CN103744865A - Mining method for commodity evaluation words of electronic articles and system thereof - Google Patents
Mining method for commodity evaluation words of electronic articles and system thereof Download PDFInfo
- Publication number
- CN103744865A CN103744865A CN201310695873.4A CN201310695873A CN103744865A CN 103744865 A CN103744865 A CN 103744865A CN 201310695873 A CN201310695873 A CN 201310695873A CN 103744865 A CN103744865 A CN 103744865A
- Authority
- CN
- China
- Prior art keywords
- public praise
- word
- positive
- module
- new
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a mining method for commodity evaluation words of electronic articles and a system thereof. The mining method comprises the steps of selecting the articles comprising public evaluation through general test data, setting up a general prediction model, predicting the test data in industrial fields according to the general prediction model, judging and mining new positive and negative public evaluation words, accumulating the new positive and negative public evaluation words and old positive and negative public evaluation words, fetching new public evaluation declarative sentences, setting a new prediction model according to the new public evaluation declarative sentences, predicting the test data in the industrial fields again, repeating the steps of setting up a new prediction model and mining the new positive and negative public evaluation words and the new public evaluation declarative sentences, guiding all the accumulated positive and negative public evaluation words into a general vocabulary data bank after no new positive and negative public evaluation words are mined, and using the positive and negative public evaluation words as shared vocabularies. According to the method, commodity evaluation words can be mined, the general vocabulary data bank can be set up, and the evaluation of particular commodities from the network can be screened out fast and accurately according to the commodity evaluation words.
Description
Technical field
The present invention evaluates the word method of excavation and system thereof about a kind of commodity for electronic article, refer to that especially a kind of commodity of can excavating out evaluate word, and set up out a universal word data bank, while evaluating to survey the commodity of electronic article on network, can evaluate word sooner and filter out more accurately the evaluation of public opinion to particular commodity on network according to commodity.
Background technology
The main pipeline that world-wide web has been propagated for commodity information, consumer can inquire the price pre-sales in online, the rate of exchange, and except price, a larger part be on inquiry network public opinion to use gains in depth of comprehension and the evaluation of the commodity bought of wish, even there is consumer more can after consumption, online seek technical search, or the negative response of commenting, if therefore owner can provide the assistance of user's necessity or provide an explanation at above-mentioned time point, not only can reduce the discontented of user, more can pro improve the confidence of consumer to this product, to increase the consumer group who uses this product.
Therefore some dealer can screen to the article on network the evaluation of public opinion to particular commodity on network, but because different field, Different Industries are evaluated the definition of terms and usage difference all to some extent for some, even some industry (field) has exclusive evaluation word, and needing domain expert to spend a large amount of manpowers, these words just can obtain, and the generic term of general self-defining, because integrity degree is not high, so also may not be certain really can be applicable to whole industries, use.
Therefore, if can evaluate via commodity the excavation of word, set up out gradually the universal word data bank of a sustainable renewal, can cross-cuttingly using, and after the article repeated test of different field, more can continue to add new generic term, so can reach Sustainable Growth and study, so that universal word data bank can be common to various field or industry is used, so should be a best solution.
Summary of the invention
Object of the present invention is to provide a kind of commodity for electronic article to evaluate the word method of excavation and system thereof, can excavate out commodity and evaluate word, and set up out a universal word data bank, to carry out the commodity of electronic article while evaluating on network, can evaluate word sooner and filter out more accurately the evaluation of public opinion to particular commodity on network according to commodity.
