CN109657248A - A kind of comment and analysis method, apparatus, equipment and storage medium - Google Patents
A kind of comment and analysis method, apparatus, equipment and storage medium Download PDFInfo
- Publication number
- CN109657248A CN109657248A CN201811585336.3A CN201811585336A CN109657248A CN 109657248 A CN109657248 A CN 109657248A CN 201811585336 A CN201811585336 A CN 201811585336A CN 109657248 A CN109657248 A CN 109657248A
- Authority
- CN
- China
- Prior art keywords
- viewpoint
- comment
- expression pattern
- analyzed
- polarity
- 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
- 238000004458 analytical method Methods 0.000 title claims abstract description 39
- 239000013598 vector Substances 0.000 claims description 31
- 230000015654 memory Effects 0.000 claims description 23
- 238000000034 method Methods 0.000 claims description 21
- 239000011159 matrix material Substances 0.000 claims description 14
- 238000012546 transfer Methods 0.000 claims description 11
- 230000008901 benefit Effects 0.000 abstract description 10
- 238000012552 review Methods 0.000 abstract description 8
- 238000012545 processing Methods 0.000 abstract description 5
- 230000005611 electricity Effects 0.000 description 9
- 230000008569 process Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 230000001360 synchronised effect Effects 0.000 description 5
- 230000000875 corresponding effect Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 241000282326 Felis catus Species 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 230000008451 emotion Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000035807 sensation Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000009412 basement excavation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 239000013078 crystal Substances 0.000 description 1
- 238000000151 deposition Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 235000021022 fresh fruits Nutrition 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000035800 maturation Effects 0.000 description 1
- 230000036651 mood Effects 0.000 description 1
- 235000013324 preserved food Nutrition 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Game Theory and Decision Science (AREA)
- Economics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present embodiments relate to technical field of data processing, and in particular to a kind of comment and analysis method, apparatus, equipment and storage medium.The comment and analysis method includes: to obtain at least one comment to be analyzed；The viewpoint of the comment to be analyzed is determined according to reference viewpoint expression pattern collection and the polarity of the viewpoint of the comment to be analyzed is determined according to preset viewpoint polarity standard；Wherein, described with reference to viewpoint expression pattern collection is according to multiple set of reference viewpoint expression patterns commented on and be trained.Advantage of the embodiment of the present invention are as follows: obtained according to multiple consumer reviews with reference to viewpoint collection, the viewpoint in comment to be analyzed is determined using reference viewpoint, and the polarity of viewpoint is determined using viewpoint polarity standard, so as to obtain consumer point expressed by the comment to be analyzed, referred to for goods providers.
Description
Technical field
The present embodiments relate to data processing fields, and in particular to a kind of comment and analysis method, apparatus, equipment and storage
Medium.
Background technique
It, generally can be with reference to having bought the commodity before not only consumer buys a kind of commodity for shopping at network
The comment that consumer delivers, and the supplier of the commodity is also interested in the viewpoint in comment, because consumer is to commodity
View can reflect the advantage and disadvantage of the commodity, so as to so that businessman targetedly takes corrective measure.
By taking smartwatch as an example, smartwatch is a kind of wearable product of intelligence, and the consumer for much having bought wrist-watch can be
Its wrist-watch bought is commented in electric business website, such as: it " receives after wrist-watch till now, I am generally still satisfied, total body-sensing
Feel, should summarize very much, function is more powerful, and stand-by time is very long.In purchasing process, customer service attitude is still
It is very good.Purchasing process customer service is relatively good, as after-sale service, comments on again after later period experience."
Above it is exactly a typical comment of the consumer to wrist-watch, expresses the meaning to this product correlated characteristic of wrist-watch
See.
Consumer reviews, research staff, sale people of the comment of consumer to product can be generated hour to hour and day to day on the net
Member and product personnel have very strong directive significance, such as can be by the comment of analysis consumer, it is known that user is most
Be concerned about wrist-watch which feature, to which be characterized in being discontented with very much etc..
But in the prior art, comment and analysis can only be completed by manpower, be a very time-consuming job.
Summary of the invention
For this purpose, the embodiment of the present invention provides a kind of comment and analysis method, apparatus, equipment and storage medium, it is existing to solve
Time-consuming and laborious, low efficiency, problem at high cost caused by comment and analysis is completed by manpower in technology.
To achieve the goals above, embodiments of the present invention provide the following technical solutions:
In the first aspect of embodiments of the present invention, a kind of comment and analysis method is provided, comprising: obtain at least one
Item comment to be analyzed；The viewpoint of the comment to be analyzed is determined according to reference viewpoint expression pattern collection and according to preset viewpoint
Polarity standard determines the polarity of the viewpoint of the comment to be analyzed；Wherein, the reference viewpoint expression pattern collection is according to multiple
Comment on the set for the reference viewpoint expression pattern being trained.
