule in the antecedent was the point number appears in dishes with the corresponding product of the weights and then the dishes with the main component of the frequency points have attribute dishes and appearing in the this ratio is then weighted summation averaging algorithm may also be a preset fixed incremental value of the number of columns may be other calculation methods. 对于特征属性权值函数中的自变量的计算方法也是同理的。 Method for calculating the weight function characteristic attribute of the argument is the same reason.

[0035] 本发明还提供了一种智能化推荐点菜系统,其包括: [0035] The present invention further provides an intelligent recommendation ordering system, comprising:

[0036] 关联规则生成模块,用于根据获取的历史点菜数据库中所有菜品的属性数据,对所述所有菜品的属性数据之间进行数据挖掘生成关联规则集,所述关联规则集由多条关联规则组成,并计算所述关联规则集中每条关联规则的置信度; [0036] association rule generation module for ordering history database acquired attribute data for all the dishes, the data attribute data between all the dishes generate mining association rules set, the association rule set consists of a plurality of association of rules, and calculating the degree of confidence associated rule set for each association rules;

[0037] 其中,所述关联规则由关联规则前件和关联规则后件组成,所述关联规则后件由所述关联规则前件推导得出,定义所述历史点菜数据库中所有菜品的属性数据的数据集为P,在数据集P中有Nl次出现所述关联规则前件的同时,又有N2次出现所述关联规则后件,则所述关联规则的置信度为N2/N1W00X,用于表示所述关联规则在置信度为N2/ 肌*100%的概率上是可信的; [0037] wherein the association rule by the association rules front member and rear member composed of association rules, after the association rule member by the first member of the association rules deduced, ordering history database defining the attributes of all the dishes dataset is P, there occurs while the association rule Nl front views of the member, there are N2 times after the association rules in the data set member P, then the association rule confidence level N2 / N1W00X, It is used to indicate the confidence of the association rule to be authentic in the N2 / * muscle probability of 100%;

[0038] 第一菜品接收模块,用于接收用户输入的第一菜品,并将所述输入的第一菜品存储至已点菜品的数据库; [0038] a first receiving module dishes, the dishes for receiving a first user input, and the input of the first point has been stored to the dishes dishes database;

[0039] 推荐价值生成模块,用于从所述关联规则集中,针对所述所有菜品中的菜品X,寻 [0039] Recommended value generation module for centrally from the association rule, for all the dishes in the dish X, seek

找匹配以所述已点菜品数据库中的已点菜品的一种属性数据或者一种以上组合的属性数 Get the number of attributes to match the database has been dishes point dot data an attribute of dishes or in combination one or more

据为关联规则前件,以所述菜品X的一种属性数据或者一种以上组合的属性数据为关联规 According to association rule antecedent attribute data of the dishes in an X, or a combination of one or more data attributes for association rules

则后件的属性关联规则集,并将获得的所述属性关联规则集中的每条关联规则的置信度相 After the attribute member association rule sets, and the confidence of the rule association property is obtained for each of the set of rules associated with

加计算得到所述菜品X的推荐价值; Adding said calculated recommended value of X dishes;

[0040] 其中,X为所述所有菜品中的其中的一个菜品; [0040] wherein, X is a dish in which all the dishes in;

[0041 ] 输出模块,用于将所述所有菜品的推荐价值进行排序,选取推荐价值排名靠前的N [0041] Output module, the recommended value for all sorts of dishes, the recommended value of the selected top-ranking N

个菜品作为推荐菜品,并输出所述N个推荐菜品,其中,所述N为大于0的整数; As a recommended dishes dishes, and outputs the N Recommended dishes, wherein said N is an integer greater than 0;

[0042] 判断模块,用于判断是否接收到用户输入的菜品,若是,则将所述输入的菜品存储 [0042] determination means for determining whether user input is received dishes, the dishes if stored, then the input

至所述已点菜品的数据库,并向推荐价值生成模块发送寻找匹配的执行指令;否则,结束推 The dishes have to point the database, send for a matching execution of instructions and recommend value generation module; otherwise, end push

荐点菜。 Recommended a la carte.

[0043] 优选的,所述所有菜品的属性数据包括以下属性类别:主成分属性、特征属性。 [0043] Preferably, all of the dishes attribute data includes attribute categories: a main component attribute, wherein properties. [0044] 更优选的,所述关联规则生成模块包括: [0044] More preferably, the association rule generation module comprises:

[0045] 主成分属性数据的关联规则生成单元,用于根据获取的历史点菜数据库中所有菜品的主成分属性数据,对所述所有菜品的主成分属性数据之间进行数据挖掘生成主成分关联规则集,所述主成分关联规则集由多条主成分关联规则组成,并计算所述主成分关联规则集中每条主成分关联规则的置信度; [0045] The main component of association rules attribute data generation unit for, among the data of the attribute data for all the dishes ordering history database according to the main component in the main component of the acquired attribute data associated with all of the dishes generate a main component Mining rule set, the rule set associated with the main component of a plurality of rules associated with the main component composition, and calculating the principal components associated with each ruleset main component confidence association rules;

[0046] 特征属性数据的关联规则生成单元,用于根据获取的历史点菜数据库中所有菜品的特征属性数据,对所述所有菜品的特征属性数据之间进行数据挖掘生成特征关联规则集,所述特征关联规则集由多条特征关联规则组成,并计算所述特征关联规则集中每条特征关联规则的置信度; Association rules [0046] feature attribute data generating means, based on the history database ordering attribute data acquired in the characteristic of all the dishes, attribute data between characteristic data of the all the dishes mining association rules generate a feature set, the wherein said association rule characterized by a plurality of sets of rules is associated, and wherein the association rule set is calculated for each feature confidence association rule;

[0047] 所述所述推荐价值生成模块包括: [0047] The value of the recommendation generation module comprises:

[0048] 权值分配单元,用于对主成分属性和特征属性分别设立主成分属性权值函数和特征属性权值函数; [0048] The weight allocating unit, for the main component attributes and characteristics of the main component attribute properties were established weight function and the weight function characteristic attribute;

[0049] 主成分属性推荐置信度生成单元,用于从所述主成分关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则前件,以所述菜品X的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则后件的主成分属性关联规则集,并将获得的所述主成分属性关联规则集中的每条主成分关联规则的置信度与主成分权值相乘,即得到所述所有菜品中的菜品X的主成分属性推荐置信度,其中,所述主成分权值是根据所述主成分属性权值函数计算得到; [0049] The main component attribute recommended confidence generating unit for correlation set from said main component for all the dishes in the dish X, looking for a match to the main points have been dishes point database dishes main component composition attribute data attribute data or attribute data of one or more principal components combined correlation rule antecedent, the dishes in a main component attribute data X or one or more thereof as a main component attribute association rules rear member association rule sets, and the confidence weights and the principal component is obtained by multiplying the main component attributes associated rule set of association rules for each principal component, a main component to obtain the attributes of all the dishes in the dish recommendation X confidence degrees, wherein the main component is calculated according to the weight of the main component attribute weight function;

[0050] 特征属性推荐置信度生成单元,用于从所述特征关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种特征属性数据或者一种以上组合的特征属性数据为关联规则前件,以所述菜品X的一种特征属性数据或者一种以上 [0050] Recommended confidence characteristic attribute generating unit for correlation set from the feature for all the dishes in the dish X, wherein attribute data to find a match to the database has been dishes point dishes point or a combination of one or more feature attribute data for association rule antecedent to the features of the dishes or the attribute data X is one or more

8组合的特征属性数据为关联规则后件的特征属性关联规则集,并将获得的所述特征属性关联规则集中的每条特征关联规则的置信度与特征权值相乘,即得到所述所有菜品中的菜品X的特征属性推荐置信度,其中,所述特征权值是根据所述特征属性权值函数计算得到; [0051] 推荐价值生成单元,用于将所述菜品X的主成分属性推荐置信度与特征属性推荐置信度相加,即得到所述菜品X的推荐价值。 Wherein attribute data for the combined 8 wherein the association rule element attributes associated rule set, multiplying the confidence weights and characterized in the characteristic attributes and the obtained association rule set association rules for each feature, i.e., to obtain all the recommended dishes confidence characteristic properties of the X dishes, wherein said characteristic weights are calculated according to the weight function characteristic attribute; [0051] recommended value generating unit, a main component for the attributes of the dishes X recommended confidence confidence recommendation characteristic attribute added to obtain the recommended value of X dishes.

[0052] 本发明利用关联规则数据挖掘的方法对历史点菜数据库中所有菜品的属性数据之间进行数据挖掘生成关联规则集,并计算关联规则集中每条关联规则的置信度;接收用户输入的第一菜品,并将输入的第一菜品存储至已点菜品数据库;从关联规则集中,针对所有菜品中的菜品X,寻找匹配以已点菜品数据库中的已点菜品的一种属性数据或者一种以上组合的属性数据为关联规则前件,以菜品X的一种属性数据或者一种以上组合的属性数据为关联规则后件的属性关联规则集,并将获得的属性关联规则集中的每条关联规则的置信度相加计算得到菜品X的推荐价值;将所有菜品中的每一个菜品的推荐价值进行排序, 选取推荐价值排名靠前的N个菜品作为推荐菜品,并输出N个推荐菜品,其中,N为大于O的整数;判断是否接收到用户输入的菜品,若是,则将输入 Method [0052] The present invention makes use of data mining association rules is performed between ordering history database and attribute data for all the dishes generate data mining association rules set, and calculates the concentration degree of confidence for each association rule association rule; receiving a user input first dishes, and inputs the first point has been stored to the dishes dishes database; concentrated from association rule, for all the dishes dishes X, for a matching point has an attribute data to the database dishes have a point or dishes attribute data for the combination of the above association rule antecedent attribute data attribute data in a dish, or X is a combination of one or more attributes associated rule set of the member association rules, and the rules associated with the attribute set of each of the obtained confidence association rules is calculated by adding the value of the recommended dishes of X; and a recommended value of each dish in all sorts of dishes, selected recommendation ranking value as the N recommended dishes dishes, and output N recommended dishes, wherein, N is an integer greater than O; dishes to determine whether the received user input, and if yes, enter 菜品存储至已点菜品数据库并继续步骤C;否则,结束推荐点菜。 Dishes stored in the database have carte menu and continue with step C; otherwise, the end of the recommended order. 与现有技术相比较,本发明能够使推荐的菜品更具科学性和合理性。 Compared with the prior art, the present invention enables a more scientific and rational recommended dishes. 本发明中将菜品的属性具体分为主成分属性和特征属性,用来反映菜品内在的本质特征,使得生成的关联规则更具科学性,为决策者提供真正有建设意义的决策参考。 The present invention will be divided into the main dishes attribute specific component properties and characteristics of properties to reflect the intrinsic nature features dishes such association rules generated more scientific and provide a reference for the truly constructive policy makers. 本发明中针对不同的属性还分别设立了递增的权值函数,用来逐步减小先前已点菜品对候选菜品的重复性影响。 The present invention also for different attributes were set up incremental weights function to gradually reduce repetitive impact point dishes previously candidate dishes.