Can reach the above-mentioned commodity for electronic article and evaluate the word method of excavation and system thereof, wherein for the commodity of electronic article, evaluate the word method of excavation, the steps include:
1. first with several universal test data, pick out several the articles that contain public praise, and take out after several positive and negative public praise words and several public praise declarative sentences, then according to positive and negative public praise word and public praise declarative sentence, set and set up out a general forecast model;
2. after, several industrial field test datas are predicted according to general forecast model, and judgement and excavate the new positive and negative public praise word for industrial field;
3. after more existing positive and negative public praise word and new positive and negative public praise word being added up, the positive and negative public praise word of accumulative total is taken out after several new public praise declarative sentences in several the articles that contain public praise, according to the positive and negative public praise word of accumulative total and new public praise declarative sentence, set the rule that word is marked again, to set up out a new forecast model; And
4. last, according to new forecast model, several industrial field test datas are predicted again again, and again judge and excavate new positive and negative public praise word and new public praise declarative sentence, afterwards, repeat new forecast model foundation and new positive and negative public praise word and new public praise declarative sentence excavation, and in not excavating after new positive and negative public praise word, finish excavation program.
In particular, after described end excavation program, the positive and negative public praise word all adding up is imported in a universal word data bank, with the vocabulary as shared, use.
In particular, described article is network article.
Commodity for electronic article of the present invention are evaluated word and are excavated system, comprise that a test data input module, a general forecast model are set up module, module, a vocabulary accumulative total module, a declarative sentence acquisition module, a forecast model training module excavated in a proprietary vocabulary, wherein this test data input module is in order to input test data, and this test data at least comprises several universal test data and several industrial field test datas; And this general forecast model is set up module and is comprised a bite an inscription on a tablet chapter and select module, be connected with this test data input module, this public praise article is selected module in order to pick out several the articles that contain public praise in several universal test data; Module is selected in one public praise word, selects module be connected with this public praise article, can arrange out several positive and negative public praise words in several the articles that contain public praise; One declarative sentence acquisition module, selects module and this public praise word with this public praise article and selects module and be connected, and uses positive and negative public praise word in several the articles that contain public praise, to take out several public praise declarative sentences; One general forecast model training module, selects module with this public praise word and this declarative sentence acquisition module is connected, in order to set according to positive and negative public praise word and public praise declarative sentence the rule that word is marked, to set up out a general forecast model;
This proprietary vocabulary is excavated module and is connected with this test data input module and this general forecast model training module, can several industrial field test datas be predicted according to general forecast model, and judgement and take out the new positive and negative public praise word for industrial field; And this vocabulary accumulative total module is excavated module and this public praise word with this proprietary vocabulary to select module and be connected, in order to existing positive and negative public praise word and new positive and negative public praise word are added up; This declarative sentence acquisition module, selects module with this public praise article and this vocabulary accumulative total module is connected, and uses the positive and negative public praise word of accumulative total in several the articles that contain public praise, to take out several new public praise declarative sentences; This forecast model training module, excavating module with this vocabulary accumulative total module, declarative sentence acquisition module and this proprietary vocabulary is connected, in order to the positive and negative public praise word according to accumulative total and new public praise declarative sentence, set the rule that word is marked, to set up out a new forecast model, and this proprietary vocabulary excavation module can be predicted several industrial field test datas according to new forecast model again, and repeat judgement and take out the new positive and negative public praise word for industrial field.
In particular, the described commodity for electronic article are evaluated word and are excavated system, more include a universal word data bank being connected with this vocabulary accumulative total module, wherein in this universal word data bank, there is shared positive and negative public praise word, and this proprietary vocabulary is while excavating module judgement without new positive and negative public praise word, sees through this vocabulary accumulative total module the positive and negative public praise word adding up is imported in this universal word data bank.
In particular, described article is network article.
Beneficial effect of the present invention is, the evaluation word method of excavation of the commodity for electronic article provided by the present invention and system thereof, and while mutually comparing with other located by prior art, its advantage is as follows:
1. the present invention can evaluate via commodity the excavation of word, set up out gradually the universal word data bank of a sustainable renewal, can cross-cuttingly using, and after the article repeated test of different field, more can continue to add new generic term, universal word data bank so can reach Sustainable Growth and study, so that can be common to various field or industry is used.
2. by universal word data bank of the present invention, in the time will filtering out public opinion on network to the evaluation of particular commodity to some fields, do not need again to collect public praise word, and after can directly using or first excavate with the article of this industry by universal word data bank, can in this new field, directly come into operation.
