技术领域technical field
本发明涉及电子商务领域，具体设计一种基于改进贝叶斯算法的网购风险等级评估方法。The invention relates to the field of electronic commerce, and specifically designs an online shopping risk level evaluation method based on an improved Bayesian algorithm.
背景技术Background technique
如今，随着网络的发达和普及，基于信息技术、计算机技术和网络技术的电子商务应用，由于其存在方便、价格低、选择面大等诸多优势，越来越受到广大消费者的青睐，改变了企业的生产经营方式，也改变了人们的消费方式。然而，由于电子商务的虚拟性、商家的恶性竞争、产品假冒伪劣等，如何辨识网上购物存在的风险，特别是产品质量风险，一直是消费者关心的问题。Nowadays, with the development and popularization of the Internet, e-commerce applications based on information technology, computer technology and network technology are more and more favored by consumers due to their convenience, low price, and wide selection. It has changed the way of production and operation of enterprises, and also changed the way of consumption of people. However, due to the virtual nature of e-commerce, the vicious competition of merchants, and counterfeit and shoddy products, how to identify the risks of online shopping, especially product quality risks, has always been a concern of consumers.
当前各大电子商务平台为了对更好地监督和管理网店的运营情况，也为了帮助消费者了解购买的商品和其所在网店的情况，根据消费者的反馈，制定了许多的指标，比如买家评论、评论得分、月销量、卖家服务、物流、退款退货情况等。At present, in order to better supervise and manage the operation of online stores, and to help consumers understand the purchased goods and the conditions of their online stores, major e-commerce platforms have formulated many indicators based on consumer feedback, such as Buyer reviews, review scores, monthly sales, seller services, logistics, refunds and returns, etc.
但是，由于指标的数量较大，部分指标也没有给出相比于同行业的水平，并且没有一个综合的水平划分，导致消费者无法直观、有效地评估网购活动中存在地风险情况。However, due to the large number of indicators, some indicators do not give the level compared with the same industry, and there is no comprehensive level division, which makes it impossible for consumers to intuitively and effectively assess the risks in online shopping activities.
于是，为了解决消费者对面大量指标而无法有效了解其反映的情况，帮助消费者了解此次网购中存在的风险情况，设计了一种基于改进贝叶斯算法的网购风险等级评估方法，用户只要提供相应的指标，就能得到网购相应的风险等级。Therefore, in order to solve the problem that consumers face a large number of indicators and cannot effectively understand the situation reflected by them, and help consumers understand the risks existing in this online shopping, an online shopping risk level assessment method based on the improved Bayesian algorithm is designed. Users only need to By providing the corresponding indicators, the corresponding risk level of online shopping can be obtained.
发明内容Contents of the invention
(一)要解决的技术问题(1) Technical problems to be solved
本发明提供了一种基于改进贝叶斯算法的网购风险等级评估方法，对电商平台提供的大量指标进行了综合风险等级划分，统计各指标在综合风险等级中的分布情况，并且让消费者可以通过网购对象的指标了解到存在的风险情况，提升了电商平台提供的指标的价值，对电商平台和消费者都具有很好的应用价值。The present invention provides an online shopping risk level assessment method based on the improved Bayesian algorithm, which divides a large number of indicators provided by the e-commerce platform into comprehensive risk levels, counts the distribution of each indicator in the comprehensive risk level, and allows consumers to The existing risk situation can be known through the indicators of online shopping objects, which enhances the value of the indicators provided by the e-commerce platform, and has good application value for both the e-commerce platform and consumers.
本发明提出的利用相似性的加权贝叶斯算法，有效地解决了朴素贝叶斯算法中各指标独立在现实情况下难以满足的问题，并采用各最终风险等级权重独立计算的方法，提高了分类的准确性。The weighted Bayesian algorithm using similarity proposed by the present invention effectively solves the problem that each index in the Naive Bayesian algorithm is difficult to satisfy in reality independently, and adopts the method of independent calculation of each final risk level weight, which improves the classification accuracy.
本发明提出的改进传统贝叶斯算法权重，深刻考虑到了加入权重后导致先验概率存在明显差异时对结果的影响过大的问题，对权重进行了改进，进一步提高了分类的准确性。The weight of the improved traditional Bayesian algorithm proposed by the present invention deeply considers the problem of excessive influence on the result when the prior probability is significantly different after adding the weight, and improves the weight to further improve the accuracy of classification.