附图说明 BRIEF DESCRIPTION

[0053] 利用附图对本发明作进一步说明,但附图中的实施例不构成对本发明的任何限制。 [0053] using the drawings The present invention is further illustrated, but the embodiments in the drawings do not constitute any limitation on the present invention.

[0054] 图1为本发明一种智能化推荐点菜方法在一个优选实施例中的方法流程图; [0055] 图2为本发明一种智能化推荐点菜系统在一个优选实施例中的结构示意图; [0056] 图3为本发明一种智能化推荐点菜系统在一个优选实施例中的详细的结构示意图。 [0054] Figure 1 is an intelligent invention à la flowchart of a method recommended in the method of a preferred embodiment embodiment; [0055] FIG. 2 of the present invention, an intelligent recommendation ordering system in a preferred embodiment of the structural diagram; detailed schematic diagram of a configuration of an intelligent recommendation ordering system in a preferred embodiment [0056] FIG. 3 of the present invention.

具体实施方式 detailed description

[0057] 结合以下实施例对本发明作进一步描述: [0058] 实施例一: [0057] The following embodiments in conjunction with the present invention is further described: [0058] Example a:

[0059] 本发明的一种智能化推荐点菜方法的实施例如图1所示,为本发明的一种方法流程图。 [0059] The embodiment of an intelligent recommended ordering method of the invention, for example, as shown in FIG. 1, a flowchart of a method of the present invention.

[0060] 具体的,一种智能化推荐点菜方法,包括以下步骤: [0060] Specifically, an intelligent recommendation ordering method comprising the steps of:

[0061] 步骤S01、根据获取的历史点菜数据库中所有菜品的属性数据,对所述所有菜品的属性数据之间进行数据挖掘生成关联规则集,所述关联规则集由多条关联规则组成,并计算所述关联规则集中每条关联规则的置信度; [0061] Step S01, the history of ordering the database according to the acquired attribute data for all the dishes, the data of the attribute data between all of the dishes generate mining association rules set, the association rule set composed of a plurality of association rules, and calculating confidence of the association rule set for each of the association rules;

[0062] 其中,所述关联规则由关联规则前件和关联规则后件组成,所述关联规则后件由所述关联规则前件推导得出,定义所述历史点菜数据库中所有菜品的属性数据的数据集为P,在数据集P中有Nl次出现所述关联规则前件的同时,又有N2次出现所述关联规则后件,则所述关联规则的置信度为N2/N1W00X,用于表示所述关联规则在置信度为N2/ 肌*100%的概率上是可信的; [0062] wherein the association rule by the association rules front member and rear member composed of association rules, after the association rule member by the first member of the association rules deduced, ordering history database defining the attributes of all the dishes dataset is P, there occurs while the association rule Nl front views of the member, there are N2 times after the association rules in the data set member P, then the association rule confidence level N2 / N1W00X, It is used to indicate the confidence of the association rule to be authentic in the N2 / * muscle probability of 100%;

[0063] 调取原始的事务数据库(可以是某餐饮店真实的历史消费记录),在原始的事务 [0063] the transfer of the original transaction database (restaurants may be a true historical record of consumption), the original transaction

数据库中对数据进行预处理,首先剔除用户不感兴趣的饮品和非主食的数据项,经处理后 After preprocessing of the data in the database, and the first non-staple drinks excluding items not of interest to the user, the treated

的数据库可以定义为历史点菜数据库,然后利用餐饮业领域的知识得到历史点菜数据库中 The database can be defined as a la carte history database, and then use the knowledge in the field of history a la carte restaurant industry to get the database

每一个菜品的属性数据,这里的属性数据可以根据餐饮业领域的客观知识判断得出,也可 Attribute data for each dish, where the attribute data can be judged according to objective knowledge in the field of food and beverage industry results, also

以是根据厨师等专家的主观判断得出,然后将获取的历史点菜数据库中所有菜品的属性数 All the dishes so the number of properties based on subjective judgments derived chefs and other experts, then get a la carte history database

据导入计算机,再对所有菜品的属性数据之间进行数据挖掘生成关联规则。 According to import computer, and then generate a data mining association rules between the attribute data of all the dishes.

[0064] 关联规则挖掘是数据挖掘的典型方法。 [0064] The method of mining association rules is a typical data mining. 关联规则是指从大量的数据中挖掘出的描 Refers to the association rule is excavated from a large number of data described

述数据项之间的关系的有无、表示数据库中的一组对象之间的某种关联关系的规则。 The relationship between the presence or absence of said data items, some association rule indicates the relationship between the database and a set of objects. 本发 Present

明并不是简单的将菜品的名称作为数据项进行数据挖掘,而是对菜品的内在的属性数据作 Ming is not simply the name of the dishes for data mining as a data item, but the intrinsic properties of the data for the dishes

为数据项进行数据挖掘,这种数据挖掘生成的关联规则可以得出真正有意义的规则或者说 Data mining is a data item that generated the data mining association rules can draw meaningful rules or

是内在的依赖关系,即菜品的特性之间的依赖关系。 Is the inherent dependencies, namely the dependencies between features dishes. 另外,随着原始事物数据库的更新,重 In addition, with the original update the database of things, heavy

新获取历史点菜数据库中菜品的属性数据,从而对所述菜品的属性数据进行定期的数据挖 Newly acquired property data history database a la carte dishes, which attribute data of the dishes of regular data mining

掘,使得推荐菜品的功能跟得上时代的需求。 Dig, making the needs of the times to keep up with the function of the recommended dishes.

[0065] 另,引进一个具体的例子对关联规则中的几个名称的进行解释: [0065] Also, the introduction of a specific example to explain the rules of association of several names:

[0066] 比如关联规则为"属性AA二〉属性BB"其中,定义所述历史点菜数据库中所有菜 [0066] association rules such as "AA two attribute> attribute BB" wherein, defining the ordering history database all dishes

品的属性数据的数据集为P,在数据集P中有Nl次出现所述属性AA的同时,又有N2次出现 Product attribute data for the data set P, P has the data set simultaneously Nl occurrences of the attribute AA, there appears N2 times

属性BB,则所述关联规则的置信度为N2/N1W00X,用于表示所述关联规则在置信度为N2/ Properties BB, if the confidence of the association rules N2 / N1W00X, for indicating the degree of confidence in the association rule N2 /

肌*100%的概率上是可信的,它是指特定个体对待特定命题真实性相信的程度。 Muscle * 100% probability is credible, it refers to a specific individual to treat certain degree of authenticity to believe the proposition. 即在N1次 That is N1 times

出现属性AA的记录中,有N2次同时出现了属性BB,则关联规则"属性AA二〉属性BB"在 Record appearance attribute AA, there also appeared N2 times property BB, the association rule "two AA Properties> Properties BB" in

N2/NW100X的概率上是可信的。 The probability of N2 / NW100X is credible.

[0067] 定义属性AA为关联规则前件,属性BB为关联规则后件。 [0067] AA is defined as the property rules associated with the front member, the association rules for the attribute member BB.

[0068] 其中,所述关联规则由关联规则前件和关联规则后件组成,所述关联规则后件由所述关联规则前件推导得出,定义所述历史点菜数据库中所有菜品的属性数据的数据集为P,在数据集P中有Nl次出现所述关联规则前件的同时,又有N2次出现所述关联规则后件,则所述关联规则的置信度为N2/N1W00X,用于表示所述关联规则在置信度为N2/ 肌*100%的概率上是可信的, [0068] wherein the association rule by the association rules front member and rear member composed of association rules, after the association rule member by the first member of the association rules deduced, ordering history database defining the attributes of all the dishes dataset is P, there occurs while the association rule Nl front views of the member, there are N2 times after the association rules in the data set member P, then the association rule confidence level N2 / N1W00X, is used to indicate the confidence in the association rule is the N2 / muscle * 100% probability is authentic,

[0069] 在关联规则挖掘的方法上,本发明采用了最新的关联规则挖掘算法FPgrowth。 [0069] In the method of mining association rules, the present invention employs the latest association rule mining algorithm FPgrowth. Collection

用频繁模式树的数据结构,在效率上比关联规则挖掘的传统算法Apriori更高。 With frequent pattern tree data structure, higher in efficiency than the association rule mining algorithm traditional Apriori.

[0070] 步骤S02、接收用户输入的第一菜品,并将所述输入的第一菜品存储至已点菜品数 [0070] step S02, the dishes receive a first user input, and the input of the first point has been stored to the number of dishes dishes

据库; Databases;

[0071] 步骤S03、从所述关联规则集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种属性数据或者一种以上组合的属性数据为关联规则前件,以所述菜品X的一种属性数据或者一种以上组合的属性数据为关联规则后件的属性关联规则集,并将获得的所述属性关联规则集中的每条关联规则的置信度相加计算得到所述菜品X的推荐价值; [0071] step S03, the association from the rule set, for all the dishes in the dish X, looking for matches the attribute data in the database has been dishes point dot data an attribute of dishes or in combination one or more after the attributes associated rule set of association rule antecedent attribute data of the dishes in an X or a combination of one or more data attributes for association rules member, and the associated attributes of each set of rules obtained in association rules confidence is calculated by adding the value of the recommended X of dishes;

[0072] 其中,X为所述所有菜品中的其中的一个菜品; [0072] wherein, X is a dish in which all the dishes in;

10[0073] 由此可以计算得到所述所有菜品中的每一个菜品的推荐价值。 10 [0073] can be calculated to obtain the recommended value of each of all the dishes dishes. 根据已点菜品的属性数据,通过生成的关联规则可以得出顾客已点的菜品具有什么样的属性,喜欢包括这种属性的菜品的顾客,还可能喜欢具有什么其他属性的菜品。 According to the attribute data points have dishes, by generating association rules can be drawn dishes Customers point of what properties have, like customers include dishes such attributes may also like what dishes have other attributes.

[0074] 步骤S04、将所述所有菜品中的每一个菜品的推荐价值进行排序,选取推荐价值排名靠前的N个菜品作为推荐菜品,并输出所述N个推荐菜品,其中,所述N为大于O的整数; 根据已点菜品,可以找出若干条以它们的属性数据为关联规则前件的关联规则,那么所有菜品中有的菜会与这其中的多条关联规则相关,每条关联规则都有一定的置信度,显然相关的置信度之和越高,推荐价值越高。 [0074] step S04, the recommended value of each of the dish in all the dishes to be sorted, selected recommendation ranking value as the N Recommended dishes dishes, and outputs the N Recommended dishes, wherein the N is an integer greater than O; the dishes have points, can identify several pieces of attribute data in their association rule antecedent association rules, then all dishes some of the cabbage association rules associated with a plurality of which, each association rule has a certain degree of confidence, confidence is clearly related to the sum of the higher, the higher the recommended value.

[0075] 另,在此同时保证主成分属性相同的菜品不超过2个,可以利用选择函数if条件判别函数,进行定义函数选取。 [0075] Also, at the same time ensure the same main component attribute dishes no more than 2, may be utilized if condition selection function discriminant function defined function selected.