3. the present invention, in setting up general forecast model, still needs manpower to coordinate and carries out operating computer mark, but the program of setting up after general forecast model operates simulation by system oneself, therefore can allow the cost of human input drop to minimum.
4. the input test data of the present invention is huger, data quantity is larger, can to last during for industry result out can be more accurate.
Accompanying drawing explanation
Fig. 1: the present invention evaluates the schematic flow sheet of the word method of excavation and system thereof for the commodity of electronic article.
Fig. 2: the present invention evaluates the overall architecture schematic diagram of the word method of excavation and system thereof for the commodity of electronic article.
Embodiment
Relevant for aforementioned and other technology contents, feature and effect of the present invention, in the following detailed description coordinating with reference to graphic preferred embodiment, can clearly present.
Refer to Fig. 1, for the present invention evaluates the schematic flow sheet of the word method of excavation and system thereof for the commodity of electronic article, known in figure, the steps include:
1. first with several universal test data, pick out several the articles that contain public praise, and take out after several positive and negative public praise words and several public praise declarative sentences, then according to positive and negative public praise word and public praise declarative sentence, set and set up out a general forecast model 101;
2. after, several industrial field test datas are predicted according to general forecast model, and judgement and excavate the new positive and negative public praise word 102 for industrial field;
3. after more existing positive and negative public praise word and new positive and negative public praise word being added up, the positive and negative public praise word of accumulative total is taken out after several new public praise declarative sentences in several the articles that contain public praise, according to the positive and negative public praise word of accumulative total and new public praise declarative sentence, set the rule that word is marked again, to set up out a new forecast model 103; And
4. last, according to new forecast model, several industrial field test datas are predicted again again, and again judge and excavate new positive and negative public praise word and new public praise declarative sentence, afterwards, repeat new forecast model foundation and new positive and negative public praise word and new public praise declarative sentence excavation, and in not excavating after new positive and negative public praise word, finish excavation program 104.
Refer to Fig. 2 known, should evaluate word for the commodity of electronic article and excavate system, comprise test data input module 11, an one general forecast model and set up the proprietary vocabulary excavation of module 12, module 13, a vocabulary accumulative total module 14, a declarative sentence acquisition module 15, forecast model training module 16, a universal word data bank 17 and a Field Words data bank 18, wherein this test data input module 11 is in order to input test data, and the test data that can input comprises several universal test data or several industrial field test datas;
And this general forecast model is set up module 12 and is comprised a bite an inscription on a tablet chapter and select module 121, module 122 is selected in one public praise word, one declarative sentence acquisition module 123 and a general forecast model training module 124, it is to set up out a general forecast model out that this general forecast model is set up module 12 main, therefore when this test data input module 11 is selected this public praise article of several universal test data inputs after module 121, by this public praise article, select module 121 and in several universal test data, pick out several the articles that contained public praise, for example, in one piece of article, " my family's Hung treasured is drunk osmanthus Ge Xinmeili at the beginning following content, but can be just always, as diarrhoea, the S26 of Jiu Huan nest afterwards, having drunk S26 really can compare puffiness and (see now photo at that time, all feel he is good strong), and there is his milk that can overflow a period of time, so change again new angel, pacify food in one's mouth milk powder (I classmate's child also drinks this tank), after having changed new angel, drink well, now one-year-oldly changed new angel's goat milk, provide feedwater water with reference to hello! ", owing to occurring osmanthus lattice, Xin Meili, diarrhoea in this piece of article, well etc. evaluating term, therefore this piece of article can be picked out,
Afterwards, by selecting with this public praise article the public praise word that module 121 is connected, select module 122, in these several the articles that contain public praise, arrange out several positive and negative public praise words, there to be following content to illustrate, " our family drank S26 originally, and I also feel meeting puffiness to drink now aristogenesis instead ", wherein puffiness can be labeled as to negative sense public praise word, drink instead and be labeled as forward public praise word;
