CN109977396A - Emotion identification method, device, computer equipment and the computer storage medium of corpus participle - Google Patents
Emotion identification method, device, computer equipment and the computer storage medium of corpus participle Download PDFInfo
- Publication number
- CN109977396A CN109977396A CN201910119320.1A CN201910119320A CN109977396A CN 109977396 A CN109977396 A CN 109977396A CN 201910119320 A CN201910119320 A CN 201910119320A CN 109977396 A CN109977396 A CN 109977396A
- Authority
- CN
- China
- Prior art keywords
- corpus
- field
- emotion
- words
- participle
- 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
- 230000008451 emotion Effects 0.000 title claims abstract description 393
- 238000000034 method Methods 0.000 title claims abstract description 43
- 230000008909 emotion recognition Effects 0.000 claims abstract description 24
- 230000011218 segmentation Effects 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012216 screening Methods 0.000 claims description 10
- 230000007935 neutral effect Effects 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 238000004891 communication Methods 0.000 description 4
- 230000002045 lasting effect Effects 0.000 description 4
- 239000000463 material Substances 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 241000233805 Phoenix Species 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000013499 data model Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000007726 management method 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
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/242—Dictionaries
-
- 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
Abstract
This application discloses emotion identification method, device, computer equipment and the computer storage mediums of a kind of corpus participle, are related to emotion recognition technical field, the accuracy rate of emotion recognition in text can be improved.The described method includes: the corpus of each data source segments in acquisition field；The significantly default emotion set of words of Sentiment orientation in field is filtered out from the corpus of each data source participle, and the default emotion set of words is merged with the emotion set of words in sentiment dictionary, obtains the emotion set of words in field；Utilize the semantic dependency between the specific emotion word of Sentiment orientation in the corpus participle between point in mutual information calculating field in addition to default emotion set of words and the emotion set of words in the field；The probability on clear Sentiment orientation is segmented according to the corpus in the field in addition to default emotion set of words, determines the Sentiment orientation of the corpus participle in field in addition to default emotion set of words.
Description
Technical field
The present invention relates to emotion recognition technical fields, emotion identification method, device, calculating more particularly, to corpus participle
Machine equipment and computer storage medium.
Background technique
With the continuous development of Internet technology, user is exchanged by network, and that issues in the network platform is a large amount of
Message and comment, by analyze these message and comment in hide Sentiment orientation, reference can be provided for other consumers,
Or businessman provides information, to improve service, increases user satisfaction.
Sentiment orientation, which can be understood as user, which expresses the attitude that itself viewpoint is held to certain object, is support, opposes, neutral,
The positive emotion that is often referred to, negative emotion, neutral emotion.Such as " praise " and " praise " is all commendatory term, expresses positive feelings
Sense, and " dirty " and " ugliness " is all derogatory term, expresses negative emotion.Currently, the analysis of Sentiment orientation mainly passes through in advance
The sentiment dictionary of building searches the emotion word in text, and identifies the Sentiment orientation of user according to the emotion word found,
The Sentiment orientation in text can be analyzed by sentiment dictionary, it is long-pending for such as analyzing the emotion of microblogging the preceding paragraph comment expression
Pole or passiveness.
However, during building sentiment dictionary, since the basic emotion word in sentiment dictionary is very limited, and emotion
The text of identification is more complicated and variation is cracking, if carrying out emotion recognition using emotion word basic in sentiment dictionary,
The emotion word in text can not be recognized accurately, cause sentiment analysis inaccurate.
Summary of the invention
In view of this, the present invention provides emotion identification method, device, computer equipment and the calculating of a kind of corpus participle
Machine storage medium, main purpose are to solve the problems, such as sentiment analysis inaccuracy in presently relevant technical text.