In one embodiment of the invention, the method also includes: the viewpoint of the comment to be analyzed is mapped to pre-
Bidding is signed；The polarity for counting the viewpoint of all comments to be analyzed for being mapped to the default label, obtains statistical result.
In one embodiment of the invention, the default label is the label of user preset；The method also includes: it will
The statistical result is sent to the user.
In one embodiment of the invention, described to be obtained with reference to viewpoint expression pattern collection by following steps: to obtain more
Item comment；Candidate viewpoint expression pattern collection is constructed according to the content of the comment, candidate's viewpoint set of patterns includes multiple times
Select viewpoint expression pattern；Multiple candidate viewpoint expression patterns are ranked up；The reference is obtained according to the result of the sequence
Viewpoint expression pattern collection.
In one embodiment of the invention, described to include: according to the candidate viewpoint set of patterns of the content of comment building
The comment is segmented, and carries out part-of-speech tagging；Search the noun of adjective left or right side；The noun found and institute
It states adjective and forms a candidate viewpoint expression pattern.
In one embodiment of the invention, described be ranked up to multiple candidate viewpoint expression patterns includes: according to more
The conjunction of a candidate's viewpoint expression pattern is ranked up multiple candidate viewpoints；The conjunction is the institute in the comment
State the Chinese character combination between noun and the adjective.
In one embodiment of the invention, described be ranked up to multiple candidate viewpoint expression patterns includes: transfer square
Battle array × viewpoint expression pattern vector=conjunction scores vector, transposed matrix × conjunction scores vector=update of transfer matrix
Viewpoint expression pattern vector afterwards；Iteration is updated to what viewpoint expression pattern vector sum the N+1 times that n-th is calculated obtained
When difference between viewpoint expression pattern vector is less than preset threshold, stops calculating, obtain the result of the sequence；Wherein, turn
Moving matrix indicates that the number that each conjunction occurs in different expression patterns, N are positive integer.
In the second aspect of embodiments of the present invention, a kind of comment and analysis device is provided, comprising: acquiring unit,
For obtaining at least one comment to be analyzed；Determination unit, it is described to be analyzed for being determined according to reference viewpoint expression pattern collection
The viewpoint of comment and determined according to preset viewpoint polarity standard the comment to be analyzed viewpoint polarity；Wherein, described
With reference to the set that viewpoint expression pattern collection is according to multiple reference viewpoint expression patterns commented on and be trained.
In the third aspect of embodiments of the present invention, a kind of electronic equipment, including processor and memory are provided；
Wherein, the memory store code；The processor executes the code, for executing comment and analysis described in first aspect
Method.
In the fourth aspect of embodiments of the present invention, a kind of computer-readable storage medium for storing program is provided
Matter, described program includes instruction, when described instruction is computer-executed, the computer is made to execute comment described in first aspect
Analysis method.
Embodiment according to the present invention, comment and analysis method, apparatus, equipment and storage medium provided by the invention have
Following advantage: obtaining determining the viewpoint in comment to be analyzed using reference viewpoint with reference to viewpoint collection according to multiple consumer reviews,
And the polarity of viewpoint is determined using viewpoint polarity standard, so as to obtain consumer point expressed by the comment to be analyzed,
It is referred to for goods providers.
Detailed description of the invention
It, below will be to embodiment party in order to illustrate more clearly of embodiments of the present invention or technical solution in the prior art
Formula or attached drawing needed to be used in the description of the prior art are briefly described.It should be evident that the accompanying drawings in the following description is only
It is merely exemplary, it for those of ordinary skill in the art, without creative efforts, can also basis
The attached drawing of offer, which is extended, obtains other implementation attached drawings.
Fig. 1 is a kind of flow chart for comment and analysis method that one embodiment of the invention provides；
Fig. 2 is a kind of structural schematic diagram for comment and analysis device that another embodiment of the present invention provides；
Fig. 3 is the structural schematic diagram for a kind of electronic equipment that another embodiment of the present invention provides.
In figure: 21. acquiring units, 22. determination units, 31. processors, 32. memories.
Specific embodiment
Embodiments of the present invention are illustrated by particular specific embodiment below, those skilled in the art can be by this explanation
Content disclosed by book is understood other advantages and efficacy of the present invention easily, it is clear that described embodiment is the present invention one
Section Example, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art are not doing
Every other embodiment obtained under the premise of creative work out, shall fall within the protection scope of the present invention.