[0076] 另,步骤D中的输出N个推荐菜品的数量可以根据用户的要求进行更改,可以选取N的数量为5,若没有生成5个符合条件且推荐价值大于0的推荐菜品,则可以选取菜品推荐价值排序表中排名最高的菜品或者也可以是选取餐厅特别推荐的其它菜品,凑足5个菜品作为推荐菜品输出。 [0076] Also, the number of N Recommended dishes in Step D output can be changed according to the user's requirements, may be selected number N is 5, if not generated 5 Matched and recommended value greater than the recommended dishes 0, it can be recommended dishes worth ordering select the highest ranked table dishes or may also be selected restaurant especially recommended other dishes, gather five dishes recommended dishes as output.

[0077] 步骤S05、判断是否接收到用户输入的菜品,若是,则将所述输入的菜品存储至所述已点菜品数据库并继续步骤C ;否则,结束推荐点菜。 [0077] Step S05, determines whether a user input is received dishes, the dishes if stored, then the point has been input to the database and the dishes continues with step C; Otherwise, the recommended order.

[0078] 另,步骤E中的用户输入的菜品可以是从推荐菜品中选择的菜品,也可以是用户自选的菜品,还可以是餐厅推荐的常点菜品等。 [0078] Also, step E dishes user input may be selected from the recommended dishes in dishes, may be user selectable dishes, may also often points Dining dishes and the like.

[0079] 具体的,所述所有菜品的属性数据包括以下属性类别:主成分属性和特征属性。 [0079] Specifically, the attribute data for all the dishes attribute categories comprises: a main component attributes and characteristics of attributes. The

述特征属性包括以下属性子类别:口味、烹饪方法、辅料、调料、菜系、产地等。 Wherein said attribute includes the following attributes subcategories: taste, cooking methods, materials, spices, cuisine, etc. origin. 主成分属性 Principal Component Properties

为具体的菜品包含的主要成分,比如:黑椒牛柳炒意粉,意粉为主成分,属性数据的选取可 Concrete dishes containing a main component, such as: Black Pepper Beef fried pasta, pasta as a main component, the attribute data may be selected

根据实际情况选取以上列举的部分或者全部数据,还可以为其他属性子类别。 Enumerated above to select some or all of the actual situation data may also be other attributes subcategories. 具体的烹饪 Specific cooking

方法的属性子类别可以用炒、煎、爆、炸、烧、煮、蒸、炖/煨/焖、熏、烘/烤、白灼等指标来衡 Properties sub-category method can be fried, fried, explosion, fried, roasting, boiling, steaming, stewing / simmer / stew, smoked, baked / grilled, boiled and other indicators to scale

量;调料的属性子类别可以用酱、醋、辣椒油、辣椒、葱姜、麻油、味精等指标来衡量;口味的 Amount; spices property sub-categories can be used sauce, vinegar, chili oil, chili, ginger, sesame oil, monosodium glutamate and other indicators to measure; taste

属性子类别可以用辣味、麻辣、酸甜等指标来衡量凍系为鲁、川、苏、粤、闽、浙、湘、徽等菜 Properties sub-categories can be spicy, spicy, sweet and sour and other indicators to measure the freeze line as Shandong, Sichuan, Jiangsu, Guangdong, Fujian, Zhejiang, Hunan, emblem and other vegetables

系,产地为中国、外国等。 Department of Chinese origin, foreign and so on.

[0080] 所述步骤S01具体包括: [0080] The step S01 comprises:

[0081] 根据获取的历史点菜数据库中所有菜品的主成分属性数据,对所述所有菜品的主成分属性数据之间进行数据挖掘生成主成分关联规则集,所述主成分关联规则集由多条主成分关联规则组成,并计算所述主成分关联规则集中每条主成分关联规则的置信度; [0082] 根据获取的历史点菜数据库中所有菜品的特征属性数据,对所述所有菜品的特征属性数据之间进行数据挖掘生成特征关联规则集,所述特征关联规则集由多条特征关联规则组成,并计算所述特征关联规则集中每条特征关联规则的置信度; [0081] The ordering history database acquired attribute data for all the main component of the dishes, a main component to generate data mining association rules between sets attribute data of the main component for all the dishes, the main component by a multi-association rule set Article of rules associated with the main component, and calculating the principal components associated with each ruleset main component confidence association rules; [0082] the attribute data acquired in the history database ordering all features dishes, the dishes all generating feature data mining association rules between sets feature attribute data, wherein the association rule characterized by a plurality of sets of rules is associated, and calculates the feature association rule set for each feature confidence association rules;

[0083] 简单地把菜品的名称作为数据项,并不能挖掘到有用的规则,所以采用提取菜品的内在特征属性以及主成分属性,分别把菜品的内在特征属性以及主成分属性作为数据项进行一次关联规则挖掘,从而得到菜品的内在特征属性之间以及主成分之间的内在关系。 [0083] simply the name of the dishes as a data item, and does not dig into useful rules, so the use of the intrinsic characteristic properties and a main component attribute extraction dishes, respectively, the intrinsic characteristic properties of the dishes, and a main component attributes as data items a association rule mining to obtain the intrinsic relationship between the intrinsic characteristic properties of the dishes, and a main component. [0084] 所述步骤S03具体包括: [0084] The step S03 comprises:

[0085] Cl、对所述主成分属性和所述特征属性分别设立主成分属性权值函数和特征属性权值函数;[0086] C2、从所述主成分关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则前件,以所述菜品X的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则后件的主成分属性关联规则集,并将获得的所述主成分属性关联规则集中的每条主成分关联规则的置信度与主成分权值相乘,即得到所述所有菜品中的菜品X的主成分属性推荐置信度,其中,所述主成分权值是根据所述主成分属性权值函数计算得到; [0087] 另,所述主成分权值还可以是预先设置; [0085] Cl, the main component of the characteristic attributes and attribute set up a main component attribute weight function and characteristic attribute weight function; [0086] C2, from the main component correlation set, for all the dishes in X dishes, looking for a match to said main component has the attribute data point database dishes have a master-point component attribute data dishes or one or more combinations of antecedent rules for association to said one main dishes X confidence attribute data with the main component or a main component composition attribute data of one or more attributes associated rule sets of the main component member association rules, the rules associated with the main component attribute and the obtained main component of each set of association rules multiplied by the weight component, i.e., all the attributes of the main component obtained in the dishes dishes confidence recommendation X, wherein the main component is calculated according to the weight of the main component attribute weight function; [0087] another, the main component weights may also be set in advance;

[0088] C3、从所述特征关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种特征属性数据或者一种以上组合的特征属性数据为关联规则前件,以所述菜品X的一种特征属性数据或者一种以上组合的特征属性数据为关联规则后件的特征属性关联规则集,并将获得的所述特征属性关联规则集中的每条特征关联规则的置信度与特征权值相乘,即得到所述所有菜品中的菜品X的特征属性推荐置信度,其中, 所述特征权值是根据所述特征属性权值函数计算得到; [0089] 另,所述特征权值还可以是预先设置; [0088] C3, from the feature correlation set, for all the dishes in the dish X, seeking to match the characteristic properties have been in the database point dishes dishes a characteristic point data or one or more attribute combinations data for the association rule antecedent to the features of the dishes or the attribute data X wherein attribute data is characterized by one or more combinations of the association rule element attributes associated rule set, the rule association property characteristic and the obtained concentrate and the confidence values ​​for each characteristic feature weights association rule is multiplied to obtain characteristic properties of the recommended confidence in all the dishes dishes X, wherein said characteristic weights are based on the weight function characteristic attribute get; [0089] also, the feature weights may also be set in advance;

[0090] C4、将所述菜品X的主成分属性推荐置信度与特征属性推荐置信度相加,即得到所述菜品X的推荐价值。 [0090] C4, the main component of the attribute X dishes recommended recommended confidence confidence characteristic attribute added to obtain the recommended value of X dishes.

[0091] 所述步骤S03中的Cl中的主成分属性权值函数和特征属性权值函数均为递增函数。 [0091] The step S03 of the main component Cl in the attribute weight function and the weight function are characteristic attribute increasing function.

[0092] 所述步骤S03中的步骤Cl中的主成分属性权值函数为Y二Fl(tl) 二tl衬l,其中, 所述主成分属性权值函数的自变量tl的计算方法为,针对每条主成分关联规则,在关联规则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值; [0092] The main weight function component attribute Cl step step S03 is Y in the two Fl (tl) tl two lining L, wherein, calculated from the variables tl main component attribute is weight function, association rules for each piece of the main component, the main component of the rule association property of the antecedent point number of dishes have been appearing in the main component and the number of dishes and the attribute appearing in the point is the ratio of ;

[0093] 所述特征属性权值函数Y二F2(t2) = t2衬2,其中,所述特征属性权值函数的自变 [0093] The characteristic attribute weight function Y = F2 (t2) = t2 liner 2, wherein the characteristic attribute of the independent variable weight function

量t2的计算方法为,针对每条特征关联规则,在关联规则前件中的特征属性在所述已点菜 T2 is an amount calculated for each feature association rule, wherein the rule association property antecedent has the ordering

品中出现的序号之和与所述特征属性在所述已点菜品中出现的次数之和的比值。 And the number of characteristic attributes of the article appearing in the number of dishes and the ratio has occurred points.

[0094] 另,所述主成分属性权值函数和所述特征属性权值函数为相同的权值函数Y = [0094] Also, the main component attribute weight function and the weight function is a characteristic property of the same weight function Y =

F(t) =t*t。 F (t) = t * t. 设立主成分属性权值函数和特征属性权值函数是为了标识某条关联规则在推 Establish a main component attribute weight function and the weight function is a characteristic attribute for identifying association rules in a push bar

荐菜品过程中的重要程度,而选取主成分属性权值函数与特征属性权值函数相同,是考虑 Recommended dishes during the importance, the value of the function is selected wherein the same attribute weight function attribute weight main component, is considered

到主成分属性与特征属性的重要性相当。 The main component of the importance attributes and characteristics equivalent properties.

[0095] 另,引进一个具体的例子对主成分属性权值函数中的自变量tl的计算方法进行进一步解释: [0095] Also, the introduction of a specific example for further explanation of the calculation of principal components tl argument attribute weight function is:

[0096] 某条主成分关联规则:(主成分属性A1、B1、C1)=> (主成分属性D1) [0096] Article of a main component of the main component attribute association rules :( A1, B1, C1) => (main component attribute D1)

[0097] 在已点菜品{a、b、c〜}中,若主成分属性Al在已点菜品中第a个菜品出现,类似 [0097] In point has dishes {a, b, c~}, when the main component attribute Al a-th point dishes appear in the dishes, like

的,主成分属性Bl在已点菜品中第b个菜品出现,主成分属性Cl在已点菜中第c个菜品出 Principal Component b th Bl properties appear in the dishes point dishes, a main component attribute Cl c-th dishes have been shown in the ordering

现,则主成分属性权值函数中的自变量tl = (a+b+c)/3。 Now, the main component attribute weight function arguments tl = (a + b + c) / 3.