And mark and arrange out after several positive and negative public praise words, by selecting module 121 and this public praise word with this public praise article, select the declarative sentence acquisition module 123 that module 122 is connected, use the positive and negative public praise word of mark in several the articles that contain public praise, to take out several public praise declarative sentences, this declarative sentence acquisition module 123 can be obtained the position at commodity and public praise place, wherein can contain general complete sentence, for example, in the following article that contains public praise:
<div class=" body dfs " ><p> can remove to participate in mother classroom </p>
<p> pregnant woman also needs some activity～</p>
<p> participates in mother classroom and learns first as mother </p>
The mommy that <p> is familiar with other first can work as classmate～</p>
The course that <p> has can be taken lagnappe, and I feel it is also good surcharge～</p>
The course in <p> mommy classroom is full many </p> also
<p> similarly is that regeneration edge, aristogenesis, mamaway participated in </p> well well of experience
The knowledge that <p> acquires is really benefit from very much～</p>
Wherein " aristogenesis " and " well " are all positive and negative public praise word, therefore by " aristogenesis " and " well ", filtering out declarative sentence in this piece of article " similarly is regeneration edge, eugenic, mamaway participated in experience well well ", and declarative sentence acquisition has more long sentence punctuate rule and short sentence punctuate rule, wherein can first check sentence length, with judgement, will carry out long sentence punctuate (is to set each length to be greater than over 25 characters in the present embodiment, carry out long sentence punctuate) or short sentence is made pauses in reading unpunctuated ancient writings, and (setting each length is less than over 25 characters, carry out short sentence punctuate), long sentence punctuate rule is to make pauses in reading unpunctuated ancient writings with punctuation mark, intercept positive and negative public praise word covering scope, otherwise, short sentence punctuate rule is the front and back sentence expansion with short sentence, to intercept the covering scope of positive and negative public praise word.
Afterwards, by selecting with this public praise word the general forecast model training module 124 that module 122 and this declarative sentence acquisition module 123 is connected, according to positive and negative public praise word and public praise declarative sentence, set the rule that word is marked, this general forecast model training module 124 is used Conditional Random Fields(CRF) system is learnt, and carry out the action training that marks of word to input, after the knowledge of having been used while previously having marked with algorithm study, imitate the mark previously word being carried out, to set up out a general forecast model.
And after general forecast model sets up out, by the proprietary vocabulary being connected with this test data input module 11 and this general forecast model training module 124, excavate module 13, several the industrial field test datas of this test data input module being inputted according to general forecast model are predicted, if not belong to previously obtained positive and negative public praise word after prediction, take out as the new positive and negative public praise word for industrial field, afterwards, by selecting with this proprietary vocabulary excavation module 13 and this public praise word the vocabulary accumulative total module 14 that module 122 is connected, existing positive and negative public praise word and new positive and negative public praise word are added up, and the new positive and negative public praise word being added up more can directly be stored in a Field Words data bank 18, so that the Field Words data bank of different field can add up the proprietary word of different field,
Afterwards, select module 121 and this vocabulary declarative sentence acquisition module 15 that totally module 14 is connected with this public praise article, several new public praise declarative sentences are taken out in the positive and negative public praise word that can re-use accumulative total in several the articles that contain public praise, and again by excavating with this vocabulary accumulative total module 14, declarative sentence acquisition module 15 and this proprietary vocabulary the forecast model training module 16 that module 13 is connected, by the positive and negative public praise word of accumulative total and new public praise declarative sentence, set the rule that word is marked, to set up out a new forecast model;
Afterwards, this proprietary vocabulary is excavated module 13 and according to a new forecast model logarithm industrial field test data, is again predicted, and repeat judgement and take out the new positive and negative public praise word for industrial field, and this proprietary vocabulary is while excavating module 13 judgements without new positive and negative public praise word, seeing through this vocabulary accumulative total module 14 imports positive and negative public praise word and the declarative sentence of accumulative total in one universal word data bank 17, with the vocabulary data as general, use, and after excavating via the test data of different field, can make universal word data bank 17 more complete, and when surveying the commodity evaluation of electronic article on network in the future, according to general commodity, evaluate word and will filter out sooner and more accurately the evaluation of public opinion to particular commodity on network.