Comment and analysis method provided in an embodiment of the present invention is a kind of opining mining scheme.Opining mining is a kind of natural language
Say processing task, be the excavations of emotion informations such as the theme to text information, opinion holder, subjective and objective property, mood attitude and
Analysis, and then identify that the emotion of subjective texts tends to.Opining mining, in comment on commodity, can analyze in consumer
Evaluation of the consumer to each feature of the commodity out, for example, for " receiving after wrist-watch till now, I generally or compares
Relatively satisfactory, common sensation should summarize very much, and an examination color is all relatively good, and second is that function is stronger
Greatly, third is that scalability is relatively good, mainly its software and backstage this part, the 4th, be exactly it appearance it is good.As for
The power consumption problem that everybody is concerned about, this, wait with after after a period of time I do a detailed experiment, comment later giving
Valence.In purchasing process, customer service attitude is still very good.Purchasing process customer service is relatively good, as after-sale service, later period
It is commented on again after experience." this comment it can be concluded that viewpoint are as follows: it is powerful, stand-by time is long, customer service attitude is pretty good.
Next, comment and analysis method provided in an embodiment of the present invention is specifically introduced.
Embodiment 1
A kind of comment and analysis method is present embodiments provided, executing subject can have calculating, processing capacity to be any
System, unit, platform or server.
As shown in Figure 1, the comment and analysis method includes the following steps.
Step 11 obtains at least one comment to be analyzed.
An available comment to be analyzed, also available a plurality of comment to be analyzed.The a plurality of comment to be analyzed obtained
For the consumer reviews for same commodity.
It can use the consumer reviews for a certain commodity on crawler technology acquisition electric business platform.It is with day cat store
Example, day cat store are that each commodity all generates commodity sign, which corresponds to the commodity.According to commodity sign, benefit
The comment that the corresponding commodity of the commodity sign are directed on day cat store is obtained with crawler technology.
Step 12 determines the viewpoint of the comment to be analyzed according to reference viewpoint expression pattern collection and according to preset sight
Point-polarity standard determines the polarity of the viewpoint of the comment to be analyzed；Wherein, described more for basis with reference to viewpoint expression pattern collection
A set for commenting on the reference viewpoint expression pattern being trained.
In the present embodiment, viewpoint refers to the Chinese character group of view that nouns and adjectives is formed, for indicating consumer
It closes, such as " customer service attitude " and " good " composition one are for indicating the viewpoint of " customer service attitude is good ".But and not all noun
It may serve to one viewpoint of composition with adjectival combination, for example, " customer service attitude " and " cheap " cannot then form a sight
Point.
In the present embodiment, viewpoint expression pattern refers to the intercombination between nouns and adjectives, such as (" customer service "+
" good ") it is a viewpoint expression pattern, (" customer service "+" cheap ") is a viewpoint expression pattern.For (" customer service "+" good ") this
One viewpoint expression pattern can form the viewpoint of one " customer service is good ", and express mould for (" customer service "+" cheap ") this viewpoint
Formula cannot then form a viewpoint.Whether a viewpoint can be formed for specific noun and specific adjective, needed with ginseng
Viewpoint expression pattern is examined to judge.Refer to reference to viewpoint expression pattern may be considered be capable of forming a viewpoint, correctly
The combination of nouns and adjectives, such as (" customer service "+" good ").
In one example, described to be obtained with reference to viewpoint expression pattern collection by following steps: to obtain a plurality of comment；According to
The content of the comment constructs candidate viewpoint Ji Biaodamoshiji, and candidate's viewpoint expression pattern collection includes multiple candidate viewpoints
Expression pattern；Multiple candidate viewpoint expression patterns are ranked up；It is obtained according to the result of the sequence described with reference to viewpoint table
Expression patterns collection.
The method introduced in 1 that can take steps obtains a plurality of comment, and details are not described herein again.It, can after obtaining comment
First to remove the noise character in comment.Noise character can be expression, meaningless punctuation mark etc..For example, " screen is big,
Customer service attitude is pretty good.", therein, extra fullstop is meaningless punctuation mark.
It is described to be specifically included according to the candidate viewpoint expression pattern collection of the content of comment building: the comment is divided
Word, and carry out part-of-speech tagging；Search the noun of adjective left or right side；The noun found and the adjective form one
Candidate viewpoint expression pattern.