(二)技术方案(2) Technical solution
一种基于改进贝叶斯算法的网购风险等级评估方法，其特征在于，所述方法包括：A method for assessing online shopping risk levels based on an improved Bayesian algorithm, characterized in that the method includes:
步骤1，使用网络爬虫技术从电商平台上抓取某类商品的商品信息和店铺信息数据。爬虫技术是从一个初始网页的URL开始，根据设计的正则表达式抓取网页中的内容和抽取新的URL，直到完成设定的任务为止。将记录数据的网页用dom4j技术提取需要的信息节点解析为文本，持久化处理，存入数据库中。Step 1, use web crawler technology to grab product information and store information data of a certain type of product from the e-commerce platform. The crawler technology starts from the URL of an initial webpage, crawls the content in the webpage and extracts new URLs according to the designed regular expressions, until the set tasks are completed. Use the dom4j technology to extract the required information nodes from the web pages that record data, parse them into text, persist them, and store them in the database.
步骤2，选择出适合的商品信息(月销量、累计评价、收藏人气、评价得分)和店铺信息(年龄、描述、服务、物流、纠纷退款率)作为风险评估的指标集合{a1,a2,...,an}。对指标进行无量纲化处理，使指标数据间具有可比性，采用的无量纲化公式是：Step 2: Select suitable product information (monthly sales, cumulative evaluation, collection popularity, evaluation score) and store information (age, description, service, logistics, dispute refund rate) as the set of indicators for risk assessment {a1 ,a2 ,...,an }. Dimensionless processing is carried out on the indicators to make the indicator data comparable. The dimensionless formula adopted is:
从数据的实际情况出发，将每种指标数据排序，平均划分成4个风险等级，取每个等级数据集合的中位数作为相应等级的指标。Starting from the actual situation of the data, the data of each index is sorted and divided into 4 risk levels on average, and the median of the data set of each level is taken as the index of the corresponding level.
步骤3，利用欧式距离法对购买各商品的风险进行等级的划分，取最小距离对应等级作为样本等级，计算公式是：Step 3, use the Euclidean distance method to classify the risk of purchasing each commodity, and take the level corresponding to the minimum distance as the sample level. The calculation formula is:
其中，k＝1，2，3，4，代表划分的4个风险等级；ai表示各指标的数值；bki表示k等级下各指标的标准值；dk表示样本与k等级之间的距离。Among them, k=1, 2, 3, 4, representing the four risk levels divided; ai represents the value of each index; bki represents the standard value of each index under the k level; dk represents the distance between the sample and the k level distance.
指标数据进行等级离散化处理是将所有样本的指标数据换成最接近的等级标准值对应的等级。选取部分作为训练集M，部分作为测试集m。The level discretization process of index data is to replace the index data of all samples with the level corresponding to the closest level standard value. Select part as training set M and part as test set m.
步骤4，对指标数据进行等级离散化处理，即数据均以等级形式表示，是将所有样本的指标数据换成最接近的等级标准值对应的等级。Step 4, perform grade discretization processing on the indicator data, that is, the data are all expressed in grade form, which is to replace the indicator data of all samples with the grade corresponding to the closest grade standard value.
步骤5，统计出需要的参数值，通过公式计算出贝叶斯公式的先验概率P(Ck)和条件概率P(xi|Ck)，其中，先验概率是各风险等级发生的概率，条件概率是各风险等级发生情况下各指标风险等级发生的概率。Step 5, count the required parameter values, and calculate the prior probability P(Ck ) and conditional probability P(xi |Ck ) of the Bayesian formula through the formula, where the prior probability is the occurrence of each risk level Probability, conditional probability is the probability of occurrence of each index risk level under the occurrence of each risk level.
