CN105095396A - Model establishment method, quality assessment method and device - Google Patents
Model establishment method, quality assessment method and device Download PDFInfo
- Publication number
- CN105095396A CN105095396A CN201510387836.6A CN201510387836A CN105095396A CN 105095396 A CN105095396 A CN 105095396A CN 201510387836 A CN201510387836 A CN 201510387836A CN 105095396 A CN105095396 A CN 105095396A
- Authority
- CN
- China
- Prior art keywords
- text
- data
- weight
- model
- calculated
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/31—Indexing; Data structures therefor; Storage structures
- G06F16/313—Selection or weighting of terms for indexing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method and a device for establishing a model used in commodity quality assessment. The method comprises that relevancy between each candidate keyword in each piece of sample data in a sample set and commodity quality is calculated, and a keyword library is determined according to the calculated relevancy; in combination with the determined keyword library, a weight of a word in each piece of sample data in a sample is calculated; and the calculated weights are normalized to obtain training data, the training data is used to carry out trainings so as to output a model document, and the model document is used to describe the model obtained through the trainings. By the scheme, the commodity quality assessment will become more accurate.
Description
Technical field
The application relates to data mining technology, is specifically related to a kind of text based commercial quality analytical approach and device and relevant model creation method and apparatus.
Background technology
Along with the rise of ecommerce, shopping online is more and more universal.Like this, just have accumulated a large amount of data to need to process in service provider side.Such as, the text data such as evaluation, complaint, work order, goods return and replacement of magnanimity commodity is collected, Feature Words calculate, text training and quality analysis, realize the quantification of commercial quality and the prediction/identification of fake products, the quality of the judgement commodity of energy pin-point accuracy is an indispensable and even very important ring in whole operation.
In existing technology, use artificial modeling, and carry out commercial quality analysis according to the mode of the continuous sophisticated model of analysis result.Such as, collect text (commodity evaluation, complaint, work order, goods return and replacement etc.) data, then participle is carried out to text data, remove stop words, antonym.And then formulate commercial quality quantitative model according to historical data.In the quality evaluation stage, by the key word of specifying and text matches, and according to model successively parameter data, and optimize keyword and quantitative model according to analysis result and keyword hit situation.
But, the text matches keyword rule of thumb arranged in above-mentioned scheme, relatively low with the degree of correlation of commercial quality.In addition, the quantitative model recognition accuracy formulated by historical data is low, change and frequently cause cost increase.Further, model is in higher-dimension situation, and text does to higher-dimension control and maps, and computational complexity increases.In the prior art with Synchronization Analysis or multithread analyzing mode performance low, analyze magnanimity commodity time can not meet business need.
Summary of the invention
For one or more problem of the prior art, propose a kind of establishment for assessment of the method for the model of commercial quality, the method utilizing this model evaluation commercial quality and device.
In one aspect of the invention, provide the method for a kind of establishment for assessment of the model of commercial quality, comprise step: each candidate keywords in every bar sample data of calculating sample set and the degree of correlation of commercial quality, and according to the degree of correlation determination keywords database calculated; The weight of word in this sample in each sample data is calculated in conjunction with the keywords database determined; And the weight calculated is normalized, obtain training data, and utilize training data to carry out training with output model file, described model file is used for being described through the model of training and obtaining.
In another aspect of this invention, provide the device of a kind of establishment for assessment of the model of commercial quality, comprise: the degree of correlation calculating each candidate keywords and the commercial quality in every bar sample data of sample set, and the device of degree of correlation determination keywords database according to calculating; The device of the weight of word in this sample in each sample data is calculated in conjunction with the keywords database determined; And the weight calculated is normalized, obtain training data, and utilize training data to carry out training with output model file, described model file is used for being described through the device of training the model obtained.
In another aspect of the invention, provide a kind of method assessing commercial quality, comprise step: calculate the weight of word in the text in every bar text data of commodity to be assessed in conjunction with keywords database; The weight calculated is normalized, obtains training data; Based on the training data obtained, the model created is utilized to calculate quality category corresponding to these commodity and probability.