Specifically can be used Stanford University exploitation participle tool (Stanford Word Segmenter, link:
Https: //nlp.stanford.edu/software/segmenter.shtml) it is segmented.After participle, part of speech mark is used
Note tool carries out part-of-speech tagging, and the part-of-speech tagging tool that Stanford University's exploitation specifically can be used carries out part-of-speech tagging
(Stanford Log-linear Part-Of-Speech Tagger, link: https: //nlp.stanford.edu/
software/tagger.shtml)。
After having carried out participle and part-of-speech tagging, the adjective that sentiment dictionary includes, and benefit are filtered out using sentiment dictionary
The noun near adjective is searched by the way of windowing with the adjective filtered out.Specifically, an adjective left side can be searched
Noun in the preset characters number of side can also search the noun on the right side of adjective in preset characters number.Preset characters number
It can freely be arranged, such as can be 5 characters.The adjective and the name contamination found are a candidate viewpoint expression
Mode.
In the present embodiment, sentiment dictionary can use NTUSD Taiwan Univ. Chinese sentiment dictionary.
All candidate viewpoint expression patterns in a comment are obtained according to above scheme.Such as one is commented on
It " receives after wrist-watch till now, I am generally still satisfied, common sensation, should summarize, function very much
It is more powerful, it is standby very long.In purchasing process, customer service is still very good.Purchasing process customer service is relatively good, as after sale
Service is commented on again after later period experience." at least available (" function "+" powerful "), (" standby "+" length "), (" customer service "+
" good "), (" customer service "+" good ") four candidate viewpoint expression formulas.
In one example, described be ranked up to multiple candidate viewpoint expression patterns includes: according to multiple candidate viewpoints
The conjunction of expression pattern is ranked up multiple candidate viewpoints；The conjunction is the noun and institute in the comment
State the Chinese character combination between adjective.
In a comment, the Chinese character group between nouns and adjectives is combined into conjunction.Nouns and adjectives constitutes
Relationship description between conjunction and candidate viewpoint expression formula can be in this example that modification is closed by candidate viewpoint expression formula
System.For example, entitled " customer service " therein, adjective is " good ", and conjunction is for comment " customer service is still very good "
" still very "." still very " modifies (" customer service "+" good ".)
In this example, in the range of multiple comments, if a candidate viewpoint expression formula is by many different connections
Word modification, then illustrate that candidate's viewpoint expression formula is more accurate, score is relatively high, correspondingly, its sequence is also earlier.One
It is a candidate viewpoint expression formula refer to identical noun and it is identical describe word combination, without consider its which or which comment
Middle appearance, for example, the comment for " customer service is still very good ", obtained candidate viewpoint expression formula (" customer service "+" good ")；
Comment for " customer service is very good " is similarly obtained candidate viewpoint expression formula (" customer service "+" good "), i.e., in two comments
Obtained identical candidate viewpoint expression formula is the same candidate viewpoint expression formula；And for modification (" customer service "+" good ")
There are two conjunctions, respectively " still very ", " very ".
In one example, described be ranked up to multiple candidate viewpoint expression formulas includes: that transfer matrix × viewpoint is expressed
Pattern vector=conjunction scores vector, transposed matrix × conjunction scores vector=updated viewpoint expression of transfer matrix
Pattern vector；Iteration is updated to the viewpoint expression pattern vector sum the N+1 times obtained viewpoint expression pattern that n-th is calculated
When difference between vector is less than preset threshold, stops calculating, obtain the result of the sequence；Wherein, transfer matrix indicates every
The number that a conjunction occurs in different expression patterns, N are positive integer.
It specifically can be used and multiple candidate viewpoint expression patterns be ranked up using page rank (PageRank) algorithm.
A kind of Algorithms for Page Ranking that page rank algorithm is proposed by Google, but the thought of its behind is also applied for emotional expression combination
Sequence.
In this example, noun aspect, adjective opinion, candidate viewpoint expression pattern be (aspect,
Opinion) pair (can abbreviation pair), conjunction linkwords.
In this example, multiple candidate viewpoint expression patterns are ranked up using page rank (PageRank) algorithm
Thinking are as follows:
(1) if (aspect, an opinion) pair is modified by many different linkwords, then just illustrating this
A (aspect, opinion) pair is more correct, the score of this pair also can be relatively high.
(2) if there is some linkwords among very high (aspect, the opinion) pair of score, then this
Therefore the score of linkwords can also improve accordingly.
Specific calculating process is as follows:
Defining A first is linkwords scores vector, and B is transfer matrix, and C is pair scores vector.Original state A1,
B,C1.Wherein A1, C1 are the vectors for being initialized as all 1, the shape of transfer matrix B be linkswords* (aspect,
Opinoin) pair is meant that the number that each linkwords occurs in different pair.For example, the pair obtained has
Pair 1, pair 2, pair 3, linkwords have linkwords 1, linkwords 2, linkwords 3.Wherein,
Linkwords 1 occurs 3 times in pair 1, occurs in pair 23 times, occurs 0 time in pair 3, linkwords 2
Occur in pair 12 times, occur in pair 21 time, occur 1 time in pair 3, linkwords 3 goes out in pair 1
Existing 1 time, occurs 0 time in pair 2, occur in pair 31 time, then each linkwords occurs secondary in different pair
Number constitutes transfer matrix B.