[0098] 所述步骤S03中的步骤Cl中的主成分属性权值函数为Y = Fl (tl) = tl*tl*tl, 其中,所述主成分属性权值函数的自变量tl的计算方法为,针对每条主成分关联规则,在关联规则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值;[0099] 所述特征属性权值函数Y二F2(t2) = t2衬2,其中,所述特征属性权值函数的自变 [0098] The main component attribute weight function step in the step S03 Cl is Y = Fl (tl) = tl * tl * tl, where tl is calculated from the variable component of said primary attribute weight function is, association rule for each principal component, a main component of the rule association property of the antecedent point number of dishes have been appearing in the main component attribute and the number of the point has occurred and dishes ratio; [0099] the characteristic attribute weight function Y = F2 (t2) = t2 liner 2, wherein said independent variable weight function characteristic attribute

量t2的计算方法为,针对每条特征关联规则,在关联规则前件中的特征属性在所述已点菜 T2 is an amount calculated for each feature association rule, wherein the rule association property antecedent has the ordering

品中出现的序号之和与所述特征属性在所述已点菜品中出现的次数之和的比值。 And the number of characteristic attributes of the article appearing in the number of dishes and the ratio has occurred points.

[0100] 另,设立主成分属性权值函数和特征属性权值函数是为了标识某条关联规则在推 [0100] Also, the establishment of a main component attribute weight function and the weight function is a characteristic attribute for identifying association rules in a push bar

荐菜品过程中的重要程度,而选取主成分属性权值函数与特征属性权值函数不相同,是考 Recommended dishes importance of the process, and the value of the function select component main weight function characteristic attribute property rights are not the same, a test

虑到主成分属性与特征属性的重要性不相当,当认为主成分属性的重要性大于特征属性的 Into account the importance of a main component attribute does not correspond to the characteristic property, when the main component of the importance of that attribute is greater than the characteristic attribute

重要性时,主成分属性权值函数Fl (t),以及特征属性权值函数F2 (t),并同时使得这两个 When the importance of, the main component attribute weight function Fl (t), and the feature attribute weight function F2 (t), and at the same time so that the two

函数满足如下要求: Function satisfies the following requirements:

[0101] 对于任意的t, Fl(t) > F2(t); [0101] for any t, Fl (t)> F2 (t);

[0102] 当认为主成分属性的重要性小于特征属性的重要性时,主成分属性权值函数Fl (t),以及特征属性权值函数F2 (t),并同时使得这两个函数满足如下要求: [0103] 对于任意的t, Fl (t) < F2 (t); [0102] When the main component of the importance of that attribute is less than the importance of the characteristic properties of the main component attribute weight function Fl (t), and the feature attribute weight function F2 (t), and such that the two functions simultaneously satisfy the following requirements: [0103] for any t, Fl (t) <F2 (t);

[0104] 另,引进一个具体的例子对特征属性权值函数中的自变量的计算方法进行进一步解释: [0104] Also, the introduction of a specific example of the calculation method of the weight function characteristic attribute argument will be further explained:

[010S] 某条特征关联规则:(特征属性E1、F1、G1) => {特征属性111} [010S] wherein a bar association rules :( characteristic properties E1, F1, G1) => {111} characteristic attribute

[0106] 在已点菜品{e、f、g〜}中,若特征属性El在已点菜品中第e个菜品出现,类似的, [0106] In point has dishes {e, f, g~}, if the e-th characteristic properties El dishes dishes appear in the points, like,

特征属性Fl在已点菜品中第f个菜品出现,特征属性Gl在已点菜中第g个菜品出现,则特 Fl characteristic properties of the f point dishes appear in the dishes, the characteristic property of the g-th Gl dishes appear in the ordering, then Laid

征属性权值函数中的自变量t2 = (e+f+g)/3。 Syndrome attribute weight function arguments t2 = (e + f + g) / 3.

[0107] 另,以上对于主成分属性权值函数中的自变量的计算方法,可以是简单的针对每条主成分关联规则,在关联规则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值这种简单的计算,还可以是对已点菜品出现的序号分别赋菜品权值,采用在关联规则前件中的主成分属性在所述已点菜品中出现的序号与之相对应的菜品权值的乘积之和再与所述主成分属性在所述已点菜品中出现的次数之和的比值这种加权求和再求平均的算法,还可以是预先设定的一个递增数列的固定值,还可以是其他的计算方法。 [0107] Also, the above method for calculating the main component of the attribute weight argument function, may be a simple association rules for each principal component, a main component of the rule association property of said front member has points of dishes and a number of the main component attribute appearing in this simple calculation of the number and the ratio of the dishes has occurred point, it may also have a number of points appearing dishes were assigned weights dishes, using the associated main component attribute rule in the antecedent was the point number appears in dishes with the corresponding product of the weights and then the dishes with the main component of the frequency points have attribute dishes and appearing in the this ratio is then weighted summation averaging algorithm may also be a preset fixed incremental value of the number of columns may be other calculation methods. 对于特征属性权值函数中的自变量的计算方法也是同理的。 Method for calculating the weight function characteristic attribute of the argument is the same reason.

[0108] 另,该递增函数可以是线性递增函数、指数型递增函数等。 [0108] Also, the increasing function may be a linear increasing function, exponential function or the like is incremented.

[0109] 另,可以根据实际情况定义权值函数。 [0109] Also, the value of the function may be defined according to the actual weight. 但是在一般情况下,这个函数是严格单调递增的。 But in general, this function is strictly monotonically increasing.

[0110] 点菜举例: [0110] Example la carte:

[0111] 1、点第一盘菜 [0111] 1, the first point dish

[0112] 此次点菜:黑椒牛柳炒意粉 [0112] The a la carte: Black Pepper Beef Fried Spaghetti

[0113] 已点菜品数据库:黑椒牛柳炒意粉 [0113] has carte menu database: Black Pepper Beef Fried Spaghetti

[0114] 已点菜品黑椒牛柳炒意粉的属性数据为"黑椒"、"牛柳"、"意粉",其中特征属性数据为"黑椒"、"牛柳",主成分属性数据为"意粉" [0114] Black Pepper Beef fried dishes point has spaghetti attribute data is "black pepper", "beef", "spaghetti", which feature attribute data is "black pepper", "beef", the main component attribute data for the "spaghetti"

[0115] 以菜品香炒鱿鱼丝为例,计算菜品香炒鱿鱼丝的推荐价值,其中菜品香炒鱿鱼丝的属性数据为"香"、"鱿鱼丝","香"为香炒鱿鱼丝的特征属性,"鱿鱼丝"为香炒鱿鱼丝的主成分属性。 [0115] dishes to Hong fired wire, for example, to calculate the value of the recommended dishes silk incense fired, fired incense dishes which attribute data wire is "sweet", "squid", "sweet" as incense fired wire characteristic properties, " squid "as the main ingredient of incense fired wire properties.

[0116] 以已点菜品黑椒牛柳炒意粉中的特征属性"黑椒"、"牛柳"为关联规则前件、以菜 [0116] In point has characteristic properties dishes Black Pepper Beef fried pasta of the "black pepper", "beef" is the association rule antecedent to food

13品香炒鱿鱼丝中的特征属性"香"为关联规则后件的特征属性关联规则集中的特征关联规则有: Hong characteristic properties fired product wire 13 "sweet" is characterized by the association rule element attributes associated rule set of feature association rules are:

[0117] {黑椒、牛柳} => {香}置信度:0. 5 [0117] {pepper, beef} => {} incense confidence: 05

[0118] 该条特征关联规则的权值计算过程如下:在关联规则前件中,有两个数据项,均为特征属性,分别为:"黑椒"、"牛柳"。 [0118] wherein the article weights association rules is calculated as follows: In association rule antecedent, there are two data items are characteristic property, namely: "black pepper", "beef." 其中,特征属性"黑椒"在所有已点菜品中的第l个已点菜品中出现,特征属性"牛柳"在所有已点菜品中的第1个已点菜品中出现,因此,调用特征属性权值函数F1(X) =1朽(可取其它函数作为权值函数)。 Wherein the feature attribute "Pepper" dishes at all points in the l th points has occurred dishes, characterized Properties "beef" appears at all points in the first dishes point dishes has been, therefore, the call feature attribute weight function F1 (X) = 1 rot (preferably other as a function of weight function). 计算其自变量^= (1 + D/2 Calculated argument ^ = (1 + D / 2

[0119] 将特征属性关联规则{黑椒、牛柳} => {香}的置信度0.5,乘以其权值F2(1) =1*1 = l,得到了该条特征关联规则的推荐置信度为0. 5*1 = 0. 5。 [0119] The characteristic attribute association rule {pepper, beef} => {} confidence fragrance 0.5, times its weight of F2 (1) = 1 * 1 = l, wherein the article has been recommended association rules a confidence level of 0.5 * 1 = 0.5.

[0120] 以已点菜品黑椒牛柳炒意粉中的主成分属性"意粉"为关联规则前件、以香炒鱿鱼丝中的主成分属性"鱿鱼丝"为关联规则后件的主成分关联规则集中的主成分关联规则有: A main component [0120] Black Pepper Beef dishes point to have a main component attribute "spaghetti" means a fried flour association rule antecedent, the fragrant property of the main component in the fired filaments "squid" Rules for the association member rule set associated with the main component of the association rules are:

[0121] {意粉} => {鱿鱼丝}置信度:0.48 [0121]} = {Spaghetti> {} squid confidence: 0.48

[0122] 该条主成分关联规则的权值计算过程如下:在关联规则前件中,有一个数据项,均为主成分属性,为:意粉。 [0122] piece weight value calculation process main component association rules as follows: the association rule antecedent, there is a data item, both the main component attribute as: pasta. 其中,主成分属性"意粉"在所有已点菜中的第1个已点菜中出现因此,调用主成分属性权值函数F1(X) =1朽(可取其它函数作为权值函数)。 Wherein a main component attribute "spaghetti" appears therefore, call a main component attribute weight function F1 (X) = 1 rot (preferably other as a function of weight function) has been the first in the ordering of all of the ordering. 计算其自变 Calculated independent variable

[0123] 将主成分关联规则{意粉} => {鱿鱼丝}的置信度0.48,乘以其权值F1(1)= 1*1 = l,得到了该条主成分关联规则的推荐置信度0. 48*1 = 0. 48。 [0123] The main component of spaghetti association rule {} => {} confidence squid 0.48, multiplied by the weight F1 (1) = 1 * 1 = l, to give the article the main component association rules recommended Confidence of 0.48 * 1 = 0.48.

[0124] 将上述属性关联规则的推荐置信度相加计算,便得到了菜品香炒鱿鱼丝的推荐价值为0. 5+0. 48 = 0. 98。 [0124] The recommended confidence the attribute calculated by adding the association rule, they will have a recommended value of the wire is fired dishes fragrance 0.5 + 0.48 = 0.98.

[0125] 利用同样的方法,可以计算出所有菜品中的其它菜品的推荐价值。 [0125] Using the same method, the recommended value may be calculated for all the other dishes dishes. 并将所有菜品按照其推荐价值降序排列,得到如下推荐菜。 And all the dishes in descending order according to their recommended values, obtain the following recommended dishes.