In still another aspect of the invention, provide a kind of device assessing commercial quality, comprising: the device calculating the weight of word in the text in every bar text data of commodity to be assessed in conjunction with keywords database; The weight calculated is normalized, obtains the device of training data; Based on the training data obtained, the model created is utilized to calculate the device of quality category corresponding to these commodity and probability.
According to such scheme, the accuracy that text based commercial quality can be provided to analyze and performance.In addition, adopt chi dynamic calculation to be used for the keyword of matched text, greatly improve the degree of correlation of keyword and commercial quality.In other examples, antonym filtration is carried out to keyword, improve the accuracy of quantized data, spatial scaling is carried out to keywords database, be beneficial to the better model of training.
In addition, use SVM support vector machine in some schemes, adopt linear algorithm to carry out linear analysis to the nonlinear characteristic of text at high-dimensional feature space, application kernel function expansion theorem, reduces computation complexity.In addition, can use distributed task scheduling carry out keyword calculating, quantification and quality analysis can fast, the quality of nearly real-time analysis magnanimity commodity.
Accompanying drawing explanation
In order to understand the present invention better, will describe the present invention according to the following drawings:
Fig. 1 shows the schematic diagram of the network structure according to the embodiment of the present invention;
Fig. 2 shows the structural representation of the server side according to the embodiment of the present invention;
Fig. 3 shows the process flow diagram of the method for the model of the establishment assessment commercial quality according to the embodiment of the present invention; And
Fig. 4 shows the process flow diagram of the method for the assessment commercial quality according to the embodiment of the present invention.
Embodiment
To specific embodiments of the invention be described in detail below, it should be noted that the embodiments described herein is only for illustrating, is not limited to the present invention.In the following description, in order to provide thorough understanding of the present invention, a large amount of specific detail has been set forth.But, those of ordinary skill in the art be it is evident that: these specific detail need not be adopted to carry out the present invention.In other instances, in order to avoid obscuring the present invention, do not specifically describe known structure, material or method.
In whole instructions, " embodiment ", " embodiment ", " example " or mentioning of " example " are meaned: the special characteristic, structure or the characteristic that describe in conjunction with this embodiment or example are at least one embodiment of the invention involved.Therefore, the phrase " in one embodiment " occurred in each place of whole instructions, " in an embodiment ", " example " or " example " differ to establish a capital and refer to same embodiment or example.In addition, can with any suitable combination and/or sub-portfolio by specific feature, structure or property combination in one or more embodiment or example.In addition, it should be understood by one skilled in the art that term "and/or" used herein comprises any and all combinations of one or more relevant project listed.
Fig. 1 shows the schematic diagram of the network structure according to the embodiment of the present invention.As shown in Figure 1, user is surfed the Net by the mobile terminal 140 and 150 of mobile phone and so on, and such as mobile terminal 140 and 150 is connected to communication network 170 by base station 160, and then is connected to the server 110 of service provider by network 170.Similarly, user also can be surfed the Net by the computing machine 120 of desktop computer and so on.Such as, desktop computer 120 is connected to communication network 170 by the network element of router one 30 and so on, and then is connected to the server 110 of service provider.Like this, user can operate 120/140 or 150 and carry out net surfing, and shopping online etc. is carried out in the website of service provider.And in service provider side, by network by commodity displaying on network, facilitate user to browse and buy.When thousands upon thousands user carries out shopping online, the quality of service provider to the commodity sold just is needed to assess and monitor on the net.This relates to the process of mass data, needs with higher accuracy rate and higher performance to carry out the process of these data.
Fig. 2 shows the structural representation of the server side according to the embodiment of the present invention.As shown in Figure 2, in the illustrated embodiment in which, realize carrying out participle to the text of commodity association, keyword calculating, quantification, spatial scaling, model training and quality analysis.Be described on the basis of SKU dimension below, but those skilled in the art should be able to expect implementing the present invention in other dimension.
As shown in Figure 2, management end UI201 is the user interface of the manager works of convenient service provider, such as, carry out the collection of data, sends instruction carrying out quality analysis etc.In addition, management end UI can also be used to adjust keywords database or set filtrator, such as, arrange antonym needing to filter out etc.