Calculating process is as follows:
1.A2=B*C1 (updates A vector),
2.C2=BT* A2 (updating C vector),
3.A3=B*C2,
4.C3=BT*A3,
……..。
Such iteration updates, and until CN+1 is close with CN (N is positive integer), (i.e. differing between CN+1 and CN is less than pre-
If preset threshold is to carry out pre-set threshold value according to corresponding situation when threshold value) show to have restrained and can stop iteration.
C vector is exactly the score for representing each pair at this time, and the correct candidate viewpoint expression mould for coming front can be obtained by sequence
Formula is determined as with reference to viewpoint expression pattern.
It is multiple to constitute with reference to viewpoint expression pattern with reference to viewpoint expression pattern collection.
The candidate viewpoint expression pattern that the preceding M% in candidate viewpoint expression pattern sequence can be chosen is used as with reference to viewpoint table
Expression patterns.M > 0.
M value can be carried out according to specific needs.
In one example, 30 M.
When analyzing comment to be analyzed, it can be removed noise character, participle, part-of-speech tagging, and seen
Point expression pattern, concrete mode can refer to described above, and details are not described herein again.
Obtained viewpoint expression pattern is matched in reference viewpoint expression pattern concentration, if matched, then it is assumed that should
Viewpoint expression pattern is correct, and nouns and adjectives constitutes a viewpoint.If unmatched, then it is assumed that the viewpoint expression pattern
Incorrect, nouns and adjectives constitutes a viewpoint.
It should be noted that comment to be analyzed for one, available multiple viewpoint expression patterns.Multiple viewpoint expression
Independent concentrate in reference viewpoint expression pattern of each viewpoint expression pattern in mode is matched, to determine whether it is correct table
It reaches, whether nouns and adjectives constitutes a viewpoint.
The polarity of viewpoint refers to that the viewpoint is positive or negative sense, such as " customer service is pretty good " this viewpoint indicates
Meaning front, polarity are forward direction；The meaning that " charging time is long " this viewpoint indicates is negative, then its polarity is negative sense.
Also, for different commodity, the same possible polarity of viewpoint is different, for example, for viewpoint " there are many moisture ",
If the commodity being directed to are fresh fruit, the polarity of the viewpoint is forward direction；If the commodity being directed to are dried food and nuts, the viewpoint
Polarity is negative sense.
The polarity of specific various viewpoints is negative, and the polarity of which kind of viewpoint is positive, and needs according to preset viewpoint polarity standard
Judged.Different viewpoint polarity standards can be preset according to different commodity.Specifically, certain commodity can be directed to
Viewpoint polarity table is preset, wherein having recorded with reference to viewpoint expression pattern viewpoint polarity corresponding with its.
In one example, the method also includes: the viewpoint of the comment to be analyzed is mapped on default label；System
The polarity for counting the viewpoint of all comments to be analyzed for being mapped to the default label, obtains statistical result.
The supplier of commodity can preset label, and label can be customized, or pre-set
It is selected in candidate tally set.
Specially the noun in viewpoint is mapped on default label.By taking label is " continuation of the journey " as an example, from a plurality of to be analyzed
Determine that obtained viewpoint has " electricity foot ", " electricity is small ", " customer service is pretty good " in comment." electricity " may map to " continuation of the journey " this
On label.
It can specifically be calculated and be mapped by term vector.Specifically, language calculate and be needed first in voice
Letter symbol switchs to number.Term vector is exactly a kind of semantic set of number for describing some word and being contained.In the present embodiment
The tool word2vec that the term vector of word is provided by Google be calculated (address word2vec: https: //
code.google.com/archive/p/word2vec/).Through calculating, the vector of " electricity " this word may be [0.5,
0.6,0.4], the vector of " continuation of the journey " this word may be [0.4,0.5,0.5], then " electricity " and the COS distance of " continuation of the journey " are just
It is 0.9819, by calculating, the term vector of " customer service " is [0.1, -0.9,0.2], then the COS distance of " customer service " and " continuation of the journey "
It is exactly -0.4114.The codomain of cosine is [- 1,1], -1 be exactly represent two words the meaning it is altogether irrelevant, 1 is exactly to represent two
The semanteme of word is identical, therefore, " electricity " can be mapped in " continuation of the journey " according to term vector.