步骤6，利用Jaccard相关系数计算出各指标风险等级与最终风险等级之间的相关性，并且4个最终风险等级分开计算，得到4组相关性值，通过相关性的大小获得4组权重，加入指数权重后导致在先验概率P(Ck)存在明显差异时对结果的影响过大，为了解决这一问题，将权重从和为1改进为平均值为1。Step 6, using the Jaccard correlation coefficient Calculate the correlation between the risk level of each indicator and the final risk level, and calculate the 4 final risk levels separately to obtain 4 sets of correlation values, and obtain 4 sets of weights through the size of the correlation. After adding the index weight, it will lead to a priori When there is a significant difference in the probability P(Ck ), the impact on the result is too large. In order to solve this problem, the weight is improved from 1 to the average value.
步骤7，利用改进的加权贝叶斯公式计算出测试集m中样本属于各风险等级的后验概率，比较出最大的概率属于的风险等级作为相应样本的风险等级。具体的计算公式是：Step 7: Use the improved weighted Bayesian formula to calculate the posterior probability that the samples in the test set m belong to each risk level, and compare the risk level with the highest probability as the risk level of the corresponding sample. The specific calculation formula is:
其中，V表示最终风险等级，C表示风险等级集合；k＝1，2，3，4，Ck代表划分的四方法个风险等级；X表示样本；xi表示第i个风险指标对应等级；n表示风险指标的数目；Vik表示Ck等级对应的权重组；arg表示取最大值时的Ck，即取最大的概率的等级作为样本最终等级。Among them, V represents the final risk level, C represents the set of risk levels; k=1, 2, 3, 4, Ck represents the four risk levels divided; X represents the sample; xi represents the level corresponding to the i-th risk indicator; n represents the number of risk indicators; Vik represents the weight group corresponding to the Ck level; arg represents the Ck when the maximum value is taken, that is, the level with the largest probability is taken as the final level of the sample.
(三)有益效果(3) Beneficial effects
本发明提出的基于改进贝叶斯算法的网购风险等级评估方法，能够帮助消费者更好地了解网购的对象在同行业中的风险大小情况，同时对贝叶斯算法的加权和权重的改进有效地提高了风险等级评估结果的准确性。The online shopping risk level assessment method based on the improved Bayesian algorithm proposed by the present invention can help consumers better understand the risk of online shopping objects in the same industry, and is effective for the weighting and weight improvement of the Bayesian algorithm. It greatly improves the accuracy of the risk assessment results.
附图说明Description of drawings
图1是本发明的技术路线图。Fig. 1 is a technical roadmap of the present invention.
图2是本发明的风险等级标准值设定图。Fig. 2 is a diagram of risk level standard value setting in the present invention.
图3是本发明的贝叶斯算法加权及权重改进流程图。Fig. 3 is a flowchart of Bayesian algorithm weighting and weight improvement in the present invention.
图4是本发明的改进权重效果示意图。Fig. 4 is a schematic diagram of the improved weight effect of the present invention.
具体实施方式Detailed ways
为了使本发明的目的、技术方案和有点更加清楚明了，下面结合具体实施方案并参照附图，对本发明进一步详细说明。In order to make the object, technical solution and advantages of the present invention more clear, the present invention will be further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings.
在下面的描述中阐述了很多细节以便于充分理解本发明，但是，本发明还可以在采用其他不同于此描述范围内的其他方式来实施，因此，本发明的保护范围并不受下面公开的具体实施例限制。In the following description, many details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from the scope of this description. Therefore, the protection scope of the present invention is not limited by the following disclosure. Specific embodiments are limited.
图1显示了本发明一种基于改进贝叶斯算法的网购风险等级评估方法的技术路线图。Fig. 1 shows a technical roadmap of an online shopping risk level assessment method based on the improved Bayesian algorithm of the present invention.
步骤1，以手机为例，使用网络爬虫技术从天猫上抓取手机的商品信息和店铺信息数据。爬虫技术是从搜索手机转到的第一个网页的URL开始，根据设计的正则表达式抓取网页中的内容和抽取新的URL，直到完成抓取最后一级店铺详情URL中的内容为止。将记录数据的网页用dom4j技术提取需要的信息节点解析为文本，持久化处理，存入数据库中。Step 1, taking the mobile phone as an example, use web crawler technology to grab the product information and store information data of the mobile phone from Tmall. The crawler technology starts from searching the URL of the first webpage that the mobile phone turns to, crawls the content in the webpage and extracts the new URL according to the designed regular expression, until the content in the URL of the last level of store details is crawled. Use the dom4j technology to extract the required information nodes from the web pages that record data, parse them into text, persist them, and store them in the database.