Then, the article of the actual Yi Mei of the present invention this brand of Johnson & Johnson and milk powder industry carries out implementation, and implementation step is:
1. approximately 9000 pieces of articles of Yi Mei Johnson & Johnson first, pick out approximately 500 pieces, the article that contains public praise, after the rower of going forward side by side note, arrange out approximately 300 positive and negative public praise words, take out afterwards approximately 200 narrations, after training, obtain general forecast model.
2. obtain milk powder industry totally 30,000 pieces of articles (test data), the concurrent proprietary vocabulary of milk powder industrial field that excavates, is therefore used general forecast model to predict for the first time, obtains approximately 100 neologisms.
3. again again obtaining public praise after states and chats sentence, in original 500 pieces of articles, add again after 100 neologisms, obtain 400 positive and negative public praise words (300 existing positive and negative public praise words add 100 neologisms that obtain), and use these 400 positive and negative public praise words, in original 500 pieces of articles, obtain approximately 300 narrations, after training, obtain the forecast model upgrading.
4. use new forecast model, then the data of 30,000 of milk powder industry is predicted for the second time, obtain approximately 80 neologisms.
5. original 500 pieces of articles, then add after 80 neologisms, obtain 480 positive and negative public praise words, use 480 positive and negative public praise words, in original 500 pieces of articles, obtain approximately 350 narrations, after training, obtain the forecast model upgrading.
6. continue to repeat implementation step 2,3, after the 4th test, be total to obtain 200 neologisms (100+80+20).Again without new word, the program of excavation finishes at this point afterwards.
From above-mentioned implementation step, the present invention uses universal model to predict (excavation) to the article of frontier, the result of prediction can be excavated new vocabulary, yet the vocabulary of excavating can not become the vocabulary of frontier immediately, more need additionally through after Rule screening, become candidate's vocabulary, can add artificial auxiliary judgment; And the increase of frontier vocabulary will make the quantity possibility of public praise declarative sentence increase, therefore apply mechanically the technology of public praise words and phrases acquisition, can obtain more declarative sentence as training data; If the positive and negative public praise word of above-mentioned steps 3 or declarative sentence has increase in addition, just train new model, the program of repeating step 1 to 3 again, every repetition once, just can increase field dictionary, until not newly-increased, if but there is no new declarative sentence, the program of excavating just stops, and qualified vocabulary just can add shared dictionary.And via after the article of this brand of Yi Mei Johnson & Johnson and milk powder industry is implemented above, the accuracy at resulting initial stage can reach more than 95%, more outstanding than the method commonly used on the market at present.
By the above detailed description of preferred embodiments, hope can be known description feature of the present invention and spirit more, and not with above-mentioned disclosed preferred embodiment, category of the present invention is limited.On the contrary, its objective is that hope can contain in the category of the scope of the claims of being arranged in of various changes and tool equality institute of the present invention wish application.
Claims (6)
1. for the commodity of electronic article, evaluate a word method of excavation, it is characterized in that, the steps include:
First with several universal test data, pick out several the articles that contain public praise, and take out after several positive and negative public praise words and several public praise declarative sentences, then according to positive and negative public praise word and public praise declarative sentence, set and set up out a general forecast model;
Afterwards, several industrial field test datas are predicted according to general forecast model, and judgement and excavate the new positive and negative public praise word for industrial field;
After more existing positive and negative public praise word and new positive and negative public praise word being added up, the positive and negative public praise word of accumulative total is taken out after several new public praise declarative sentences in several the articles that contain public praise, according to the positive and negative public praise word of accumulative total and new public praise declarative sentence, set the rule that word is marked again, to set up out a new forecast model; And
Finally, according to new forecast model, several industrial field test datas are predicted again again, and again judge and excavate new positive and negative public praise word and new public praise declarative sentence, afterwards, repeat new forecast model foundation and new positive and negative public praise word and new public praise declarative sentence excavation, and in not excavating after new positive and negative public praise word, finish excavation program.