After noun in viewpoint is mapped on label, forward direction/negative sense accounting of the viewpoint below each label is counted.Example
Such as, the viewpoint of " electricity foot " have 9, the viewpoint of " electricity is small " have 1, for " continuation of the journey ", " electricity foot " is forward direction, " electric
Measure small " for negative sense, then the positive viewpoint counted is 9, and negative sense viewpoint is 1.
In one example, the default label is the label of user preset；The method also includes: the statistics is tied
Fruit is sent to the user.
The present embodiment can provide third party's comment and analysis service.User can preset label.What is specified according to user
After the comment and analysis of commodity, obtained statistical result can be sent by email address, the instant messaging account number etc. that user reserves
To user.
Comment and analysis method provided in this embodiment, which has the advantages that, to be obtained according to multiple consumer reviews with reference to viewpoint
Collection, determines the viewpoint in comment to be analyzed using reference viewpoint, and the polarity of viewpoint is determined using viewpoint polarity standard, so as to
To obtain consumer point expressed by the comment to be analyzed, referred to for goods providers.
Embodiment 2
A kind of comment and analysis device is present embodiments provided, as shown in Figure 2, comprising:
Acquiring unit 21, for obtaining at least one comment to be analyzed；
Determination unit 22, for determined according to reference viewpoint expression pattern collection the comment to be analyzed viewpoint and according to
Preset viewpoint polarity standard determines the polarity of the viewpoint of the comment to be analyzed；Wherein, described to refer to viewpoint expression pattern collection
For according to multiple set for commenting on the reference viewpoint expression pattern being trained.
The content that the function of each functional unit of comment and analysis device provided in this embodiment can be recorded with reference implementation example 1 is real
Existing, details are not described herein again.
It has the advantages that the present embodiment provides comment and analysis device and is obtained according to multiple consumer reviews with reference to viewpoint
Collection, determines the viewpoint in comment to be analyzed using reference viewpoint, and the polarity of viewpoint is determined using viewpoint polarity standard, so as to
To obtain consumer point expressed by the comment to be analyzed, referred to for goods providers.
Embodiment 3
A kind of electronic equipment is present embodiments provided, as shown in figure 3, including processor 31 and memory 32；Wherein,
32 store code of memory；
The processor 31 executes the code, for executing comment and analysis method described in embodiment 1.
It has the advantages that the present embodiment provides electronic equipment and is obtained according to multiple consumer reviews with reference to viewpoint collection, benefit
The viewpoint in comment to be analyzed is determined with reference viewpoint, and the polarity of viewpoint is determined using viewpoint polarity standard, so as to
To consumer point expressed by the comment to be analyzed, referred to for goods providers.
Embodiment 4
A kind of computer readable storage medium for storing program is present embodiments provided, described program includes instruction, described
When instruction is computer-executed, the computer is made to execute comment and analysis method described in embodiment 1.
In an embodiment of the present invention, processor can be a kind of IC chip, the processing capacity with signal.Place
Reason device can be general processor, digital signal processor (Digital Signal Processor, abbreviation DSP), dedicated collection
At circuit (Application Specific Integrated Circuit, abbreviation ASIC), field programmable gate array
(Field Programmable Gate Array, abbreviation FPGA) either other programmable logic device, discrete gate or crystal
Pipe logical device, discrete hardware components.
It may be implemented or execute disclosed each method, step and the logic diagram in the embodiment of the present invention.General procedure
Device can be microprocessor or the processor is also possible to any conventional processor etc..In conjunction with disclosed in the embodiment of the present invention
Method the step of can be embodied directly in hardware decoding processor and execute completion, or with hardware in decoding processor and soft
Part block combiner executes completion.Software module can be located at random access memory, and flash memory, read-only memory may be programmed read-only storage
In the storage medium of this fields such as device or electrically erasable programmable memory, register maturation.Processor reads storage medium
In information, in conjunction with its hardware complete the above method the step of.
Storage medium can be memory, such as can be volatile memory or nonvolatile memory, or may include
Both volatile and non-volatile memories.
Wherein, nonvolatile memory can be read-only memory (Read-Only Memory, abbreviation ROM), may be programmed
Read-only memory (Programmable ROM, abbreviation PROM), Erasable Programmable Read Only Memory EPROM (Erasable PROM, letter
Claim EPROM), electrically erasable programmable read-only memory (Electrically EPROM, abbreviation EEPROM) or flash memory.
Volatile memory can be random access memory (Random Access Memory, abbreviation RAM), be used as
External Cache.By exemplary but be not restricted explanation, the RAM of many forms is available, such as static random-access is deposited
Reservoir (Static RAM, abbreviation SRAM), dynamic random access memory (Dynamic RAM, abbreviation DRAM), synchronous dynamic with
Machine accesses memory (Synchronous DRAM, abbreviation SDRAM), double data speed synchronous dynamic RAM
(Double Data RateSDRAM, abbreviation DDRSDRAM), enhanced Synchronous Dynamic Random Access Memory (Enhanced
SDRAM, abbreviation ESDRAM), synchronized links dynamic random access memory (Synchlink DRAM, abbreviation SLDRAM) and directly
Rambus random access memory (DirectRambus RAM, abbreviation DRRAM).