According to the present invention on one side, a kind of emotion identification method of corpus participle is provided, this method comprises:
The corpus participle of each data source in acquisition field；
The significantly default emotion set of words of Sentiment orientation in field is filtered out from the corpus of each data source participle,
And merge the default emotion set of words with the emotion set of words in sentiment dictionary, obtain the emotion word set in field
It closes, record has the specific emotion word of Sentiment orientation in field in emotion set of words in the field；
Using in the corpus participle and the field between point in mutual information calculating field in addition to default emotion set of words
Semantic dependency in emotion set of words between the specific emotion word of Sentiment orientation, obtain in field except default emotion set of words it
Outer corpus segments the probability on clear Sentiment orientation；
The probability on clear Sentiment orientation is segmented according to the corpus in the field in addition to default emotion set of words, really
Determine the Sentiment orientation of the corpus participle in field in addition to default emotion set of words.
It is further, described that filter out Sentiment orientation in field significantly pre- from the corpus of each data source participle
If emotion set of words includes:
The corpus participle for scanning each data source one by one, by corpus participle and the emotion word in basic sentiment dictionary into
Row compares；
The inconsistent corpus participle of comparison result is filtered out from the corpus of data source each in field participle, obtains field
The unknown corpus participle of interior Sentiment orientation；
Sentiment orientation is filtered out from the unknown corpus participle of Sentiment orientation in the field by way of manually marking
Apparent corpus participle, obtains default emotion set of words.
It further, include the emotion word of positive Sentiment orientation and passive feelings in field in emotion set in the field
Feel the emotion word of tendency, the corpus participle and institute using between point in mutual information calculating field in addition to default emotion set of words
The semantic dependency in the emotion set of words in field between the specific emotion word of Sentiment orientation is stated, is obtained in field except default feelings
Feeling the probability that the corpus except set of words segments on clear Sentiment orientation includes:
Utilize the corpus participle and positive emotion tendency between point in mutual information calculating field in addition to default emotion set
Semantic dependency between emotion word obtains the corpus participle in field in addition to default emotion set in positive emotion tendency
Probability；
Utilize the corpus participle and Negative Affect tendency between point in mutual information calculating field in addition to default emotion set
Semantic dependency between emotion word obtains the corpus participle in field in addition to default emotion set in Negative Affect tendency
Probability.
Further, the corpus according in the field in addition to default emotion set of words is segmented inclines in clear emotion
Upward probability determines that the Sentiment orientation of the corpus participle in field in addition to default emotion set of words includes:
The corpus participle calculated separately in the field in addition to default emotion set of words is general in positive emotion tendency
Probability the sum of of the sum of the rate with the corpus participle in the field in addition to default emotion set of words in Negative Affect tendency；
By probability the sum of of the corpus participle in the field in addition to default emotion set of words in positive emotion tendency
Probability the sum of of the corpus participle in the field in addition to default emotion set of words in Negative Affect tendency is subtracted, according to difference
It is worth the Sentiment orientation of the corpus participle in the field of determination in addition to default emotion set of words.
Further, the corpus participle by the field in addition to default emotion set of words is inclined in positive emotion
On the sum of probability to subtract the participle of the corpus in the field in addition to default emotion set of words general in Negative Affect tendency
The sum of rate, the Sentiment orientation for determining the participle of the corpus in field in addition to default emotion set of words according to difference include:
If the difference is greater than zero, it is determined that the emotion of the corpus participle in field in addition to default emotion set of words is inclined
It is inclined to for positive emotion；
If the difference is less than zero, it is determined that the emotion of the corpus participle in field in addition to default emotion set of words is inclined
It is inclined to for Negative Affect；
If the difference is equal to zero, it is determined that the emotion of the corpus participle in field in addition to default emotion set of words is inclined
To for neutral Sentiment orientation.
Further, in the acquisition field before the corpus participle of each data source, the method also includes:
The corpus of each data source in multiple known arts is collected by big data platform；
It is trained the corpus of each data source in the multiple known art as sample data input network model,
Obtain corpus field identification model.