Timer or user (operation assistant director) 202 trigger and carry out commercial quality analysis.Collaborative work unit 204 generates task coordinate and distributes to keyword computing unit 205, word quantifying unit 207, training unit 208 and mass analysis cell 214 and runs successively, and each unit all can distributed parallel analysis.Each unit can find 1 corresponding Master and multiple Follow, as a certain Master stops service, and other Follow can select a Follow to take over the task of Master according to the spare time situation of doing of example, and obtain up-to-date task status, the consistance of assurance device reliability and data.
Keyword computing unit 205 calculates the degree of correlation of each candidate keywords in every bar sample data of sample set and commercial quality, and according to the degree of correlation determination keywords database calculated.Such as, keyword computing unit 205 utilizes chi to calculate all keywords and (originates such as: to the text participle of all sample datas, remove stop words, antonym) and the degree of correlation of commercial quality, and sort from high to low, number, and get top n as keywords database, make keywords database and commercial quality height correlation.Chi is a kind of hypothesis testing method, utilizes independence test whether to have relation to investigate Two Variables, and quantizes the degree of reliability of this judgement.Chi formula is as follows:
SKU represents the minimum available units of preserving storage controlling, is convenient to recognition value, as a commodity polychrome, then and corresponding multiple SKU.For SKU dimensional analysis:
A be keyword in fake products classification, the number of times that all texts occurs;
B be keyword not in fake products classification, the number of times that all texts occurs
C be keyword in fake products classification, the text number do not occurred;
D be keyword not in fake products classification, the text number do not occurred;
N is the text sum of sample data;
After Feature Words calculating, sequence, data layout is as following table:
| Numbering | Feature Words | Chi-square value |
| 1 | Fake products | 57.18 |
| 2 | Substandard products | 52.39 |
| 3 | Imitative goods | 38.87 |
| 4 | Seriously | 32.66 |
| 5 | There is problem | 16.30 |
| … | … | … |
Word quantifying unit 207 combines the keywords database determined and calculates the weight of word in this sample in each sample data.Such as word quantifying unit 207 to every bar sample data (comprise the quality category that commodity are corresponding: good, in, poor) carry out participle, remove stop words, antonym, then calculate each word weight in the sample in conjunction with keywords database, computing formula:
In above formula, N is the text sum that SKU is corresponding, and molecule is the frequency (repeatedly calculating once appears in a text) that keyword occurs in sample SKU, and denominator is keyword occurrence number summation in other classifications.
To sample data quantized data form as following table:
| SKU | Key words sorting | Feature number | Weighted value | Feature number | Weighted value | Feature number | Weighted value | … |
| 10001 | A | 2 | 0.81 | 6 | 0.61 | 10 | 0.01 | … |
| 10002 | B | 3 | 0.73 | 35 | 0.83 | 21 | 2.15 | … |
| 10003 | A | 5 | 1.11 | 78 | 1.23 | 156 | 0.28 | … |
| 10004 | … | … | … | … | … | … | … | … |
Training unit 208 is normalized the weight calculated, and obtains training data, and utilizes training data to carry out training with output model file, and described model file is used for being described through the model of training and obtaining.Such as the word quantized data that word quantifying unit 207 generates first is carried out spatial scaling by training unit 208, the scope (being beneficial to the better model of training) of unified quantization value, then training set 209 is generated, obtain model file 210, described model file is used for being described through the model of training and obtaining.In training set 209, a sample data generates a training data.
After carrying out spatial scaling, data layout is as following table:
| SKU | Key words sorting | Feature number | Weighted value | Feature number | Weighted value | Feature number | Weighted value | … |
| 10001 | A | 2 | 0.81 | 6 | 0.61 | 10 | 0.01 | … |
| 10002 | B | 3 | 0.73 | 35 | 0.83 | 21 | 0.85 | … |
| 10003 | A | 5 | 0.96 | 78 | 0.93 | 156 | 0.28 | … |
| 10004 | … | … | … | … | … | … | … | … |
Such as, training unit 208 uses SVM training set 209 is trained to model file 210 and exports.SVM expresses support for vector machine, is a kind of trainable machine learning method.