The storage medium of description of the embodiment of the present invention is intended to include but is not limited to depositing for these and any other suitable type
Reservoir.
Those skilled in the art are it will be appreciated that in said one or multiple examples, function described in the invention
It can be realized with hardware with combination of software.When application software, corresponding function can be stored in computer-readable medium
In or as on computer-readable medium one or more instructions or code transmitted.Computer-readable medium includes meter
Calculation machine storage medium and communication media, wherein communication media includes convenient for transmitting computer journey from a place to another place
Any medium of sequence.Storage medium can be any usable medium that general or specialized computer can access.
Above-described specific embodiment has carried out further the purpose of the present invention, technical scheme and beneficial effects
It is described in detail, it should be understood that being not intended to limit the present invention the foregoing is merely a specific embodiment of the invention
Protection scope, all any modification, equivalent substitution, improvement and etc. on the basis of technical solution of the present invention, done should all
Including within protection scope of the present invention.
Claims (10)
1. a kind of comment and analysis method characterized by comprising
Obtain at least one comment to be analyzed；
The viewpoint of the comment to be analyzed is determined according to reference viewpoint expression pattern collection and according to preset viewpoint polarity standard
Determine the polarity of the viewpoint of the comment to be analyzed；Wherein, the reference viewpoint expression pattern collection is to be carried out according to multiple comments
The set for the reference viewpoint expression pattern that training obtains.
2. comment and analysis method according to claim 1, which is characterized in that the method also includes:
The viewpoint of the comment to be analyzed is mapped on default label；
The polarity for counting the viewpoint of all comments to be analyzed for being mapped to the default label, obtains statistical result.
3. comment and analysis method according to claim 2, which is characterized in that the default label is the mark of user preset
Label；The method also includes: the statistical result is sent to the user.
4. comment and analysis method according to claim 1, which is characterized in that it is described with reference to viewpoint expression pattern collection by with
Lower step obtains:
Obtain a plurality of comment；
Candidate viewpoint expression pattern collection is constructed according to the content of the comment, candidate's viewpoint set of patterns includes multiple candidate sights
Point expression pattern；
Multiple candidate viewpoint expression patterns are ranked up；
It is obtained according to the result of the sequence described with reference to viewpoint expression pattern collection.
5. comment and analysis method according to claim 4, which is characterized in that described constructed according to the content of the comment is waited
The viewpoint set of patterns is selected to include:
The comment is segmented, and carries out part-of-speech tagging；
Search the noun of adjective left or right side；
The noun found and the adjective form a candidate viewpoint expression pattern.
6. comment and analysis method according to claim 5, which is characterized in that it is described to multiple candidate viewpoint expression patterns into
Row sorts
Multiple candidate viewpoints are ranked up according to the conjunction of multiple candidate viewpoint expression patterns；The conjunction is described
Chinese character combination in comment, between the noun and the adjective.
7. comment and analysis method according to claim 6, which is characterized in that it is described to multiple candidate viewpoint expression patterns into
Row sorts
Transfer matrix × viewpoint expression pattern vector=conjunction scores vector, transposed matrix × conjunction score of transfer matrix
Vector=updated viewpoint expression pattern vector；
Iteration is updated to the viewpoint expression pattern vector that viewpoint expression pattern vector sum the N+1 times that n-th is calculated obtains
Between difference be less than preset threshold when, stop calculate, obtain the result of the sequence；
Wherein, transfer matrix indicates that the number that each conjunction occurs in different expression patterns, N are positive integer.
8. a kind of comment and analysis device characterized by comprising
Acquiring unit, for obtaining at least one comment to be analyzed；
Determination unit, for determining the viewpoint of the comment to be analyzed according to reference viewpoint expression pattern collection and according to preset
Viewpoint polarity standard determines the polarity of the viewpoint of the comment to be analyzed；Wherein, described with reference to according to viewpoint expression pattern collection
Multiple set for commenting on the reference viewpoint expression pattern being trained.
9. a kind of electronic equipment, which is characterized in that including processor and memory；Wherein,
The memory store code；
The processor executes the code, requires the described in any item comment and analysis methods of 1-7 for perform claim.