Further, the corpus participle of each data source includes: in the acquisition field
The corpus data of each data source of tera incognita is input to corpus field identification model, obtains each number
According to the corresponding field of the corpus in source；
Summarize the corpus for belonging to each data source in same area, obtains the corpus of each data source in field；
Word segmentation processing is carried out to the corpus of each data source in the field, obtains the corpus point of each data source in field
Word.
According to the present invention on the other hand, a kind of emotion recognition device of corpus participle is provided, described device includes:
Acquiring unit, the corpus for obtaining each data source in field segment；
Screening unit, it is significantly pre- for filtering out Sentiment orientation in field from the corpus of each data source participle
If emotion set of words, and the default emotion set of words is merged with the emotion set of words in sentiment dictionary, obtain field
Interior emotion set of words, record has the specific emotion word of Sentiment orientation in field in emotion set of words in the field；
Computing unit, for using between point in mutual information calculating field in addition to default emotion set of words corpus participle with
Semantic dependency in emotion set of words in the field between the specific emotion word of Sentiment orientation obtains in field except default
Corpus except emotion set of words segments the probability on clear Sentiment orientation；
Determination unit is inclined for being segmented according to the corpus in the field in addition to default emotion set of words in clear emotion
Upward probability determines the Sentiment orientation of the corpus participle in field in addition to default emotion set of words.
Further, the screening unit includes:
Comparison module, the corpus for scanning each data source one by one segment, by corpus participle and basic emotion word
Emotion word in allusion quotation is compared；
First screening module, it is inconsistent for filtering out comparison result from the corpus of data source each in field participle
Corpus participle obtains the unknown corpus participle of Sentiment orientation in field；
Second screening module, for the corpus participle that Sentiment orientation is unknown out of described field by way of manually marking
In filter out Sentiment orientation apparent corpus participle, obtain default emotion set of words.
It further, include the emotion word of positive Sentiment orientation and passive feelings in field in emotion set in the field
Feel the emotion word of tendency, the computing unit includes:
First computing module, for being segmented using the corpus between point in mutual information calculating field in addition to default emotion set
Semantic dependency between the emotion word of positive emotion tendency obtains the corpus participle in field in addition to default emotion set
Probability in positive emotion tendency；
Second computing module, for being segmented using the corpus between point in mutual information calculating field in addition to default emotion set
Semantic dependency between the emotion word of Negative Affect tendency obtains the corpus participle in field in addition to default emotion set
Probability in Negative Affect tendency.
Further, the determination unit includes:
Third computing module is segmented for calculating separately the corpus in the field in addition to default emotion set of words in product
The sum of probability on the Sentiment orientation of pole is segmented with the corpus in the field in addition to default emotion set of words to incline in Negative Affect
The sum of upward probability；
Determining module, for the corpus participle in the field in addition to default emotion set of words to be inclined in positive emotion
On the sum of probability to subtract the participle of the corpus in the field in addition to default emotion set of words general in Negative Affect tendency
The sum of rate determines the Sentiment orientation of the participle of the corpus in field in addition to default emotion set of words according to difference.
Further, the determining module, if being specifically used for the difference is greater than zero, it is determined that except default feelings in field
The Sentiment orientation of the corpus participle except set of words is felt for positive emotion tendency；
The determining module, if being specifically also used to the difference less than zero, it is determined that except default emotion word set in field
The Sentiment orientation of corpus participle except conjunction is Negative Affect tendency；
The determining module, if being specifically also used to the difference equal to zero, it is determined that except default emotion word set in field
The Sentiment orientation of corpus participle except conjunction is neutral Sentiment orientation.
Further, described device further include:
Collector unit passes through big data platform before the corpus participle of each data source in the acquisition field
Collect the corpus of each data source in multiple known arts；
Training unit, for inputting network for the corpus of each data source in the multiple known art as sample data
Model is trained, and obtains corpus field identification model.
Further, the acquiring unit includes:
Module is obtained, identifies mould for the corpus data of each data source of tera incognita to be input to the corpus field
Type obtains the corresponding field of corpus of each data source；
Summarizing module obtains each data source in field for summarizing the corpus for belonging to each data source in same area
Corpus；
Word segmentation module carries out word segmentation processing for the corpus to each data source in the field, obtains each in field
The corpus of data source segments.