[0126] 点后推荐:香炒鱿鱼丝、香芋、香梅扣肉、重庆香椒牛肉、铁板香辣牛柳王 After the [0126] point recommendation: incense fired silk, taro, pork Xiangmei, Chongqing-pepper beef, spicy beef sizzling king

[0127] 2、点第二盘菜 [0127] 2, a second point dish

[0128] 此次点菜:重庆香椒牛肉 [0128] The a la carte: Chongqing-pepper beef

[0129] 已点菜品数据库:黑椒牛柳炒意粉、重庆香辣牛肉 [0129] has carte menu database: Black Pepper Beef Fried Spaghetti, Chongqing spicy beef

[0130] 已点菜品为黑椒牛柳炒意粉和重庆香辣牛肉,黑椒牛柳炒意粉的属性数据为"黑椒"、"牛柳"、"意粉",其中主成分属性数据为"意粉",特征属性数据为"黑椒"、"牛柳";重庆 [0130] has carte menu as Black Pepper Beef Fried Spaghetti and Chongqing spicy beef, attribute data Black Pepper Beef Fried Spaghetti is "black pepper", "beef", "spaghetti" in which the main components of property data of "spaghetti", wherein attribute data is "black pepper", "beef"; Chongqing

香辣牛肉的属性数据为"重庆"、"香"、"辣"、"牛肉",其中主成分属性数据为"牛肉",特征属性数据为"重庆"、"香"、"辣"。 Spicy beef attribute data is "Chongqing", "sweet", "spicy", "beef", the main component of which the attribute data is "beef", wherein attribute data is "Chongqing", "sweet", "hot."

[0131] 以香辣牛肉比萨为例,计算菜品香辣牛肉比萨的推荐价值,其中菜品香辣牛肉比萨的属性数据为"香"、"辣"、"比萨",其中主成分属性数据为"比萨",特征属性数据为"香"、 "辣"。 [0131] pizza with spicy beef, for example, calculates a recommended value of spicy beef dish pizza, where the attribute data spicy beef dish pizza is "sweet", "spicy", "pizza", which is the main component attribute data " pizza ", wherein attribute data is" sweet "," hot. "

[0132] 以已点菜品黑椒牛柳炒意粉、重庆香椒牛肉中的主成分属性数据"意粉"、"牛肉" 为关联规则前件、以菜品香辣牛肉比萨中的主成分属性数据"比萨"为关联规则后件的主成分关联规则集中的主成分关联规则有: [0133] {意粉、牛肉} => {比萨}置信度:0. 125 [0132] already point to Black Pepper Beef Fried pasta dishes, main component attribute data Chongqing-pepper beef "spaghetti", "beef" for the association rule antecedent to spicy beef dish pizza is the main component of property data "pizza" is a main component of the association rule set associated with the rule of the main component member association rules are: [0133] {pasta, beef} => {} Pisa the confidence: 0125

14[0134] 该条主成分关联规则的权值计算过程如下:关联规则前件中的主成分属性数据"意粉"在所有已点菜品中的第l个已点菜品(黑椒牛柳炒意粉)中出现,主成分属性数据"牛肉"在所有已点菜品中的第2个已点菜品(重庆香椒牛肉)中出现,因此,调用主成分属性权值函数Fl(x) = x朽(可取其它函数作为权值函数)。 14 [0134] weight value calculation process of association rules piece main component was as follows: component attribute data rule antecedent of "spaghetti" dishes at all points in the l has been points dishes (fried Black Pepper Beef Spaghetti) appear, the main component of attribute data "beef" appears in all points of dishes in the first two have been carte menu (Chongqing-pepper beef), and therefore, call the main component attribute weight function Fl (x) = x mortal (preferably other as a function of weight function). 计算其自变量,<formula>formula see original document page 15</formula> Calculation of its argument, <formula> formula see original document page 15 </ formula>

[0135] 将主成份关联规则{意粉、牛肉} => {比萨}的置信度O. 125,乘以其权值<formula>formula see original document page 15</formula>得到了该条主成分关联规则的推荐置信度0. 125*2. 25 = 0. 28125 [0135] The main component association rule {pasta, beef} => {} Pisa confidence O. 125, times its weight of <formula> formula see original document page 15 </ formula> article obtained main component association rules recommended confidence 0.125 * 2.25 = 0.28125

[0136] 以已点菜品黑椒牛柳炒意粉、重庆香椒牛肉中的特征属性数据"黑椒"、"牛柳"、 [0136] already point to Black Pepper Beef Fried pasta dishes, feature attribute data Chongqing-pepper beef "black pepper", "beef"

"重庆"、"香"、"辣"为关联规则前件,以菜品香辣牛肉比萨中的特征属性数据"香"、"辣"为 "Chongqing", "sweet", "spicy" for the association rule antecedent to spicy beef dish pizza in the feature attribute data "sweet", "spicy" as

关联规则后件的特征关联规则集有: Wherein the association rule set member association rules are:

[0137] ①{黑椒、牛柳} => {香}置信度:0.5 [0137] ① {pepper, beef} => {} Hong Confidence level: 0.5

[0138] ②{重庆,黑椒} => {辣}置信度:0.25 [0138] ② {Chongqing, black pepper} => {} hot confidence: 0.25

[0139] ①特征关联规则{黑椒、牛柳} => {香}的权值计算过程如下:关联规则前件中特征属性"黑椒"在所有已点菜品中的第l个已点菜品(黑椒牛柳炒意粉)中出现,特征属性"牛柳"在所有已点菜品中的第l个已点菜品(黑椒牛柳炒意粉)中出现,因此,调用特征属性权值函数F2(x) =1朽(可取其它函数作为权值函数)。 [0139] ① characterized in association rule {pepper, beef} => {} fragrant weights calculated as follows: wherein the association rule antecedent attribute "Pepper" dishes at all points in the first point has been dishes l (Black pepper beef fried Spaghetti) appear, feature attribute "beef" appears in all the dishes in the l point has been carte menu (Black pepper beef fried Spaghetti), therefore, the call feature attribute weight function F2 (x) = 1 rot (preferably other as a function of weight function). 计算其自变量^= (1+D/2=1。 Calculated argument ^ = (1 + D / 2 = 1.

[0"0] 将特征关联规则{黑椒、牛柳} => {香}的置信度0.5,乘以其权值F1(1) = 1*1=l,得到了该条特征关联规则的推荐置信度0. 5*1 = 0. 5. [0 "0] wherein the association rule {pepper, beef} => {} confidence fragrance 0.5, multiplied by the weight F1 (1) = 1 * 1 = l, wherein the article obtained association rules recommended confidence 0.5 * 1 = 0.5.

[0141] ②特征关联规则{重庆,黑椒} => {辣}的权值计算过程如下:关联规则前件中特征属性"重庆"在所有已点菜品中的第2个已点菜品(重庆香辣牛肉)中出现,特征属性"黑椒"在所有已点菜品中的第l个已点菜品(黑椒牛柳炒意粉)中出现,因此,调用特征属性权值函数F2(x) =1朽(可取其它函数作为权值函数)。 [0141] ② characterized in association rule {Chongqing, black pepper} => {} spicy weights calculated as follows: wherein the association rule antecedent attribute "Chongqing" dishes at all points in the first two points have been dishes (Chongqing spicy beef) appear, feature attribute "pepper" appears in all the dishes in the l point has been carte menu (Black pepper beef fried Spaghetti), therefore, the call feature attribute weight function F2 (x) rot = 1 (preferably other as a function of weight function). 计算其自变量,1= (2+1)/2 =1. 5。 Calculation of its argument, 1 = (2 + 1) / 2 = 1.5.

[0"2] 将特征属性关联规则{重庆,黑椒} => {辣}的置信度0.25,乘以其权值F2(1)=1. 5*1. 5 = 2. 25,得到了该条特征关联规则的推荐置信度0. 25*2. 25 = 0. 5625[0143] 将上述主成分属性关联规则以及特征属性推荐置信度相加计算,便得到了菜品香炒鱿鱼丝的推荐价值为0. 5625+0. 28125 = 0. 84375 [0 "2] wherein the association rule {Chongqing property, black pepper} => {} confidence spicy 0.25, multiplied by the weight F2 (1) = 1. 5 * 1. 5 = 2. 25, to give the recommended rules article characteristics associated confidence 0.25 * 2.25 = 0.5625 [0143] the properties of the main components and features of the property rules associated confidence calculated by adding the recommendation, it has been recommended value dishes fragrant fired filaments It was 0.5625 + 0.28125 = 0.84375

[0144] 利用同样的方法,可以计算出所有菜品中的其它菜品的推荐价值。 [0144] Using the same method, the recommended value may be calculated for all the other dishes dishes. 并将所有菜品按照其推荐价值降序排列,得到如下推荐菜。 And all the dishes in descending order according to their recommended values, obtain the following recommended dishes.

[0145] 点后推荐:香辣牛肉比萨、蒜香辣鸡尖、麦香辣鸡翅、五香辣味鸭舌、香辣花生米[0146] ……直至顾客点完所有的菜品,计算机判断是否接收到用户输入的菜品,若是则 After the [0145] point recommendation: spicy beef pizza, garlic spicy chicken sharp, spicy chicken wings wheat, five Hot & Spicy duck tongue, spicy peanuts [0146] ...... finished until the customer point of all the dishes, the computer determines whether to receive the dishes entered by the user, if the

将所述输入的菜品存储至所述已点菜品的数据库并返回以上步骤,否则结束推荐点菜。 The input to the storage dishes have carte menu of database and return the above steps, otherwise the end of the recommended order. [0147] 实施例二: [0147] Example II:

[0148] 对应地,本发明还提供了一种智能化推荐点菜系统如图2和图3所示,为本发明的一种结构示意图。 [0148] Correspondingly, the present invention further provides a schematic structural diagram of an intelligent recommendation ordering system shown in FIGS. 2 and 3, of the present invention.