10. a kind of computer readable storage medium for storing program, which is characterized in that described program includes instruction, described instruction
When being computer-executed, the computer perform claim is made to require the described in any item comment and analysis methods of 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811585336.3A CN109657248A (en) | 2018-12-24 | 2018-12-24 | A kind of comment and analysis method, apparatus, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811585336.3A CN109657248A (en) | 2018-12-24 | 2018-12-24 | A kind of comment and analysis method, apparatus, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109657248A true CN109657248A (en) | 2019-04-19 |
Family
ID=66116575
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811585336.3A Pending CN109657248A (en) | 2018-12-24 | 2018-12-24 | A kind of comment and analysis method, apparatus, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109657248A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113220823A (en) * | 2020-01-21 | 2021-08-06 | 北京中科闻歌科技股份有限公司 | Sentiment, topic and viewpoint analysis method for social media public language |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106776574A (en) * | 2016-12-28 | 2017-05-31 | Tcl集团股份有限公司 | User comment text method for digging and device |
| CN107862343A (en) * | 2017-11-28 | 2018-03-30 | 南京理工大学 | The rule-based and comment on commodity property level sensibility classification method of neutral net |
| CN108363725A (en) * | 2018-01-08 | 2018-08-03 | 浙江大学 | A kind of method of the extraction of user comment viewpoint and the generation of viewpoint label |
| US20180260860A1 (en) * | 2015-09-23 | 2018-09-13 | Giridhari Devanathan | A computer-implemented method and system for analyzing and evaluating user reviews |
-
2018
- 2018-12-24 CN CN201811585336.3A patent/CN109657248A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180260860A1 (en) * | 2015-09-23 | 2018-09-13 | Giridhari Devanathan | A computer-implemented method and system for analyzing and evaluating user reviews |
| CN106776574A (en) * | 2016-12-28 | 2017-05-31 | Tcl集团股份有限公司 | User comment text method for digging and device |
| CN107862343A (en) * | 2017-11-28 | 2018-03-30 | 南京理工大学 | The rule-based and comment on commodity property level sensibility classification method of neutral net |
| CN108363725A (en) * | 2018-01-08 | 2018-08-03 | 浙江大学 | A kind of method of the extraction of user comment viewpoint and the generation of viewpoint label |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113220823A (en) * | 2020-01-21 | 2021-08-06 | 北京中科闻歌科技股份有限公司 | Sentiment, topic and viewpoint analysis method for social media public language |
| CN113220823B (en) * | 2020-01-21 | 2024-03-01 | 北京中科闻歌科技股份有限公司 | Method and device for analyzing emotion, topic and viewpoint of social media public language |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Swathi et al. | An optimal deep learning-based LSTM for stock price prediction using twitter sentiment analysis | |
| US20210117617A1 (en) | Methods and systems for summarization of multiple documents using a machine learning approach | |
| CN103678564B (en) | Internet product research system based on data mining | |
| CN107273861A (en) | A kind of subjective question marking methods of marking, device and terminal device | |
| CN108509474A (en) | Search for the synonym extended method and device of information | |
| CN104199833B (en) | The clustering method and clustering apparatus of a kind of network search words | |
| CN105243129A (en) | Commodity property characteristic word clustering method | |
| US20120259617A1 (en) | System and method for slang sentiment classification for opinion mining | |
| CN105183833A (en) | User model based microblogging text recommendation method and recommendation apparatus thereof | |
| CN103473380B (en) | A kind of computer version sensibility classification method | |
| CN102043774A (en) | Machine translation evaluation device and method | |
| CN108319734A (en) | A kind of product feature structure tree method for auto constructing based on linear combiner | |
| CN108763321A (en) | A kind of related entities recommendation method based on extensive related entities network | |
| CN109800307A (en) | Analysis method, device, computer equipment and the storage medium of product evaluation | |
| CN104731774B (en) | Towards the personalized interpretation method and device of general machine translation engine | |
| CN103123633A (en) | Generation method of evaluation parameters and information searching method based on evaluation parameters | |
| CN105843796A (en) | Microblog emotional tendency analysis method and device | |
| CN109325146A (en) | A kind of video recommendation method, device, storage medium and server | |
| CN110413961A (en) | The method, apparatus and computer equipment of text scoring are carried out based on disaggregated model | |
| Gu et al. | Service package recommendation for mashup creation via mashup textual description mining | |
| CN110134845A (en) | Project public sentiment monitoring method, device, computer equipment and storage medium | |
| CN109255012A (en) | A kind of machine reads the implementation method and device of understanding | |
| CN107193892A (en) | A kind of document subject matter determines method and device | |
| CN107918778A (en) | A kind of information matching method and relevant apparatus | |
| CN110110332A (en) | Text snippet generation method and equipment |
Legal Events
| 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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190419 |