Another aspect according to the present invention provides a kind of computer equipment, including memory and processor, the storage
Device is stored with computer program, and the processor realizes the emotion identification method of corpus participle when executing the computer program
Step.
Another aspect according to the present invention provides a kind of computer storage medium, is stored thereon with computer program, institute
The step of stating the emotion identification method that corpus participle is realized when computer program is executed by processor.
By above-mentioned technical proposal, the present invention provides the emotion identification method and device of a kind of corpus participle, by from obtaining
It takes and filters out the significantly default emotion set of words of Sentiment orientation in field in the corpus participle of each data source, and and sentiment dictionary
In emotion set of words merge, obtain the emotion set of words in field so that in field in emotion set of words record have more
Emotion word known to comprehensive Sentiment orientation, it is further interior in addition to default emotion set of words using mutual information calculating field between point
Corpus participle and field in emotion set of words in semantic dependency between the specific emotion word of Sentiment orientation, obtain field
The interior corpus in addition to default emotion set of words segments the probability on clear Sentiment orientation, so that it is determined that except default feelings in field
The Sentiment orientation for feeling the corpus participle except set of words can accurately identify the Sentiment orientation of more corpus participles in field.With
Corpus participle emotion recognition mode is carried out by the basic emotion word in sentiment dictionary in the prior art to compare, the present invention is implemented
Example carries out emotion recognition for the corpus participle in specific area, by between point mutual information in the way of calculate the feelings of corpus participle
Sense tendency, can effectively extract the emotion word of specific area, and accurately identify the Sentiment orientation of emotion word, and with corpus
The lasting expansion of participle, the new emotion vocabulary in field can be captured effectively.
Detailed description of the invention
By reading the following detailed description of the preferred embodiment, various other advantages and benefits are common for this field
Technical staff will become clear.The drawings are only for the purpose of illustrating a preferred embodiment, and is not considered as to the present invention
Limitation.And throughout the drawings, the same reference numbers will be used to refer to the same parts.In the accompanying drawings:
Fig. 1 shows a kind of emotion identification method flow diagram of corpus participle provided in an embodiment of the present invention；
Fig. 2 shows the emotion identification method flow diagrams of another corpus participle provided in an embodiment of the present invention；
Fig. 3 shows a kind of structural schematic diagram of the emotion recognition device of corpus participle provided in an embodiment of the present invention；
Fig. 4 shows the structural schematic diagram of the emotion recognition device of another corpus participle provided in an embodiment of the present invention.
Specific embodiment
Exemplary embodiments of the present disclosure are described in more detail below with reference to accompanying drawings.Although showing the disclosure in attached drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
It is limited.On the contrary, these embodiments are provided to facilitate a more thoroughly understanding of the present invention, and can be by the scope of the present disclosure
It is fully disclosed to those skilled in the art.
The embodiment of the invention provides a kind of emotion identification method of corpus participle, it may be implemented to improve emotion in text and know
The purpose of other accuracy rate, as shown in Figure 1, this method comprises:
101, the corpus of each data source segments in acquisition field.
Wherein, field can be financial field, legal field, machinery field etc., and each data source can be neck in field
Forum, website, platform database for being commonly used in domain etc., such as field are financial field, then each data in financial field
Source can for phoenix finance, Financial Times etc., corpus here mainly include the word recorded in data source in field, sentence,
Article etc..
For the embodiment of the present invention, it can specifically be segmented, can also be passed through by the corpus in data platform assembling sphere
The corpus participle in point mode acquisition field is buried, here without limiting.
102, the significantly default emotion word of Sentiment orientation in field is filtered out from the corpus of each data source participle
Set, and the default emotion set of words is merged with the emotion set of words in sentiment dictionary, obtain the emotion in field
Set of words.