2. the commodity for electronic article as claimed in claim 1 are evaluated the word method of excavation, it is characterized in that, finish, after excavation program, the positive and negative public praise word all adding up to be imported in a universal word data bank, with the vocabulary as shared, use.
3. the commodity for electronic article as claimed in claim 1 are evaluated the word method of excavation, it is characterized in that, article is network article.
4. for the commodity of electronic article, evaluate word and excavate a system, it is characterized in that, comprise:
One test data input module, in order to input test data, this test data at least comprises several universal test data and several industrial field test datas;
One general forecast model is set up module, comprises:
A bite an inscription on a tablet chapter is selected module, is connected with this test data input module, and this public praise article is selected module in order to pick out several the articles that contain public praise in several universal test data;
Module is selected in one public praise word, selects module be connected with this public praise article, can arrange out several positive and negative public praise words in several the articles that contain public praise;
One declarative sentence acquisition module, selects module and this public praise word with this public praise article and selects module and be connected, and uses positive and negative public praise word in several the articles that contain public praise, to take out several public praise declarative sentences;
One general forecast model training module, selects module with this public praise word and this declarative sentence acquisition module is connected, in order to set according to positive and negative public praise word and public praise declarative sentence the rule that word is marked, to set up out a general forecast model;
Module excavated in one proprietary vocabulary, be connected with this test data input module and this general forecast model training module, can several industrial field test datas be predicted according to general forecast model, and judgement and take out the new positive and negative public praise word for industrial field;
One vocabulary accumulative total module, excavates module and this public praise word with this proprietary vocabulary and selects module and be connected, in order to existing positive and negative public praise word and new positive and negative public praise word are added up;
One declarative sentence acquisition module, selects module with this public praise article and this vocabulary accumulative total module is connected, and uses the positive and negative public praise word of accumulative total in several the articles that contain public praise, to take out several new public praise declarative sentences;
One forecast model training module, excavating module with this vocabulary accumulative total module, declarative sentence acquisition module and this proprietary vocabulary is connected, in order to the positive and negative public praise word according to accumulative total and new public praise declarative sentence, set the rule that word is marked, to set up out a new forecast model, and this proprietary vocabulary excavation module can be predicted several industrial field test datas according to new forecast model again, and repeat judgement and take out the new positive and negative public praise word for industrial field.
5. the commodity for electronic article as claimed in claim 4 are evaluated word and are excavated system, it is characterized in that, more include a universal word data bank being connected with this vocabulary accumulative total module, wherein in this universal word data bank, there is shared positive and negative public praise word, and this proprietary vocabulary is while excavating module judgement without new positive and negative public praise word, sees through this vocabulary accumulative total module the positive and negative public praise word adding up is imported in this universal word data bank.