[0149] 具体的,本发明的一种智能化推荐点菜系统,其包括: [0149] Specifically, an intelligent recommendation ordering system according to the present invention, which comprises:

[0150] 关联规则生成模块l,用于根据获取的历史点菜数据库中所有菜品的属性数据,对所述所有菜品的属性数据之间进行数据挖掘生成关联规则集,所述关联规则集由多条关联关联规则集中每条关联规则的置信度; [0150] L association rule generation module, based on the history database ordering attribute data acquired in all the dishes, the data attribute data between all the dishes to generate mining association rules set, the association rules set by a plurality of Article association rule set associated with each confidence association rules;

[0151] 其中,所述关联规则由关联规则前件和关联规则后件组成,所述关联规则后件由所述关联规则前件推导得出,定义所述历史点菜数据库中所有菜品的属性数据的数据集为P,在数据集P中有Nl次出现所述关联规则前件的同时,又有N2次出现所述关联规则后件,则所述关联规则的置信度为N2/N1W00X,用于表示所述关联规则在置信度为N2/肌*100%的概率上是可信的; [0151] wherein the association rule by the association rules front member and rear member composed of association rules, after the association rule member by the first member of the association rules deduced, ordering history database defining the attributes of all the dishes dataset is P, there occurs while the association rule Nl front views of the member, there are N2 times after the association rules in the data set member P, then the association rule confidence level N2 / N1W00X, It is used to indicate the confidence of the association rule to be authentic in the N2 / * muscle probability of 100%;

[0152] 第一菜品接收模块2,用于接收用户输入的第一菜品,并将所述输入的第一菜品存储至已点菜品的数据库; [0152] a first receiving module 2 dishes, the dishes for receiving a first user input, and the input of the first point has been stored to the dishes dishes database;

[0153] 推荐价值生成模块3,用于从所述关联规则集中,针对所述所有菜品中的菜品X, [0153] Recommended value generation module 3, the association rules from a centralized, all of the dishes for X in the dishes,

寻找匹配以所述已点菜品数据库中的已点菜品的一种属性数据或者一种以上组合的属性 Looking to the attribute matching point has been dishes database and an attribute data point dishes or one or more combinations

数据为关联规则前件,以所述菜品X的一种属性数据或者一种以上组合的属性数据为关联 Attribute data for the association rule antecedent attribute data to the one or more than one dish X is associated with a combination of

规则后件的属性关联规则集,并将获得的所述属性关联规则集中的每条关联规则的置信度 Confidence attributes associated rule set of the rule member and the associated attributes of each set of rules obtained in association rules

相加计算得到所述菜品X的推荐价值; Calculated by adding the value of the recommended X of dishes;

[0154] 其中,X为所述所有菜品中的其中的一个菜品; [0154] wherein, X is a dish in which all the dishes in;

[0155] 输出模块4,用于将所述所有菜品的推荐价值进行排序,选取推荐价值排名靠前的 [0155] Output module 4, the recommended value for all sorts of dishes, selected ranking value of the recommended

N个菜品作为推荐菜品,并输出所述N个推荐菜品,其中,所述N为大于0的整数; Recommended dishes dishes as the N, and outputs the N Recommended dishes, wherein said N is an integer greater than 0;

[0156] 判断模块5,用于判断是否接收到用户输入的菜品,若是,则将所述输入的菜品存 [0156] 5 determining module, for determining whether user input is received dishes, the dishes if stored, then the input

储至所述已点菜品的数据库,并向推荐价值生成模块发送寻找匹配的执行指令;否则,结束 Storage to the point of dishes have a database, send for a matching execution of instructions and recommend value generation module; otherwise, end

推荐点菜。 It recommended a la carte.

[0157] 更具体的,所述所有菜品的属性数据包括以下属性类别:主成分属性和特征属性。 [0157] More specifically, the attribute data includes the following attributes for all categories of dishes: a main component attributes and characteristics of attributes. 所述特征属性包括以下属性子类别:口味、烹饪方法、辅料、调料、菜系和产地。 The characteristic properties include the following attributes subcategories: taste, cooking methods, materials, seasoning, origin and cuisine. 主成分属性为具体的菜品包含的主要成分,比如:黑椒牛柳炒意粉,意粉为主成分,属性数据的选取可根据实际情况选取以上列举的部分或者全部数据,还可以为其他属性子类别。 Attribute main component as a main component contains specific dishes, such as: Black Pepper Beef fried pasta, pasta-based component, may select the attribute data to select part or all of the data listed above according to the actual situation, may be other properties Subcategory. 具体的烹饪方法的属性子类别可以用炒、煎、爆、炸、烧、煮、蒸、炖/煨/焖、熏、烘/烤、白灼等指标来衡量;调料的属性子类别可以用酱、醋、辣椒油、辣椒、葱姜、麻油、味精等指标来衡量;口味的属性子类别可以用辣味、麻辣、酸甜等指标来衡量凍系为鲁、川、苏、粤、闽、浙、湘、徽等菜系;产地为中国、外国等。 Properties sub-category specific cooking method can be fried, fried, explosion, fried, roasting, boiling, steaming, stewing / simmer / stew, smoked, baked / grilled, boiled and other indicators to measure; attribute subcategory seasoning can be used sauce, vinegar, chili oil, chili, ginger, sesame oil, monosodium glutamate and other indicators to measure; attribute can be used subcategory taste spicy, spicy, sweet and sour and other indicators to measure the freeze line as Shandong, Sichuan, Jiangsu, Guangdong, Fujian , Zhejiang, Hunan, emblem and other cuisine; the origin of Chinese, foreign and so on.

[0158] 进一步的,所述关联规则生成模块1包括: [0158] Further, the association rule generation module 1 comprises:

[0159] 主成分属性数据的关联规则生成单元ll,用于根据获取的历史点菜数据库中所有菜品的主成分属性数据,对所述所有菜品的主成分属性数据之间进行数据挖掘生成主成分关联规则集,所述主成分关联规则集由多条主成分关联规则组成,并计算所述主成分关联规则集中每条主成分关联规则的置信度; [0159] main component attribute data association rule generation unit ll, ordering history database according to the acquired attribute data for all the main components of the dishes, a main component to generate data mining attribute data between the main component for all the dishes association rule set, the rule set associated with the main component of a plurality of rules associated with the main component composition, and calculating the principal components associated with each ruleset main component confidence association rules;

[0160] 特征属性数据的关联规则生成单元12,用于根据获取的历史点菜数据库中所有菜品的特征属性数据,对所述所有菜品的特征属性数据之间进行数据挖掘生成特征关联规则集,所述特征关联规则集由多条特征关联规则组成,并计算所述特征关联规则集中每条特征关联规则的置信度; [0160] association rules feature attribute data generation unit 12, based on the history database ordering attribute data acquired in the characteristic of all the dishes, attribute data between characteristic data of the all the dishes mining association rules generate a feature set, wherein the association rule characterized by a plurality of sets of rules is associated, and calculates the feature association rule set for each feature confidence association rules;

[0161 ] 所述所述推荐价值生成模块3包括: [0161] The value of the recommendation generation module 3 comprising:

[0162] 权值分配单元31,用于对主成分属性和特征属性分别设立主成分属性权值函数和特征属性权值函数; [0162] the weight assignment unit 31, a main component for the attributes and characteristics of attributes set up a main component attribute weight function and the weight function characteristic attribute;

16[0163] 主成分属性推荐置信度生成单元32,用于从所述主成分关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则前件,以所述菜品X的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则后件的主成分属性关联规则集,并将获得的所述主成分属性关联规则集中的每条主成分关联规则的置信度与主成分权值相乘,即得到所述所有菜品中的菜品X的主成分属性推荐置信度,其中,所述主成分权值是根据所述主成分属性权值函数计算得到; 16 [0163] Recommended confidence main component attribute generation unit 32, a main component from the associated set, for all the dishes in the dish X, has been to find a match to the point in the database has dishes point dishes principal component principal component attribute data attribute data or attribute data of one or more principal components combined correlation rule antecedent, the dishes in a main component attribute data X or one or more combinations of the main element of the association rules confidence that the main component by multiplying the weight of each main component of component attribute association rules association rules set, and to focus the primary component attribute association rules obtained, i.e., to obtain a main component attribute all the dishes in the dishes of X recommended confidence, wherein the main component is calculated according to the weight of the main component attribute weight function;

[0164] 特征属性推荐置信度生成单元33,用于从所述特征关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种特征属性数据或者一种以上组合的特征属性数据为关联规则前件,以所述菜品X的一种特征属性数据或者一种以上组合的特征属性数据为关联规则后件的特征属性关联规则集,并将获得的所述特征属性关联规则集中的每条特征关联规则的置信度与特征权值相乘,即得到所述所有菜品中的菜品X的特征属性推荐置信度,其中,所述特征权值是根据所述特征属性权值函数计算得到; [0165] 推荐价值生成单元34,用于将所述菜品X的主成分属性推荐置信度与特征属性推荐置信度相加,即得到所述菜品X的推荐价值。 [0164] wherein the confidence recommended property generating unit 33, a feature from the correlation set, for all the dishes in the dish X, to find a match to the characteristic properties have been in the database point dishes point dishes characteristic feature attribute data or a combination of one or more data attributes for association rule antecedent to a characteristic of the attribute data X dishes or one or more combinations of features of the association rule member attributes associated rule set, and and wherein the confidence weights set for each feature of the feature attribute association rules obtained by multiplying the association rule, to obtain the characteristic properties of all the dishes have recommended confidence in dishes of X, wherein said characteristic weights are calculated according to the weight function characteristic attribute; [0165] recommended value generation unit 34, a main component for the attributes of the recommended X dishes confidence confidence characteristic attribute recommended adding X to obtain the dishes recommended value.

[0166] 本发明可以应用在菜品的智能推荐点菜、商品的智能推荐购买等各种与本发明的方法相关的领域上。 [0166] The present invention can be applied in the intelligent intelligent recommendation la carte dishes, and other items related to the purchase recommendation process of the present invention on the field.

[0167] 最后应当说明的是,以上实施例仅用以说明本发明的技术方案,而非对本发明保护范围的限制,尽管参照较佳实施例对本发明作了详细地说明,本领域的普通技术人员应当理解,可以对本发明的技术方案进行修改或者等同替换,而不脱离本发明技术方案的实质和范围。 [0167] Finally, it should be noted that the above embodiments are intended to illustrate the present invention, not to limit the scope of the present invention, although described with reference to the preferred embodiments of the present invention in detail, of ordinary skill in the art It will appreciate that modifications may be made to the technical solutions of the present invention, or equivalent replacements without departing from the spirit and scope of the technical solutions of the present invention.