Since the corpus participle of each data source contains the vocabulary of the various emotions in field, the feelings of some corpus participle
Sense tendency is obvious, and the Sentiment orientation of some corpus participle is indefinite, usually corpus participle can be divided into emotion in advance
It is inclined to that the emotion word of obvious, known Sentiment orientation, Sentiment orientation be indefinite and the emotion word etc. of unknown Sentiment orientation, and here
Default emotion set of words be the obvious emotion word of Sentiment orientation in corpus participle in field set, that is, Sentiment orientation is basic
The set of confirmable emotion word, it is known that the emotion word of Sentiment orientation is the emotion word belonged in existing sentiment dictionary, also
It is the emotion word recorded in existing sentiment dictionary, and the emotion word that Sentiment orientation is indefinite or Sentiment orientation is unknown is not belong to
Emotion word in existing sentiment dictionary, that is, the emotion word that do not recorded in existing sentiment dictionary.
For the embodiment of the present invention, sentiment dictionary collection is combined into the basic emotion word in existing sentiment dictionary, by that will preset
The emotion set of words that emotion set of words merges with sentiment dictionary in the field of building is able to record in more areas that Sentiment orientation is bright
Aobvious basic emotion vocabulary, these basic emotion vocabulary should be apparent and representative, inclines to improve to emotion
To the accuracy rate of unknown corpus participle emotion recognition.
103, it is segmented and the field using the corpus between point in mutual information calculating field in addition to default emotion set of words
Semantic dependency in interior emotion set of words between the specific emotion word of Sentiment orientation obtains in field except default emotion word set
Corpus except conjunction segments the probability on clear Sentiment orientation.
Wherein, mutual information is for measuring in corpus participle and the field in field in addition to default emotion set of words between point
Emotion set of words in correlation between the specific emotion word of Sentiment orientation, specific formula for calculation are as follows:
Wherein, P (word1 and word2) indicates the probability that two emotion word word1 and word2 occur jointly, P (word1)
Indicate that the probability that word1 individually occurs, P (word2) indicate the probability that word2 individually occurs.
Under normal conditions, different PMI (word1, word2) can show that between two emotion words word1 and word2
Correlation, if PMI (word1, word2) > 0, illustrate emotion word word1 to emotion word word2 be it is relevant, absolute value is bigger
Illustrate that the degree of correlation is higher, if PMI (word1, word2) < 0, illustrates it is to repel between emotion word word1 and emotion word word2
, and absolute value is bigger illustrates that repulsion degree is higher, if PMI (word1, word2)=0, illustrates emotion word word1 and emotion
Be between word word2 it is independent, it is uncorrelated also not repel.
For the embodiment of the present invention, word1 here is equivalent to the corpus in field in addition to default emotion set of words point
Word, is the indefinite corpus participle of Sentiment orientation, and word2 is equivalent to the specific feelings of Sentiment orientation in the emotion set of words in field
Feel word, can specifically choose the document comprising emotion word in field as referenced text, by statistics corpus participle in reference text
The frequency individually occurred in this calculates P (word1), by the statistics specific emotion word of Sentiment orientation in referenced text individually
The frequency of appearance calculates P (word2), through statistics corpus participle and Sentiment orientation specific emotion word in referenced text
The frequency occurred jointly calculates P (word1 and word2).Since the Sentiment orientation of the emotion word in field in emotion set is
Explicitly, so if the corpus participle in field in addition to default emotion set of words and emotion in the emotion set of words in field
The correlation being inclined between specific emotion word is higher, then illustrates that the corpus participle in field in addition to default emotion set of words exists
The probability specified on Sentiment orientation is higher, conversely, illustrating that the corpus participle in field in addition to default emotion set of words is defining
Probability on Sentiment orientation is lower.
In the embodiment of the present invention, it is introduced into the Sentiment orientation for calculating corpus participle, can be captured by mutual information between point
The emotion word being inclined to more different emotions, and then more accurately identify the Sentiment orientation of emotion word in text, improve text
The accuracy of sentiment analysis.