After obtaining model file, mass analysis cell 214 reads commodity text data, and use and method generation text training data that word quantifying unit, training unit are same, and calculate classification corresponding to certain commercial quality and probability according to the model trained by SVM;
Output result data is as following table:
| SKU | Result classification |
| 20001 | A |
| 20002 | B |
| 20003 | A |
| 20004 | … |
User safeguards sample data, keyword or antisense dictionary by management end UI201, arranges commercial quality analytical parameters and retrieval quality result data.
In the above-described embodiment the constructive process of model and the assessment of commercial quality are combined description, but the above embodiments also can be divided into the establishing stage of model and the evaluation stage of commercial quality by those skilled in the art.Such as, carry out after training obtains model file 210, can preserving for later use in memory at training unit 208 pairs of training sets 209.Fig. 3 shows the process flow diagram of the method for the model of the establishment assessment commercial quality according to the embodiment of the present invention.
In step S310, each candidate keywords in every bar sample data of calculating sample set and the degree of correlation of commercial quality, and according to the degree of correlation determination keywords database calculated.
Such as, utilize chi to calculate all keywords (to originate such as: to the text participle of all sample datas, remove stop words, antonym) and the degree of correlation of commercial quality, and sort from high to low, number, and get top n as keywords database, make keywords database and commercial quality height correlation.
In step S320, calculate the weight of word in this sample in each sample data in conjunction with the keywords database determined.
Such as, to every bar sample data (comprise the quality category that commodity are corresponding: good, in, poor) carry out participle, remove stop words, antonym, then calculate each word weight in the sample in conjunction with keywords database.
In step S330, be normalized, obtain training data, and utilize training data to carry out training with output model file the weight calculated, described model file is used for being described through the model of training and obtaining.
Such as, first the word quantized data of generation is carried out spatial scaling, the scope (being beneficial to the better model of training) of unified quantization value, then generate training set, obtain model file, this model file is used for being described through the model of training and obtaining.In training set, a sample data generates a training data.
In addition, in commercial quality evaluation stage, the model file that can preserve is to carry out quality evaluation.Such as, in the quality evaluation stage, the text data of the commodity that will assess is processed in word quantifying unit 212 and training unit 213, then carry out quality analysis.Here word quantifying unit 212 is identical with the function of training unit 208 with word quantifying unit 207 with training unit 213, and the data of just process are different.Like this, mass analysis cell 214 carries out follow-up quality analysis, obtains the quality analysis results 215 of these commodity.
Fig. 4 shows the process flow diagram of the method for the assessment commercial quality according to the embodiment of the present invention.As shown in Figure 4, in step S410, calculate the weight of word in the text in every bar text data of commodity to be assessed in conjunction with keywords database.Such as in conjunction with the word weight in the text in the text data 211 of keywords database commodity to be assessed.
In step S420, the weight calculated is normalized, obtains training data.In step S430, based on the training data obtained, the model created is utilized to calculate quality category corresponding to these commodity and probability.
The method provided by the invention of above-described embodiment, adopts the keyword of method dynamic calculation for matched text of such as chi and so on, greatly improves the degree of correlation of keyword and commercial quality, and without the need to manual maintenance.In addition, in single keyword in above-described embodiment and the computation model of the SKU degree of correlation, the probability that this degree of correlation and keyword occur in SKU is directly proportional, and is inversely proportional to, has perfectly showed the correlativity of keyword and SKU with keyword occurrence number summation in other classifications
In addition, word quantizing process carries out antonym filtration to keyword, improves the accuracy of quantized data, carries out spatial scaling to keywords database, is beneficial to the better model of training, improves accuracy.
In certain embodiments, use SVM support vector machine, adopt linear algorithm to carry out linear analysis to the nonlinear characteristic of text at high-dimensional feature space, application kernel function expansion theorem, relative traditional data mining method, reduces computation complexity.And along with the change of sample data, model timing regenerates, and reduces the loaded down with trivial details degree of human intervention.