步骤2，选择出手机信息(月销量、累计评价、收藏人气、评价得分)和店铺信息(年龄、描述、服务、物流、纠纷退款率)作为风险评估的指标集合{a1,a2,...,an}，其中，n＝9。由于纠纷退款率是逆指标，先进行减法一致性处理1-xi。对指标进行无量纲化处理，使指标数据间具有可比性，采用的无量纲化公式是：Step 2, select mobile phone information (monthly sales, cumulative evaluation, collection popularity, evaluation score) and store information (age, description, service, logistics, dispute refund rate) as the set of indicators for risk assessment {a1 ,a2 , ..., an }, wherein, n=9. Since the dispute refund rate is an inverse index, the subtraction consistency processing 1-xi is performed first. Dimensionless processing is carried out on the indicators to make the indicator data comparable. The dimensionless formula adopted is:
例如手机所有样本中月销量的最大值为50000，月销量的最小值为10，那么如果某一样本的月销量为5000，无量纲化后为(5000-10)/(50000-10)＝0.0998。For example, the maximum monthly sales volume of all mobile phone samples is 50,000, and the minimum monthly sales volume is 10. If the monthly sales volume of a certain sample is 5,000, after dimensionless, it will be (5000-10)/(50000-10)=0.0998 .
如图2所示，从数据的实际情况出发，将每种指标数据排序，平均划分成4个风险等级，取每个等级数据集合的中位数作为相应等级的标准值。As shown in Figure 2, starting from the actual situation of the data, the data of each indicator is sorted and divided into 4 risk levels on average, and the median of the data set of each level is taken as the standard value of the corresponding level.
步骤3，利用欧式距离法对购买各商品的风险进行等级的划分，取最小距离对应等级作为样本等级，计算公式是：Step 3, use the Euclidean distance method to classify the risk of purchasing each commodity, and take the level corresponding to the minimum distance as the sample level. The calculation formula is:
其中，k＝1，2，3，4，代表划分的4个风险等级；ai表示各指标的数值；bki表示k等级下各指标的标准值；dk表示样本与k等级之间的距离。如表1所示。Among them, k=1, 2, 3, 4, representing the four risk levels divided; ai represents the value of each index; bki represents the standard value of each index under the k level; dk represents the distance between the sample and the k level distance. As shown in Table 1.
表1欧式距离法符号含义表Table 1 Symbol meaning table of Euclidean distance method
指标数据进行等级离散化处理是将所有样本的指标数据换成最接近的等级标准值对应的等级。选取部分作为训练集M，部分作为测试集m。The level discretization process of index data is to replace the index data of all samples with the level corresponding to the closest level standard value. Select part as training set M and part as test set m.
步骤4，对指标数据进行等级离散化处理，即数据均以等级形式表示，是将所有样本的指标数据换成最接近的等级标准值对应的等级。例如某样本中a1的值与b11、b21、b31、b41中b11最接近，那么a1的值变为1。Step 4, perform grade discretization processing on the indicator data, that is, the data are all expressed in grade form, which is to replace the indicator data of all samples with the grade corresponding to the closest grade standard value. For example, the value of a1 in a certain sample is the closest to b11 among b11 , b21 , b31 , and b41 , then the value of a1 becomes 1.
步骤5，统计出需要的参数值，通过公式计算出贝叶斯公式的先验概率P(Ck)和条件概率P(xi|Ck)，其中，先验概率是各最终风险等级发生的概率，条件概率是各最终风险等级发生情况下各指标风险等级发生的概率。Step 5, count the required parameter values, and calculate the prior probability P(Ck ) and conditional probability P(xi |Ck ) of the Bayesian formula through the formula, where the prior probability is the occurrence of each final risk level The conditional probability is the probability of occurrence of each index risk level under the occurrence of each final risk level.