104, general on clear Sentiment orientation according to the corpus participle in the field in addition to default emotion set of words
Rate determines the Sentiment orientation of the corpus participle in field in addition to default emotion set of words.
For the embodiment of the present invention, specifying Sentiment orientation can be to be divided into actively tendency, Negative Affect tendency or nothing
Sentiment orientation can also specifically be divided for a certain Sentiment orientation certainly, for example, positive emotion tendency be divided into primary, in
Grade and advanced, the corresponding aggressiveness level of primary positive emotion is lower, and it is placed in the middle that intermediate positive emotion correspond to aggressiveness level, advanced positive
It is higher that emotion corresponds to aggressiveness level, and the embodiment of the present invention is without limiting.
Due to the Sentiment orientation of the emotion word of clear Sentiment orientation be equivalent to it is determining, except default emotion in field here
Corpus participle except set of words is equivalent to corpus participle to be identified, and positive emotion tendency is opposition with Negative Affect tendency
Sentiment orientation, by being segmented corpus to be identified participle with corpus to be identified true in the probability being determined as in positive emotion tendency
The difference being set between the probability value in Negative Affect tendency determines the Sentiment orientation of corpus participle to be identified according to difference.
Specifically, if difference is positive value, illustrate corpus participle to be identified in the emotion word for being determined as positive emotion
Probability on corresponding Sentiment orientation is high, further determines that the Sentiment orientation of corpus participle to be identified is actively to be inclined to, if poor
Value is negative value, then illustrating that corpus to be identified segments the probability on the corresponding Sentiment orientation of emotion word for being determined as Negative Affect
Height further determines that the Sentiment orientation of corpus participle to be identified for passive tendency, if difference is zero, illustrates language to be identified
Probability of the material participle on the corresponding Sentiment orientation of emotion word for being determined as ameleia tendency is high, further determines that corpus to be identified
The Sentiment orientation of participle is ameleia tendency.It can specifically be indicated with formula are as follows:
The present invention provides a kind of emotion identification method of corpus participle, by from the corpus participle for obtaining each data source
The significantly default emotion set of words of Sentiment orientation in field is filtered out, and is merged with the emotion set of words in sentiment dictionary,
The emotion set of words in field is obtained, so that record has emotion known to more fully Sentiment orientation in emotion set of words in field
Word further utilizes the corpus participle between point in mutual information calculating field in addition to default emotion set of words and the emotion in field
Semantic dependency in set of words between the specific emotion word of Sentiment orientation obtains in field in addition to default emotion set of words
Corpus segments the probability on clear Sentiment orientation, so that it is determined that the corpus in field in addition to default emotion set of words segmented
Sentiment orientation can accurately identify the Sentiment orientation of more corpus participles in field.With in the prior art pass through sentiment dictionary in
Basic emotion word compared to carry out corpus participle emotion recognition mode, the embodiment of the present invention is for the corpus point in specific area
Word carry out emotion recognition, by between point mutual information in the way of come calculate corpus participle Sentiment orientation, can effectively extract spy
Determine the emotion word in field, and accurately identify the Sentiment orientation of emotion word, and with the lasting expansion that corpus segments, in field
New emotion vocabulary can be captured effectively.
The embodiment of the invention provides the emotion identification method of another corpus participle, may be implemented to improve emotion in text
The purpose of the accuracy rate of identification, as shown in Figure 2, which comprises
201, the corpus of each data source in multiple known arts is collected by big data platform.
Wherein, each data source can be forum, the website etc. in known art, can be monitored by big data platform
Know daily record data, the tables of data etc. in field in each data source correspondence database, and collects each data source correspondence database
In daily record data, tables of data etc., to obtain the corpus of each data source in known art, which can be text, sentence
Son or article etc..
202, it is carried out the corpus of each data source in the multiple known art as sample data input network model
Training, obtains corpus field identification model.
Wherein, network model can be the convolutional neural networks model of multilayer, and known art here can be different
Field, such as machinery field, financial field etc..