In certain embodiments, use distributed task scheduling carry out keyword calculating, quantification and quality analysis can fast, the quality of near real-time analysis magnanimity commodity, in a word, the present invention to improve commodity quality the accuracy of quantized data and consistance greatly relative to prior art, analysis speed is fast, and reliability is high.
As understood by those skilled in the art, in the above-described embodiments, collaborative work unit can omit, and has certain influence to analytical performance, but does not affect the net result of this method or device.In addition keyword and antisense dictionary can be manual maintenances, and under line, the model of Erecting and improving is optimized it.In addition, above-mentioned method and apparatus can carry out quality analysis (SPU, SKU, shop, operator etc.) for the different dimensions of commodity.SPU is the set of standardized information of one group of reusable, easily retrieval, and this set description characteristic of a product is the least unit of merchandise news polymerization.
In addition, in word quantifying unit, relatedness computation model also can be distinguished according to the type of text (evaluation, complaint, work order, goods return and replacement).
According to some embodiments, above-mentioned commercial quality classification can self-defining, does not affect this method and device to the judgement of commercial quality.
Above detailed description, by using schematic diagram, process flow diagram and/or example, has set forth numerous embodiments of a kind of text based commercial quality analytical approach and device.When this schematic diagram, process flow diagram and/or example comprise one or more function and/or operation, it will be understood by those skilled in the art that each function in this schematic diagram, process flow diagram or example and/or operation can by various structure, hardware, software, firmware or in fact their combination in any come to realize separately and/or jointly.In one embodiment, some parts of theme described in embodiments of the invention can be realized by special IC (ASIC), field programmable gate array (FPGA), digital signal processor (DSP) or other integrated forms.But, those skilled in the art will recognize that, some aspects of embodiment disclosed herein can realize in integrated circuits on the whole or partly equally, be embodied as one or more computer programs of running on one or more computing machine (such as, be embodied as the one or more programs run in one or more computer system), be embodied as one or more programs of running on the one or more processors (such as, be embodied as the one or more programs run on one or more microprocessor), be embodied as firmware, or be embodied as in fact the combination in any of aforesaid way, and those skilled in the art are according to the disclosure, the ability of design circuit and/or write software and/or firmware code will be possessed.In addition, those skilled in the art will recognize that, the mechanism of theme described in the disclosure can be distributed as the program product of various ways, and regardless of the actual particular type of signal bearing medium being used for performing distribution, the exemplary embodiment of theme described in the disclosure is all applicable.The example of signal bearing medium includes but not limited to: recordable-type media, as floppy disk, hard disk drive, compact-disc (CD), digital universal disc (DVD), numerical tape, computer memory etc.; And transmission type media, as numeral and/or analogue communication medium (such as, optical fiber cable, waveguide, wired communications links, wireless communication link etc.).
Although exemplary embodiment describe the present invention with reference to several, should be appreciated that term used illustrates and exemplary and nonrestrictive term.Spirit or the essence of invention is not departed from because the present invention can specifically implement in a variety of forms, so be to be understood that, above-described embodiment is not limited to any aforesaid details, and explain widely in the spirit and scope that should limit in claim of enclosing, therefore fall into whole change in claim or its equivalent scope and remodeling and all should be claim of enclosing and contained.
Claims (12)
1. create the method for assessment of the model of commercial quality, comprise step:
Each candidate keywords in every bar sample data of calculating sample set and the degree of correlation of commercial quality, and according to the degree of correlation determination keywords database calculated;
The weight of word in this sample in each sample data is calculated in conjunction with the keywords database determined; And
Be normalized the weight calculated, obtain training data, and utilize training data to carry out training with output model file, described model file is used for being described through the model of training and obtaining.
2. the method for claim 1, wherein utilizes chi to calculate the degree of correlation of all candidate keywords and commercial quality, and extracts top n as keywords database:
A be keyword in fake products classification, the number of times that all texts occurs;
B be keyword not in fake products classification, the number of times that all texts occurs
C be keyword in fake products classification, the text number do not occurred;
D be keyword not in fake products classification, the text number do not occurred;
N is the text sum of sample data.
3. the method for claim 1, is also included in before calculating weight and carries out participle to every bar sample data, and remove stop words and antonym.