步骤6，利用Jaccard相关系数即为A∩B所有样本某指标风险等级和对应最终风险等级相同的数目，A∪B是样本总数，计算出各指标风险等级与最终风险等级之间的相关性，得到相关性值，通过相关性的大小获得权重。由于贝叶斯算法中后验概率是先验概率和条件概率之间的乘积，所以乘积权重对样本属于4个风险等级的概率比较没有影响，所以加入指数权重。加入权重后的公式为：Step 6, using the Jaccard correlation coefficient That is, A∪B is the same number of the risk level of a certain indicator of all samples as the corresponding final risk level, A∪B is the total number of samples, and the correlation between the risk level of each indicator and the final risk level is calculated to obtain the correlation value. The size of the sex gets the weight. Since the posterior probability in the Bayesian algorithm is the product of the prior probability and the conditional probability, the product weight has no effect on the probability of the sample belonging to the four risk levels, so the exponential weight is added. The formula after adding the weight is:
加入小数权重后，导致P(xi|Ck)在k取不同值时差异明显缩小，在先验概率P(Ck)存在明显差异时对结果的影响过大。为了解决这一问题，将权重从和为1改进为平均值为1。原权重为(V1,V2,V3,V4)，求出平均值改进后权重为优点是既可以体现指标间重要性的差异，又可以防止条件概率影响过小，从而有效降低先验概率差异对结果的干扰。如图3所示，以P1＝0.3v1*0.4v2*0.5v3*0.6v4和P2＝0.6v1*0.5v2*0.4v3*0.3v4为例，(v1,v2,v3,v4)＝(0.1,0.2,1.8,1.9),(0.2,0.3,1.7,1.8)...(1.9,1.8,0.2,0.1),得到P1/P2的变化曲线，从图中可以看出权重影响了原本相同的P1和P2的大小关系，而且当权重在0.5-1.5之间时，倍率变化在0.5-2之间，该权重区间符合实际情况，该倍率也可以达到更好的分类效果。After adding the decimal weight, the difference of P(xi |Ck ) is obviously reduced when k takes different values, and the influence on the result is too large when there is a significant difference in the prior probability P(Ck ). To solve this problem, the weights are improved from a sum of 1 to an average of 1. The original weight is (V1 , V2 , V3 , V4 ), find the average The improved weight is The advantage is that it can not only reflect the difference in the importance of indicators, but also prevent the conditional probability from being too small, thereby effectively reducing the interference of the prior probability difference on the results. As shown in Figure 3, taking P1=0.3v1 *0.4v2 *0.5v3 *0.6v4 and P2=0.6v1 *0.5v2 *0.4v3 *0.3v4 as an example, (v1,v2,v3,v4)=(0.1, 0.2, 1.8, 1.9), (0.2, 0.3, 1.7, 1.8)...(1.9, 1.8, 0.2, 0.1), get the change curve of P1/P2, it can be seen from the figure that the weight affects the same P1 The size relationship with P2, and when the weight is between 0.5-1.5, the magnification change is between 0.5-2, the weight range is in line with the actual situation, and the magnification can also achieve better classification results.
步骤7，将4个最终风险等级分开计算，得到4组相关性值，通过相关性的大小获得4组权重。权重的改进过程如图4所示。利用改进的加权贝叶斯公式计算出测试集m中样本属于各风险等级的后验概率，比较出最大的概率属于的风险等级作为相应样本的风险等级。具体的计算公式是：In step 7, the 4 final risk levels are calculated separately to obtain 4 sets of correlation values, and 4 sets of weights are obtained through the magnitude of the correlation. The weight improvement process is shown in Figure 4. Use the improved weighted Bayesian formula to calculate the posterior probability of the samples in the test set m belonging to each risk level, and compare the risk level to which the maximum probability belongs as the risk level of the corresponding sample. The specific calculation formula is:
其中，V表示最终风险等级，C表示风险等级集合；k＝1，2，3，4，Ck代表划分的四方法个风险等级；X表示样本；xi表示第i个风险指标对应等级；n表示风险指标的数目；Vik表示Ck等级对应的权重组；arg表示取最大值时的Ck，即取最大的概率的等级作为样本最终等级。Among them, V represents the final risk level, C represents the set of risk levels; k=1, 2, 3, 4, Ck represents the four risk levels divided; X represents the sample; xi represents the level corresponding to the i-th risk indicator; n represents the number of risk indicators; Vik represents the weight group corresponding to the Ck level; arg represents the Ck when the maximum value is taken, that is, the level with the largest probability is taken as the final level of the sample.