It, specifically can be according to each data source since the corpus of data source each in known art is corresponding with different fields
The field of corpus being marked, such as data source is financial field, then the data collected from the data source are financial field, into
One step is trained the corpus for being marked with known art as sample data input network model, can construct each data source
Corpus and known art mapping relations.
Specifically the corpus of each data source inputs what network model was trained as sample data in using known art
In the process, corresponding by corpus data in the first layer structure extraction difference known art in the multilayered structure of network model
Local data's feature；Summarize corpus data in different known arts by second of structure in the multilayered structure of network model
Corresponding local data's feature obtains various dimensions local data feature；Pass through the third time in the multilayered structure of network model
Structure carries out dimension-reduction treatment to various dimensions local data feature, obtains the corresponding data characteristics of corpus data in different field；It is logical
The four-layer structure in network model in multilayered structure is crossed to divide the corresponding data characteristics of corpus data in different field
Class obtains the corpus for identifying each data source in different known arts.
Since corpus participle there may be different emotion or expression in different field, pass through existing feelings
Sense dictionary can not accurately analyze the emotion of corpus participle.For the embodiment of the present invention, collected by big data platform multiple known
The corpus of each data source in field can belong to the corpus in field, the corpus in each field for different field building
The corpus that each data source is collected in the field is recorded, to the sentiment analysis that corpus segments be navigated in different field, energy
Enough emotions for more thering is accurate identification corpus to segment state in the text.
203, the corpus data of each data source of tera incognita is input to corpus field identification model, obtained each
The corresponding field of the corpus of a data source.
For the embodiment of the present invention, the corpus data of each data source of tera incognita collects the corpus number for having every field
According to and identification model record in corpus field has mapping relations between the corpus and field of each data, by by tera incognita
The corpus of each data source be input in the identification model of corpus field and can identify the corresponding neck of the corpus of each data source
Domain, for example, A corpus belongs to financial field, B corpus belongs to communications field etc..
204, summarize the corpus for belonging to each data source in same area, obtain the corpus of each data source in field.
Due to feature or application scenarios that the corpus of different field has the field exclusive, belong to identical neck by summarizing
The corpus of each data source in domain, the corpus of each data source in available each field, to be directed to the spy of different field
Sign or application scenarios are analyzed and processed the corpus of each data source in each field.
205, word segmentation processing is carried out to the corpus of each data source in the field, obtains the language of each data source in field
Material participle.
It, specifically can be with for the embodiment of the present invention since the corpus of article or sentence type is difficult to carry out sentiment analysis
Word segmentation processing is carried out using corpus of the participle tool to data source each in field, obtains the corpus point of each data source in field
Word, and then each corpus is segmented and carries out sentiment analysis.
Here participle tool can be existing participle tool, can be realized by different segmentation methods, can be with needle
Different segmentation methods, such as rule-based segmentation methods, segmentation methods and base based on statistics are selected actual demand
In the segmentation methods of semantic understanding, the embodiment of the present invention is without limiting.
206, the corpus participle for scanning each data source one by one, by corpus participle and the emotion in basic sentiment dictionary
Word is compared.
Wherein, basic emotion vocabulary is the emotion word of known Sentiment orientation recorded in existing sentiment dictionary, by by language
Material participle is compared with the emotion word in basic sentiment dictionary, it can be determined that whether there is in the corpus participle of each data source
Determine the emotion word of Sentiment orientation, if the comparison results are consistent, then illustrates that corpus participle is the feelings of existing determining Sentiment orientation
Word is felt, for example, otherwise, illustrating that corpus participle is the uncertain emotion word of Sentiment orientation.
207, the inconsistent corpus participle of comparison result is filtered out from the corpus of data source each in field participle, is obtained
The corpus participle that Sentiment orientation is unknown in field.
For the embodiment of the present invention, incline due to there may be existing emotion in the corpus participle of data source each in field
To specific emotion word, and the specific emotion word of Sentiment orientation passes through number each out of field without carrying out emotion recognition here
According to the inconsistent corpus participle of comparison result is filtered out in the corpus participle in source, to obtain the unknown language of Sentiment orientation in field
Material participle.