4. the method for claim 1, wherein calculates each word weight in the sample according to the following formula:
Above formula Middle molecule is the frequency that keyword occurs in sample set, and denominator is keyword occurrence number summation in other classifications, and N is text sum.
5. the method for claim 1, wherein utilizes algorithm of support vector machine to obtain model file based on training data.
6. create the device for assessment of the model of commercial quality, comprising:
Calculate the degree of correlation of each candidate keywords and the commercial quality in every bar sample data of sample set, and the device of degree of correlation determination keywords database according to calculating;
The device of the weight of word in this sample in each sample data is calculated in conjunction with the keywords database determined; And
Be normalized the weight calculated, obtain training data, and utilize training data to carry out training with output model file, described model file is used for being described through the device of training the model obtained.
7. assess a method for commercial quality, comprise step:
The weight of word in the text in every bar text data of commodity to be assessed is calculated in conjunction with keywords database;
The weight calculated is normalized, obtains training data;
Based on the training data obtained, the model created is utilized to calculate quality category corresponding to these commodity and probability.
8. method as claimed in claim 7, wherein said keywords database is the degree of correlation by each candidate keywords in every bar sample data of calculating sample set and commercial quality, and determine according to the degree of correlation calculated.
9. method as claimed in claim 7, wherein calculates each word weight in the text according to the following formula:
Above formula Middle molecule is the frequency that keyword occurs in the text, and denominator is keyword occurrence number summation in other classifications, and N is text sum.
10. method as claimed in claim 7, is also included in before calculating weight and carries out participle to every bar text data, and remove stop words and antonym.
11. 1 kinds of devices assessing commercial quality, comprising:
The device of the weight of word in the text in every bar text data of commodity to be assessed is calculated in conjunction with keywords database;
The weight calculated is normalized, obtains the device of training data;
Based on the training data obtained, the model created is utilized to calculate the device of quality category corresponding to these commodity and probability.
12. devices as claimed in claim 11, wherein the mode of task assesses the quality of commodity in a distributed manner.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510387836.6A CN105095396A (en) | 2015-07-03 | 2015-07-03 | Model establishment method, quality assessment method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510387836.6A CN105095396A (en) | 2015-07-03 | 2015-07-03 | Model establishment method, quality assessment method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105095396A true CN105095396A (en) | 2015-11-25 |
Family
ID=54575833
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510387836.6A Pending CN105095396A (en) | 2015-07-03 | 2015-07-03 | Model establishment method, quality assessment method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105095396A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105790866A (en) * | 2016-03-31 | 2016-07-20 | 中国联合网络通信集团有限公司 | Base station grading method and base station grading device |
| CN106919545A (en) * | 2015-12-25 | 2017-07-04 | 皇家飞利浦有限公司 | Method and apparatus for processing document |
| CN107274043A (en) * | 2016-04-07 | 2017-10-20 | 阿里巴巴集团控股有限公司 | Quality evaluating method, device and the electronic equipment of forecast model |
| CN107908783A (en) * | 2017-12-07 | 2018-04-13 | 百度在线网络技术（北京）有限公司 | Retrieve appraisal procedure, device, server and the storage medium of text relevant |
| CN107944761A (en) * | 2017-12-18 | 2018-04-20 | 广东广业开元科技有限公司 | Early warning and monitoring analysis method is complained based on artificial intelligence protection of consumers' rights index enterprise |