6. the commodity for electronic article as claimed in claim 4 are evaluated word and are excavated system, it is characterized in that, article is network article.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310695873.4A CN103744865A (en) | 2013-12-18 | 2013-12-18 | Mining method for commodity evaluation words of electronic articles and system thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310695873.4A CN103744865A (en) | 2013-12-18 | 2013-12-18 | Mining method for commodity evaluation words of electronic articles and system thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103744865A true CN103744865A (en) | 2014-04-23 |
Family
ID=50501883
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310695873.4A Pending CN103744865A (en) | 2013-12-18 | 2013-12-18 | Mining method for commodity evaluation words of electronic articles and system thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103744865A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104484329A (en) * | 2014-10-31 | 2015-04-01 | 浙江工商大学 | Consumption hot spot tracking method and device based on time series change analysis of comment Key word |
| CN107995676A (en) * | 2017-11-24 | 2018-05-04 | 北京小米移动软件有限公司 | terminal control method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040186774A1 (en) * | 2000-07-31 | 2004-09-23 | Sung Lee | Method and system for earning, storing, and using credits in exchange for satisfying predetermined conditions on a website |
| TW201118619A (en) * | 2009-11-30 | 2011-06-01 | Inst Information Industry | An opinion term mining method and apparatus thereof |
| CN102760153A (en) * | 2011-04-21 | 2012-10-31 | 帕洛阿尔托研究中心公司 | Incorporating lexicon knowledge to improve sentiment classification |
-
2013
- 2013-12-18 CN CN201310695873.4A patent/CN103744865A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20040186774A1 (en) * | 2000-07-31 | 2004-09-23 | Sung Lee | Method and system for earning, storing, and using credits in exchange for satisfying predetermined conditions on a website |
| TW201118619A (en) * | 2009-11-30 | 2011-06-01 | Inst Information Industry | An opinion term mining method and apparatus thereof |
| CN102760153A (en) * | 2011-04-21 | 2012-10-31 | 帕洛阿尔托研究中心公司 | Incorporating lexicon knowledge to improve sentiment classification |
Non-Patent Citations (1)
| Title |
|---|
| 丁晟春 等: ""基于CRF模型的半监督学习迭代观点句识别研究"", 《情报学报》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104484329A (en) * | 2014-10-31 | 2015-04-01 | 浙江工商大学 | Consumption hot spot tracking method and device based on time series change analysis of comment Key word |
| CN104484329B (en) * | 2014-10-31 | 2018-07-06 | 浙江工商大学 | Consumption hot spot method for tracing and device based on comment centre word timing variations analysis |
| CN107995676A (en) * | 2017-11-24 | 2018-05-04 | 北京小米移动软件有限公司 | terminal control method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Law et al. | Tourism demand forecasting: A deep learning approach | |
| Geng et al. | A dynamic analysis on global natural gas trade network | |
| Moen et al. | Evolution of the PDCA cycle | |
| Ivarsson et al. | Things we still haven’t learned (so far) | |
| CN111046171B (en) | Emotion discrimination method based on fine-grained labeled data | |
| Al-Harthy | Stochastic oil price models: Comparison and impact | |
| Beven | Advice to a young hydrologist | |
| CN104572915B (en) | One kind is based on the enhanced customer incident relatedness computation method of content environment | |
| CN103744865A (en) | Mining method for commodity evaluation words of electronic articles and system thereof | |
| CN108932637A (en) | A kind of training method, device and the electronic equipment of aspect mining model | |
| CN105894138A (en) | Optimum weighted composite prediction method for shipment amount of manufacturing industry | |
| Johnston et al. | Have the human geographical can (n) ons fallen silent; or were they never primed? | |
| Plakandaras et al. | US inflation dynamics on long-range data | |
| CN101876974A (en) | System and method for classifying text feeling polarities | |
| CN107563135A (en) | A kind of optimum structure equation model automatic generation method | |
| Yuping et al. | Calculating China’s historical economic aggregate: A GDP-centered overview | |
| Sampson et al. | Big Data for Specific Places in Hydrologic Modeling | |
| Dahlke et al. | Divergent transition pathways for artificial intelligence: a longitudinal and multi-level perspective using structural topic modeling | |
| Petrosyan | Procedures Designing Composite Progressive Indicators | |
| Andreica et al. | Proposal of a SMEs forecast management support system | |
| Mudalige et al. | The Effect of Entrepreneurial Orientation and Brand Orientation on SME Internalization and Performance: A Conceptual Framework | |
| Porter et al. | Integrating Environmental Monitoring and Observing Systems in Support of Science to Inform Decision-Making: Case Studies for the Southeast | |
| Boe Hüttel et al. | Deep Evidential Learning for Bayesian Quantile Regression | |
| Melton et al. | OpenET: Filling the Biggest Gap in Water Data for the Western United States | |
| Metaxas | Intelligent systems in operational quality and industrial excellence |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20140423 |
|
| WD01 | Invention patent application deemed withdrawn after publication |