Claims (9)

  1. 一种智能化推荐点菜方法,其特征在于,包括以下步骤:A、根据获取的历史点菜数据库中所有菜品的属性数据,对所述所有菜品的属性数据之间进行数据挖掘生成关联规则集,所述关联规则集由多条关联规则组成,并计算所述关联规则集中每条关联规则的置信度;其中,所述关联规则由关联规则前件和关联规则后件组成,所述关联规则后件由所述关联规则前件推导得出,定义所述历史点菜数据库中所有菜品的属性数据的数据集为P,在数据集P中有N1次出现所述关联规则前件的同时,又有N2次出现所述关联规则后件,则所述关联规则的置信度为N2/N1*100%,用于表示所述关联规则在置信度为N2/N1*100%的概率上是可信的;B、接收用户输入的第一菜品,并将所述输入的第一菜品存储至已点菜品数据库;C、从所述关联规则集中,针对所述所有菜品中的菜品X, An intelligent recommendation ordering method comprising the steps of: A, ordering the database according to the historical data of all the acquired attribute dishes, the attribute data between all the dishes to generate data mining association rules set , the association rule set consisting of a plurality of association rules, and calculating the association rules in the rule set associated with each confidence; wherein the association rule association rules by the rear member and the front member association rules composition, the association rule said rear member by the first member deduced association rule, defining the dataset history database ordering attribute data for all the dishes P, while there appears before the association rule element N1 times in the data set P, N2 times have occurred after the association rule member, then the association rule confidence level N2 / N1 * 100%, for indicating the association rule is a degree of confidence in the probability N2 / N1 * 100% of channel; a first dishes B, receiving a user input, and the input of the first point has been stored to the dishes dishes database; C, from the associated rule set, all of the dishes for X in the dishes, 寻找匹配以所述已点菜品数据库中的已点菜品的一种属性数据或者一种以上组合的属性数据为关联规则前件,以所述菜品X的一种属性数据或者一种以上组合的属性数据为关联规则后件的属性关联规则集,并将获得的所述属性关联规则集中的每条关联规则的置信度相加计算得到所述菜品X的推荐价值;其中,X为所述所有菜品中的其中的一个菜品;D、将所述所有菜品中的每一个菜品的推荐价值进行排序,选取推荐价值排名靠前的N个菜品作为推荐菜品,并输出所述N个推荐菜品,其中,所述N为大于0的整数;E、判断是否接收到用户输入的菜品,若是,则将所述输入的菜品存储至所述已点菜品数据库并继续步骤C;否则,结束推荐点菜。 The point is to find a match in the database dishes point has an attribute data of dishes or in combination of more than one attribute data is associated antecedent rules, attribute data of the dishes in an X or a combination of one or more of the properties after the data attributes associated rule set of association rules member, and adding the confidence associated attributes of each set of rules obtained in association rules recommended value of the calculated X, dishes; wherein, X is all the dishes one of the dishes; D, recommended value of each of the dishes in all sorts of dishes, selected recommendation ranking value as the N recommended dishes dishes, and outputs the N recommended dishes, wherein N is an integer greater than 0; E, determining whether user input is received dishes, the dishes if stored, then the point has been input to the database and the dishes continues with step C; otherwise, the recommended order.
  2. 2. 根据权利要求1所述的智能化推荐点菜方法,其特征在于,所述所有菜品的属性数据包括以下属性类别:主成分属性和特征属性。 2. The method of claim intelligent recommendation ordering according to claim 1, wherein the attribute data includes the following attributes for all categories of dishes: a main component attributes and characteristics of attributes.
  3. 3. 根据权利要求2所述的智能化推荐点菜方法,其特征在于, 所述步骤A具体包括:根据获取的历史点菜数据库中所有菜品的主成分属性数据,对所述所有菜品的主成分属性数据之间进行数据挖掘生成主成分关联规则集,所述主成分关联规则集由多条主成分关联规则组成,并计算所述主成分关联规则集中每条主成分关联规则的置信度;根据获取的历史点菜数据库中所有菜品的特征属性数据,对所述所有菜品的特征属性数据之间进行数据挖掘生成特征关联规则集,所述特征关联规则集由多条特征关联规则组成,并计算所述特征关联规则集中每条特征关联规则的置信度;所述步骤C具体包括:Cl、对所述主成分属性和所述特征属性分别设立主成分属性权值函数和特征属性权值函数;C2、从所述主成分关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述 3. The method of claim intelligent recommendation ordering according to claim 2, wherein said step A comprises: according to the historical database acquired ordering attribute data for all the main components of dishes, all the main dishes data generated between the component attribute data mining association rules set main component, the main component of the association rules set by a plurality of rules associated with the main component composition, and calculating the principal components associated with each ruleset main component confidence association rules; the historical database ordering attribute data acquired in the characteristic of all the dishes, a characteristic of the attribute data between all the dishes generate feature data mining association rules set, wherein the association rule set of a plurality of feature association of rules, and calculating a characteristic feature associated with each rule set associated confidence rules; step C comprises: Cl, the main component of the characteristic attributes and attribute set up a main component attribute weight function and the weight function characteristic attribute ; C2, from the main component correlation set, for all the dishes in the dish X, looking for a match to the 点菜品数据库中的已点菜品的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则前件,以所述菜品X的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则后件的主成分属性关联规则集,并将获得的所述主成分属性关联规则集中的每条主成分关联规则的置信度与主成分权值相乘,即得到所述所有菜品中的菜品X的主成分属性推荐置信度,其中,所述主成分权值是根据所述主成分属性权值函数计算得到;C3、从所述特征关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种特征属性数据或者一种以上组合的特征属性数据为关联规则前件,以所述菜品X的一种特征属性数据或者一种以上组合的特征属性数据为关联规则后件的特征属性关联规则集,并将获得的所述特征属性关联规则集 A master-component attribute data points in the database are dot dishes or dishes main component composition attribute data for one or more associated rules of a former member, said attribute data in a main component X dishes or one or more combinations of the primary component attribute data attributes associated rule set after a main component member association rule, and the confidence weight of the main component is obtained by multiplying the main component and the attributes associated rule set of association rules for each main component, i.e. to give the recommended main component attribute dishes confidence in all of the dishes X, wherein the main component of the attribute weights are obtained according to the weight function calculating the main component; a C3, from the feature correlation set, all of the dishes against X dishes, looking for a match to the feature point has the attribute data in the database dishes have a characteristic attribute data point dishes or one or more combinations of antecedent rules for the association, in a characteristic of the attribute data X dishes or one or more combinations of feature attribute data is characterized by the association rule element attributes associated rule sets, and the characteristic properties of the obtained association rule set 的每条特征关联规则的置信度与特征权值相乘,即得到所述所有菜品中的菜品X的特征属性推荐置信度,其中,所述特征权值是根据所述特征属性权值函数计算得到;C4、将所述菜品X的主成分属性推荐置信度与特征属性推荐置信度相加,即得到所述菜品X的推荐价值。 And the confidence values ​​for each characteristic feature weights association rule is multiplied to obtain characteristic properties of the recommended confidence in all the dishes dishes X, wherein said characteristic weights are based on the weight function characteristic attribute get; C4, the main component attribute dishes X recommendation recommended confidence confidence characteristic attribute added to obtain the recommended value of X dishes.
  4. 4. 根据权利要求3所述的智能化推荐点菜方法,其特征在于,所述步骤C1中的主成分属性权值函数和特征属性权值函数均为递增函数。 4. The method of claim intelligent recommendation ordering according to claim 3, wherein said step C1, the main component of the weight function and property characteristic attribute weights are a function of an increasing function.
  5. 5. 根据权利要求4所述的智能化推荐点菜方法,其特征在于,所述步骤C1中的主成分属性权值函数为Y = Fl(tl) = tl衬l,其中,所述主成分属性权值函数的自变量tl的计算方法为,针对每条主成分关联规则,在关联规则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值;所述特征属性权值函数Y = F2(t2) = t2衬2,其中,所述特征属性权值函数的自变量t2的计算方法为,针对每条特征关联规则,在关联规则前件中的特征属性在所述已点菜品中出现的序号之和与所述特征属性在所述已点菜品中出现的次数之和的比值。 The intelligent recommendation ordering method according to claim 4, wherein said step C1, the main component of the attribute weight function Y = Fl (tl) = tl L lining, wherein the main component calculation of tl argument attribute weight function is, association rule for each principal component, a main component of the rule association property of the antecedent point number of dishes have been appearing on the main components and properties and the number of dishes have a ratio of the appearing point; wherein the attribute weight function Y = F2 (t2) = t2 liner 2, wherein the characteristic attribute argument weight function is calculated t2, wherein for each association rule, wherein the rule association property of the antecedent point number is the sum of the characteristic attributes appearing in the dishes and the number of dishes point has occurred ratio.
  6. 6. 根据权利要求4所述的智能化推荐点菜方法,其特征在于,所述步骤C1中的主成分属性权值函数为Y = Fl (tl) = tl衬l衬l,其中,所述主成分属性权值函数的自变量tl的计算方法为,针对每条主成分关联规则,在关联规则前件中的主成分属性在所述已点菜品中出现的序号之和与所述主成分属性在所述已点菜品中出现的次数之和的比值;所述特征属性权值函数Y = F2(t2) = t2衬2,其中,所述特征属性权值函数的自变量t2的计算方法为,针对每条特征关联规则,在关联规则前件中的特征属性在所述已点菜品中出现的序号之和与所述特征属性在所述已点菜品中出现的次数之和的比值。 The intelligent recommendation ordering method according to claim 4, wherein said step C1, the main component of the attribute weight function Y = Fl (tl) = tl l liner lining l, wherein said the method of calculating the main component of tl argument attribute weight function is, association rule for each principal component, a main component of the rule antecedent attributes associated sequence number appearing in the dishes and the main component is the point and the number of attributes appearing in a ratio of the dishes have point; wherein the attribute weight function Y = F2 (t2) = t2 liner 2, wherein said weight calculation feature attribute value of the argument of the function t2 is, for each feature association rule, wherein the rule association property of the antecedent point number is the sum of the characteristic attributes appearing in the dishes and the number of dishes point has occurred ratio.
  7. 7. —种智能化推荐点菜系统,其特征在于包括:关联规则生成模块,用于根据获取的历史点菜数据库中所有菜品的属性数据,对所述所有菜品的属性数据之间进行数据挖掘生成关联规则集,所述关联规则集由多条关联规则组成,并计算所述关联规则集中每条关联规则的置信度;其中,所述关联规则由关联规则前件和关联规则后件组成,所述关联规则后件由所述关联规则前件推导得出,定义所述历史点菜数据库中所有菜品的属性数据的数据集为P,在数据集P中有Nl次出现所述关联规则前件的同时,又有N2次出现所述关联规则后件,则所述关联规则的置信度为N2/N1W00X,用于表示所述关联规则在置信度为N2/N1W00X的概率上是可信的;第一菜品接收模块,用于接收用户输入的第一菜品,并将所述输入的第一菜品存储至已点菜品的数据库;推荐价值生成模块,用 7. - kind of intelligent recommendation ordering system, comprising: association rules generating module, for ordering the database according to the historical data of all the acquired attribute dishes, all of attribute data between the data mining dishes generating a set of association rules, the association rule set consisting of a plurality of association rules, and calculating the degree of confidence associated rule set for each association rules; wherein the association rule association rules by the rear member and the front member composed of association rules, prior to said rear member by the association rule before the association rule deduced member, defining the dataset history database ordering attribute data for all the dishes P, there is Nl times in the data set in association rule P Meanwhile member, there are N2 times after the association rule member, then the association rule confidence level N2 / N1W00X, for indicating the degree of confidence in the association rule probability N2 / N1W00X is authentic ; dishes a first receiving module, for receiving a first user input of dishes, and the dishes first input point has been stored to the database of dishes; recommended value generation module, with 从所述关联规则集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种属性数据或者一种以上组合的属性数据为关联规则前件,以所述菜品X的一种属性数据或者一种以上组合的属性数据为关联规则后件的属性关联规则集,并将获得的所述属性关联规则集中的每条关联规则的置信度相加计算得到所述菜品X的推荐价值;其中,X为所述所有菜品中的其中的一个菜品;输出模块,用于将所述所有菜品的推荐价值进行排序,选取推荐价值排名靠前的N个菜品作为推荐菜品,并输出所述N个推荐菜品,其中,所述N为大于0的整数;判断模块,用于判断是否接收到用户输入的菜品,若是,则将所述输入的菜品存储至所述已点菜品的数据库,并向推荐价值生成模块发送寻找匹配的执行指令;否则,结束推荐点菜。 From the associated rule set, for all the dishes in the dish X, has been to find a match to the database dishes point has an attribute data point dishes or one or more combinations of attribute data for association rule antecedent after the association attributes in a rule set of attribute data of the dishes or the attribute data X is one or more members in combination to association rules, and the rules associated with the attribute confidence obtained set of association rules for each calculated by adding recommended value X obtained in the dishes; wherein, X is where all of the dishes in a dish; an output module, the recommended value for all of the dishes are sorted ranking of the selected recommended value of N dishes as recommended dishes, and outputs the N recommended dishes, wherein said N is an integer greater than 0; determining module, for determining whether a user input is received dishes, and if so, then the input to the storage dishes the database has been mentioned point dishes, and recommend value generation module sends for a matching instruction execution; otherwise, the end of the recommended order.
  8. 8. 根据权利要求7所述的智能化推荐点菜系统,其特征在于,所述所有菜品的属性数据包括以下属性类别:主成分属性、特征属性。 8. The intelligent recommendation ordering system according to claim 7, wherein the attribute data includes the following attributes for all categories of dishes: a main component attribute, wherein properties.
  9. 9. 根据权利要求8所述的智能化推荐点菜系统,其特征在于,所述关联规则生成模块包括:主成分属性数据的关联规则生成单元,用于根据获取的历史点菜数据库中所有菜品的主成分属性数据,对所述所有菜品的主成分属性数据之间进行数据挖掘生成主成分关联规则集,所述主成分关联规则集由多条主成分关联规则组成,并计算所述主成分关联规则集中每条主成分关联规则的置信度;特征属性数据的关联规则生成单元,用于根据获取的历史点菜数据库中所有菜品的特征属性数据,对所述所有菜品的特征属性数据之间进行数据挖掘生成特征关联规则集,所述特征关联规则集由多条特征关联规则组成,并计算所述特征关联规则集中每条特征关联规则的置信度;所述所述推荐价值生成模块包括:权值分配单元,用于对主成分属性和特征属性分别设立主成分 9. The intelligent recommendation ordering system according to claim 8, wherein the association rule generation module comprising: a main component attribute data association rule generation means for ordering all the dishes in accordance with the historical database acquired the main component of the attribute data, the attribute data between a main component for all the dishes to generate principal component data mining association rules set, the main components of a plurality of association rule sets association rules the main component composition, and calculating the main component association rules between feature attribute data generating means, based on the history database ordering attribute data acquired in the characteristic of all the dishes, wherein the dishes all attribute data; association rule set of association rules for each main component of the confidence generating feature data mining association rules set, wherein the association rule set of a plurality of feature association of rules, and calculating said confidence feature association rule set of association rules for each feature; the recommendation value generation module comprises: weight assignment unit, for the main component attributes and characteristics of the main component were set properties 属性权值函数和特征属性权值函数;主成分属性推荐置信度生成单元,用于从所述主成分关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则前件,以所述菜品X的一种主成分属性数据或者一种以上组合的主成分属性数据为关联规则后件的主成分属性关联规则集,并将获得的所述主成分属性关联规则集中的每条主成分关联规则的置信度与主成分权值相乘,即得到所述所有菜品中的菜品X的主成分属性推荐置信度,其中,所述主成分权值是根据所述主成分属性权值函数计算得到;特征属性推荐置信度生成单元,用于从所述特征关联集中,针对所述所有菜品中的菜品X,寻找匹配以所述已点菜品数据库中的已点菜品的一种特征属性数据或者一种以上组合 Attribute weight function and the weight function characteristic attribute; main component attribute recommended confidence generating unit for correlation set from the main component for all the dishes in the dish X, has been to find a match to the database dishes point a main component the main component attribute data attribute data has a master-point component attribute data dishes or one or more combinations of the correlation rule antecedent, the one main component to the attribute data X dishes, or a combination of one or more attributes associated rule set of association rules after the main component member, with the main component and the confidence weights obtained by multiplying the main component attributes associated rule set of association rules for each principal component, i.e., to obtain all the dishes in recommended main component attribute confidence dishes X, wherein the main component of the attribute weights was calculated according to the weight function of the main component; property feature confidence recommendation generation means, wherein from said correlation set, for the All the above dishes dishes X, seeking to match a characteristic of the attribute data in the database has been dishes point points or one or more combinations of dishes 的特征属性数据为关联规则前件,以所述菜品X的一种特征属性数据或者一种以上组合的特征属性数据为关联规则后件的特征属性关联规则集,并将获得的所述特征属性关联规则集中的每条特征关联规则的置信度与特征权值相乘,即得到所述所有菜品中的菜品X的特征属性推荐置信度,其中,所述特征权值是根据所述特征属性权值函数计算得到;推荐价值生成单元,用于将所述菜品X的主成分属性推荐置信度与特征属性推荐置信度相加,即得到所述菜品X的推荐价值。 Feature attribute data wherein attribute data is associated with the rule antecedent to a characteristic of the attribute data X dishes or one or more combinations of features of the association rule member attributes associated rule sets, and the characteristic properties of the obtained wherein the confidence weights and multiplying each set of association rules in association rule features, to obtain the characteristic properties of all the dishes have recommended confidence in dishes of X, wherein said characteristic weights are based on the feature attribute weights. function value calculated; recommended value generating unit, a main component for the attributes of the recommended X dishes confidence confidence recommendation characteristic attribute added to obtain the recommended value of X dishes.
CN 200910193222 2009-10-22 2009-10-22 Method and system for intelligently recommending ordering dishes CN101697162B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 200910193222 CN101697162B (en) 2009-10-22 2009-10-22 Method and system for intelligently recommending ordering dishes