| CN108304868A (en) * | 2018-01-25 | 2018-07-20 | 阿里巴巴集团控股有限公司 | Model training method, data type recognition methods and computer equipment |
| CN108399545A (en) * | 2017-02-06 | 2018-08-14 | 北京京东尚科信息技术有限公司 | E-commerce platform quality determining method and device |
| CN108429776A (en) * | 2017-02-13 | 2018-08-21 | 阿里巴巴集团控股有限公司 | Method for pushing, device, client, interactive device and the system of network object |
| CN109473149A (en) * | 2018-11-09 | 2019-03-15 | 天津开心生活科技有限公司 | Data Quality Assessment Methodology, device, electronic equipment and computer-readable medium |
| CN110363251A (en) * | 2019-07-23 | 2019-10-22 | 杭州嘉云数据科技有限公司 | A kind of SKU image classification method, device, electronic equipment and storage medium |
| CN111310431A (en) * | 2018-11-27 | 2020-06-19 | 浙江智美知识产权服务有限公司 | Quantitative intelligent processing method |
| CN111598585A (en) * | 2020-04-30 | 2020-08-28 | 山东贵合信息科技有限公司 | Agricultural product quality tracing method, equipment and medium |
| CN111600993A (en) * | 2020-05-13 | 2020-08-28 | 北京华峰创业科技有限公司 | Method and device for stroke reminding according to short message |
| WO2021179138A1 (en) * | 2020-03-09 | 2021-09-16 | 图灵通诺（北京）科技有限公司 | Method and system for analyzing commodities on supermarket shelf |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663025A (en) * | 2012-03-22 | 2012-09-12 | 浙江盘石信息技术有限公司 | Illegal online commodity detection method |
| CN103699523A (en) * | 2013-12-16 | 2014-04-02 | 深圳先进技术研究院 | Product classification method and device |
| CN103778548A (en) * | 2012-10-19 | 2014-05-07 | 阿里巴巴集团控股有限公司 | Goods information and keyword matching method, and goods information releasing method and device |
| CN103778205A (en) * | 2014-01-13 | 2014-05-07 | 北京奇虎科技有限公司 | Commodity classifying method and system based on mutual information |
| CN103823893A (en) * | 2014-03-11 | 2014-05-28 | 北京大学 | User comment-based product search method and system |
| CN103853824A (en) * | 2014-03-03 | 2014-06-11 | 沈之锐 | In-text advertisement releasing method and system based on deep semantic mining |
-
2015
- 2015-07-03 CN CN201510387836.6A patent/CN105095396A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102663025A (en) * | 2012-03-22 | 2012-09-12 | 浙江盘石信息技术有限公司 | Illegal online commodity detection method |
| CN103778548A (en) * | 2012-10-19 | 2014-05-07 | 阿里巴巴集团控股有限公司 | Goods information and keyword matching method, and goods information releasing method and device |
| CN103699523A (en) * | 2013-12-16 | 2014-04-02 | 深圳先进技术研究院 | Product classification method and device |
| CN103778205A (en) * | 2014-01-13 | 2014-05-07 | 北京奇虎科技有限公司 | Commodity classifying method and system based on mutual information |
| CN103853824A (en) * | 2014-03-03 | 2014-06-11 | 沈之锐 | In-text advertisement releasing method and system based on deep semantic mining |
| CN103823893A (en) * | 2014-03-11 | 2014-05-28 | 北京大学 | User comment-based product search method and system |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106919545A (en) * | 2015-12-25 | 2017-07-04 | 皇家飞利浦有限公司 | Method and apparatus for processing document |
| CN105790866A (en) * | 2016-03-31 | 2016-07-20 | 中国联合网络通信集团有限公司 | Base station grading method and base station grading device |
| CN105790866B (en) * | 2016-03-31 | 2018-07-24 | 中国联合网络通信集团有限公司 | Base station rankings method and device |
| CN107274043B (en) * | 2016-04-07 | 2021-05-07 | 阿里巴巴集团控股有限公司 | Quality evaluation method and device of prediction model and electronic equipment |
| CN107274043A (en) * | 2016-04-07 | 2017-10-20 | 阿里巴巴集团控股有限公司 | Quality evaluating method, device and the electronic equipment of forecast model |
| CN108399545A (en) * | 2017-02-06 | 2018-08-14 | 北京京东尚科信息技术有限公司 | E-commerce platform quality determining method and device |
| CN108429776B (en) * | 2017-02-13 | 2021-09-14 | 阿里巴巴集团控股有限公司 | Network object pushing method, device, client, interaction equipment and system |
| CN108429776A (en) * | 2017-02-13 | 2018-08-21 | 阿里巴巴集团控股有限公司 | Method for pushing, device, client, interactive device and the system of network object |
| CN107908783A (en) * | 2017-12-07 | 2018-04-13 | 百度在线网络技术（北京）有限公司 | Retrieve appraisal procedure, device, server and the storage medium of text relevant |