应理解，本发明的上述具体实施方式是用于示例性说明本发明的原理，而不构成对本发明的限制。本发明所附权利要求范围和边界、或者这种范围和边界的等同形式内的全部变化和修改例。It should be understood that the above specific embodiments of the present invention are used to illustrate the principles of the present invention, but not to limit the present invention. All changes and modifications within the scope and boundaries of the appended claims of the invention, or within the equivalents of such scope and boundaries.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810008036.2ACN108229826A (en) | 2018-01-04 | 2018-01-04 | A kind of net purchase risk class appraisal procedure based on improvement bayesian algorithm |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810008036.2ACN108229826A (en) | 2018-01-04 | 2018-01-04 | A kind of net purchase risk class appraisal procedure based on improvement bayesian algorithm |
| Publication Number | Publication Date |
|---|---|
| CN108229826Atrue CN108229826A (en) | 2018-06-29 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810008036.2APendingCN108229826A (en) | 2018-01-04 | 2018-01-04 | A kind of net purchase risk class appraisal procedure based on improvement bayesian algorithm |
| Country | Link |
|---|---|
| CN (1) | CN108229826A (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109033269A (en)* | 2018-07-10 | 2018-12-18 | 卓源信息科技股份有限公司 | A kind of Distributed Area talent supply and demand subject data crawling method |
| CN110866665A (en)* | 2018-08-27 | 2020-03-06 | 中国石油化工股份有限公司 | Risk quantification assessment method and system for petrochemical production process |
| CN111724008A (en)* | 2019-03-18 | 2020-09-29 | 阿里巴巴集团控股有限公司 | Commodity identification and data processing method and equipment |
| CN113159571A (en)* | 2021-04-20 | 2021-07-23 | 中国农业大学 | Cross-border foreign species risk level determination and intelligent identification method and system |
| CN114638298A (en)* | 2022-03-15 | 2022-06-17 | 天津大学 | Aircraft attack behavior prediction method and electronic device |
| CN114693188A (en)* | 2022-05-31 | 2022-07-01 | 四川骏逸富顿科技有限公司 | Risk supervision method, system and equipment for drug retail industry |
| CN114819333A (en)* | 2022-04-24 | 2022-07-29 | 广西电网有限责任公司 | Bayesian network-based multi-stress intelligent ammeter reliability estimation method |
| CN115544902A (en)* | 2022-11-29 | 2022-12-30 | 四川骏逸富顿科技有限公司 | Drugstore risk level identification model generation method and pharmacy risk level identification method |
| CN118332248A (en)* | 2024-06-12 | 2024-07-12 | 西安航空学院 | Minors-oriented website risk probability group decision method and device |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103095485A (en)* | 2012-10-26 | 2013-05-08 | 中国航天科工集团第二研究院七〇六所 | Network risk assessment method based on combination of Bayesian algorithm and matrix method |
| CN106384123A (en)* | 2016-09-01 | 2017-02-08 | 中国地质大学（武汉） | Feature weighting filter method based on correlation and Naive Bayes classification method |
| CN106570525A (en)* | 2016-10-26 | 2017-04-19 | 昆明理工大学 | Method for evaluating online commodity assessment quality based on Bayesian network |
| CN107480895A (en)* | 2017-08-19 | 2017-12-15 | 中国标准化研究院 | A kind of reliable consumer goods methods of risk assessment based on Bayes enhancing study |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103095485A (en)* | 2012-10-26 | 2013-05-08 | 中国航天科工集团第二研究院七〇六所 | Network risk assessment method based on combination of Bayesian algorithm and matrix method |
| CN106384123A (en)* | 2016-09-01 | 2017-02-08 | 中国地质大学（武汉） | Feature weighting filter method based on correlation and Naive Bayes classification method |
| CN106570525A (en)* | 2016-10-26 | 2017-04-19 | 昆明理工大学 | Method for evaluating online commodity assessment quality based on Bayesian network |
| CN107480895A (en)* | 2017-08-19 | 2017-12-15 | 中国标准化研究院 | A kind of reliable consumer goods methods of risk assessment based on Bayes enhancing study |
| Title |
|---|
| 邹敏: ""基于数据挖掘的电子商务产品质量风险评估技术研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》* |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109033269A (en)* | 2018-07-10 | 2018-12-18 | 卓源信息科技股份有限公司 | A kind of Distributed Area talent supply and demand subject data crawling method |