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 200910193222 CN101697162B (en) 2009-10-22 2009-10-22 Method and system for intelligently recommending ordering dishes

Publications (2)

Publication Number Publication Date
CN101697162A true true CN101697162A (en) 2010-04-21
CN101697162B CN101697162B (en) 2011-08-10

Family

ID=42142267

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 200910193222 CN101697162B (en) 2009-10-22 2009-10-22 Method and system for intelligently recommending ordering dishes

Country Status (1)

Country Link
CN (1) CN101697162B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102651052A (en) * 2012-03-29 2012-08-29 陶杰 Intelligent combo type dish ordering method
CN102915235A (en) * 2011-11-11 2013-02-06 何春望 Dish ordering and mixing software
CN103207912A (en) * 2013-04-15 2013-07-17 武汉理工大学 Method and device for intelligent service resource combination recommendation based on attribute multilevel association
CN104376021A (en) * 2013-08-16 2015-02-25 捷达世软件(深圳)有限公司 File recommending system and method
CN104794660A (en) * 2014-01-20 2015-07-22 中国移动通信集团公司 Electronic ordering method, ordering server and electronic ordering system

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101196769A (en) 2006-12-06 2008-06-11 上海市闵行中学;项 敏 Intelligent control cooling method for cabinet inside
CN100533345C (en) 2007-08-22 2009-08-26 深圳美凯电子股份有限公司 Computer intelligent power supply system and computer circuit breaking intelligent processing method

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102915235A (en) * 2011-11-11 2013-02-06 何春望 Dish ordering and mixing software
CN102651052A (en) * 2012-03-29 2012-08-29 陶杰 Intelligent combo type dish ordering method
CN103207912A (en) * 2013-04-15 2013-07-17 武汉理工大学 Method and device for intelligent service resource combination recommendation based on attribute multilevel association
CN103207912B (en) * 2013-04-15 2016-04-27 武汉理工大学 A combination of intelligent recommendation method and system resources based on service attributes associated with multiple layers
CN104376021A (en) * 2013-08-16 2015-02-25 捷达世软件(深圳)有限公司 File recommending system and method
CN104794660A (en) * 2014-01-20 2015-07-22 中国移动通信集团公司 Electronic ordering method, ordering server and electronic ordering system

Also Published As

Publication number Publication date Type
CN101697162B (en) 2011-08-10 grant

Similar Documents

Publication Publication Date Title
Kok et al. Measuring the environmental load of household consumption using some methods based on input–output energy analysis: a comparison of methods and a discussion of results
Oxenfeldt Consumer knowledge: Its measurement and extent
US20020087532A1 (en) Cooperative, interactive, heuristic system for the creation and ongoing modification of categorization systems
US20100250336A1 (en) Multi-strategy generation of product recommendations
Buckley et al. The convenience food market in Great Britain: Convenience food lifestyle (CFL) segments
US7698261B1 (en) Dynamic selection and ordering of search categories based on relevancy information
Sproles New evidence on price and product quality
US7228283B1 (en) Aesthetic profile collection
US20140089781A1 (en) Color-based designs
US20010025279A1 (en) Method and system for planning customized menu
US20060179045A1 (en) Retail store recommendation engine
US20050192869A1 (en) Recipe providing system and method thereof
US6366910B1 (en) Method and system for generation of hierarchical search results
Curtis et al. Consumer preferences for western-style convenience foods in China
JP2000331004A (en) Retrieval card system
Lassen et al. Successful strategies to increase the consumption of fruits and vegetables: results from the Danish ‘6 a day’Work-site Canteen Model Study
US20020059108A1 (en) Electronic commerce goods data search method and system with the addition of distributor strategy
US20150379733A1 (en) Automatic image-based recommendations using a color palette
CN102222164A (en) Food sensory quality evaluation method and system thereof
US7647309B1 (en) Browse relevance
Woo et al. Influence of dietary pattern on the development of overweight in a Chinese population
US20110119130A1 (en) Methods, systems, and apparatus to manage meal planning and coupon dispensing
Lakoff Identity à la carte: You are what you eat
Pavesic Prime numbers: Finding your menu's strengths
Ngapo et al. Consumer choices of pork chops: results from three panels in France

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
COR Change of bibliographic data

Free format text: CORRECT: ADDRESS; FROM: 310012 HANGZHOU KSENSE INFORMATION TECHNOLOGY DEVELOPMENT CO., LTD., ROOM 1301, VENTURE BUILDING, NO.199, WENSAN ROAD, XIHU DISTRICT, HANGZHOU CITY, ZHEJIANG PROVINCE TO: 523000 GUANGDONG KSENSE INFORMATION TECHNOLOGY CO., LTD., 809, 8/F, BUILDING B, PHASE 1, FIRST INTERNATIONAL, URBAN CENTRAL AREA, NANCHENG DISTRICT, DONGGUAN CITY, GUANGDONG PROVINCE

C41 Transfer of patent application or patent right or utility model
ASS Succession or assignment of patent right

Owner name: GUANGDONG KSENSE INFORMATION TECHNOLOGY CO., LTD.

Free format text: FORMER OWNER: GUANGDONG KSENSE INFORMATIONTECHNOLOGY DEVELOPMENT CO., LTD.

Effective date: 20110201

C14 Grant of patent or utility model
C41 Transfer of patent application or patent right or utility model
COR Change of bibliographic data

Free format text: CORRECT: ADDRESS; FROM: 523000 DONGGUAN, GUANGDONG PROVINCE TO: 510663 GUANGZHOU, GUANGDONG PROVINCE

EXPY Termination of patent right or utility model