| CN107944761A (en) * | 2017-12-18 | 2018-04-20 | 广东广业开元科技有限公司 | Early warning and monitoring analysis method is complained based on artificial intelligence protection of consumers' rights index enterprise |
| CN108304868A (en) * | 2018-01-25 | 2018-07-20 | 阿里巴巴集团控股有限公司 | Model training method, data type recognition methods and computer equipment |
| CN109473149B (en) * | 2018-11-09 | 2021-01-15 | 天津开心生活科技有限公司 | Data quality evaluation method and device, electronic equipment and computer readable medium |
| CN109473149A (en) * | 2018-11-09 | 2019-03-15 | 天津开心生活科技有限公司 | Data Quality Assessment Methodology, device, electronic equipment and computer-readable medium |
| CN111310431A (en) * | 2018-11-27 | 2020-06-19 | 浙江智美知识产权服务有限公司 | Quantitative intelligent processing method |
| CN110363251A (en) * | 2019-07-23 | 2019-10-22 | 杭州嘉云数据科技有限公司 | A kind of SKU image classification method, device, electronic equipment and storage medium |
| WO2021179138A1 (en) * | 2020-03-09 | 2021-09-16 | 图灵通诺（北京）科技有限公司 | Method and system for analyzing commodities on supermarket shelf |
| US11748787B2 (en) | 2020-03-09 | 2023-09-05 | Yi Tunnel (Beijing) Technology Co., Ltd. | Analysis method and system for the item on the supermarket shelf |
| CN111598585A (en) * | 2020-04-30 | 2020-08-28 | 山东贵合信息科技有限公司 | Agricultural product quality tracing method, equipment and medium |
| CN111600993A (en) * | 2020-05-13 | 2020-08-28 | 北京华峰创业科技有限公司 | Method and device for stroke reminding according to short message |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105095396A (en) | Model establishment method, quality assessment method and device | |
| CN110222272B (en) | Potential customer mining and recommending method | |
| Cai et al. | Multimodal Data Guided Spatial Feature Fusion and Grouping Strategy for E‐Commerce Commodity Demand Forecasting | |
| CN108960719B (en) | Method and device for selecting products and computer readable storage medium | |
| CN102890803B (en) | The defining method of the abnormal process of exchange of electronic goods and device thereof | |
| CN108985830B (en) | Recommendation scoring method and device based on heterogeneous information network | |
| CN104463637A (en) | Commodity recommendation method and device based on electronic business platform and server | |
| Hadden et al. | Churn prediction using complaints data | |
| CN112200601B (en) | Item recommendation method, device and readable storage medium | |
| Kabir et al. | Analysis of different predicting model for online shoppers’ purchase intention from empirical data | |
| CN105335518A (en) | Method and device for generating user preference information | |
| CN104834651A (en) | Method and apparatus for providing answers to frequently asked questions | |
| CN104850617A (en) | Short text processing method and apparatus | |
| CN111966886A (en) | Object recommendation method, object recommendation device, electronic equipment and storage medium | |
| CN110647995A (en) | Rule training method, device, equipment and storage medium | |
| CN111738805A (en) | Behavior log-based search recommendation model generation method, device and storage medium | |
| CN110956278A (en) | Method and system for retraining machine learning models | |
| CN102985939A (en) | Art evaluation engine and method for automatic development of an art index | |
| Desai et al. | Business intelligence visualization using deep learning based sentiment analysis on amazon review data | |
| Bahtiar et al. | Comparison of naïve Bayes and logistic regression in sentiment analysis on marketplace reviews using rating-based labeling | |
| CN104572623B (en) | A kind of efficient data analysis and summary method of online LDA models | |
| JP5835802B2 (en) | Purchase forecasting apparatus, method, and program | |
| Nasiri et al. | Aspect category detection on indonesian e-commerce mobile application review | |
| Silva et al. | A categorical clustering of publishers for mobile performance marketing | |
| Singh et al. | Twitter sentiment analysis for stock prediction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20151125 |
|
| RJ01 | Rejection of invention patent application after publication |