| CN110866665A (en)* | 2018-08-27 | 2020-03-06 | 中国石油化工股份有限公司 | Risk quantification assessment method and system for petrochemical production process |
| CN110866665B (en)* | 2018-08-27 | 2024-01-30 | 中国石油化工股份有限公司 | Risk quantitative assessment method and system for petrochemical production process |
| CN111724008A (en)* | 2019-03-18 | 2020-09-29 | 阿里巴巴集团控股有限公司 | Commodity identification and data processing method and equipment |
| CN113159571A (en)* | 2021-04-20 | 2021-07-23 | 中国农业大学 | Cross-border foreign species risk level determination and intelligent identification method and system |
| CN113159571B (en)* | 2021-04-20 | 2024-08-27 | 中国农业大学 | A method and system for determining risk levels and intelligently identifying cross-border alien species |
| CN114638298A (en)* | 2022-03-15 | 2022-06-17 | 天津大学 | Aircraft attack behavior prediction method and electronic device |
| CN114819333A (en)* | 2022-04-24 | 2022-07-29 | 广西电网有限责任公司 | Bayesian network-based multi-stress intelligent ammeter reliability estimation method |
| CN114693188A (en)* | 2022-05-31 | 2022-07-01 | 四川骏逸富顿科技有限公司 | Risk supervision method, system and equipment for drug retail industry |
| CN115544902A (en)* | 2022-11-29 | 2022-12-30 | 四川骏逸富顿科技有限公司 | Drugstore risk level identification model generation method and pharmacy risk level identification method |
| CN118332248A (en)* | 2024-06-12 | 2024-07-12 | 西安航空学院 | Minors-oriented website risk probability group decision method and device |
| Publication | Publication Date | Title |
|---|---|---|
| CN108229826A (en) | A kind of net purchase risk class appraisal procedure based on improvement bayesian algorithm | |
| US12073444B2 (en) | Method and system for programmatic analysis of consumer reviews | |
| CN103778214B (en) | A Product Attribute Clustering Method Based on User Comments | |
| TWI567673B (en) | Sorting methods and devices for search results | |
| CN104462333B (en) | Shopping search is recommended and alarm method and system | |
| CN110222272A (en) | A kind of potential customers excavate and recommended method | |
| CN102339448B (en) | Group purchase platform information processing method and device | |
| CN109711955B (en) | Poor evaluation early warning method and system based on current order and blacklist base establishment method | |
| WO2022081267A1 (en) | Product evaluation system and method of use | |
| HK1202942A1 (en) | Method for recommending information and device thereof | |
| CN109213936B (en) | Commodity searching method and device | |
| CN110647696A (en) | Method and device for sorting business objects | |
| CN110647679A (en) | Search result comprehensive ordering method under o2o scene | |
| CN107679898A (en) | A kind of Method of Commodity Recommendation and device | |
| CN111612339B (en) | Analysis method of emotional tendency of online sales commodities based on big data | |
| CN113254775A (en) | Credit card product recommendation method based on client browsing behavior sequence | |
| CN105468628B (en) | A kind of sort method and device | |
| WO2017090095A1 (en) | Information processing device, information processing method, and information processing program | |
| Wei et al. | Understanding the information characteristics of consumers’ online reviews: the evidence from Chinese online apparel shopping | |
| CN116188086A (en) | A consumer evaluation processing method for cross-border e-commerce | |
| CN111754287A (en) | Article screening method, apparatus, device and storage medium | |
| CN105809465A (en) | Information processing method and device | |
| He et al. | Understanding users' coupon usage behaviors in E-Commerce environments | |
| CN110020195A (en) | Article recommended method and device, storage medium, electronic equipment | |
| CN110968670B (en) | Method, device, equipment and storage medium for acquiring attributes of popular commodities |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | Application publication date:20180629 | |
| RJ01 | Rejection of invention patent application after publication |