208, emotion is filtered out from the unknown corpus participle of Sentiment orientation in the field by way of manually marking
It is inclined to apparent corpus participle, obtains default emotion set of words, and by the feelings in the default emotion set of words and sentiment dictionary
Sense set of words merges, and obtains the emotion set of words in field.
It is understood that the unknown corpus participle of Sentiment orientation is the emotion word do not recorded in existing sentiment dictionary,
And the corpus participle being commonly used in some fields may be the obvious emotion word of Sentiment orientation, and these corpus segment
Without carrying out emotion recognition, filtered out from the unknown corpus participle of Sentiment orientation in field by way of manually marking here
The apparent corpus participle of Sentiment orientation, obtains default emotion set of words.
The embodiment of the present invention is corresponded to, generally comprises the general emotion vocabulary of every field in sentiment dictionary, and does not include
Emotion vocabulary in field, significantly preset by dedicated Sentiment orientation in the field that will screen here emotion set of words with it is each
Emotion set of words in the general sentiment dictionary in field summarizes, and can obtain the emotion set of words in more fully field.
209, it is segmented using the corpus between point in mutual information calculating field in addition to default emotion set and is inclined with positive emotion
To emotion word between semantic dependency, obtain in field in addition to default emotion set corpus participle incline in positive emotion
Upward probability.
For the embodiment of the present invention, the corpus participle preset except emotion set of words is usually the indefinite feelings of Sentiment orientation
Feel word, needs to be judged by the specific emotion word of Sentiment orientation in the emotion set of words in field, and the emotion in field
The specific emotion word of Sentiment orientation includes the emotion word of positive emotion word and Negative Affect tendency in set of words, here with point
Between the corpus participle except set and the semantic dependency between positive emotion word, the semanteme phase are preset in mutual information calculating field
Closing property is equivalent to probability of the corpus participle to be identified in positive emotion tendency.
210, it is segmented using the corpus between point in mutual information calculating field in addition to default emotion set and is inclined with Negative Affect
To emotion word between semantic dependency, obtain in field in addition to default emotion set corpus participle incline in Negative Affect
Upward probability.
For the embodiment of the present invention, here with the corpus participle preset in mutual information calculating field between point except set with
Semantic dependency between Negative Affect word, the semantic dependency are equivalent to corpus participle to be identified in Negative Affect tendency
Probability.
211, the corpus participle in the field in addition to default emotion set of words is calculated separately in positive emotion tendency
The sum of probability and probability of the corpus participle in Negative Affect tendency in the field in addition to default emotion set of words it
With.
It is understood that the corpus participle in field in addition to default emotion set of words may be positive emotion tendency
Emotion word, it is also possible to for the emotion word of Negative Affect tendency, calculate separately all corpus to be identified in field and segment in positive feelings
All emotions to be identified segment the sum of the probability value on Negative Affect in the sum of probability in sense tendency and field.
212, by probability of the corpus participle in positive emotion tendency in the field in addition to default emotion set of words
The sum of subtract probability the sum of of the corpus participle in Negative Affect tendency in the field in addition to default emotion set of words, root
The Sentiment orientation of the participle of the corpus in field in addition to default emotion set of words is determined according to difference.
Tending to Negative Affect tendency due to positive emotion is opposite Sentiment orientation, if except default emotion word in field
Probability the sum of of the corpus participle in positive emotion tendency except set is greater than in field in addition to default emotion set of words
Probability the sum of of the corpus participle in Negative Affect tendency illustrates corpus participle to be identified in positive feelings then difference is greater than zero
Probability in sense tendency is greater than the probability on Negative Affect, determines that the Sentiment orientation of corpus participle to be identified inclines for positive emotion
To；If probability the sum of of the corpus participle in field in addition to default emotion set of words in positive emotion tendency is less than field
Probability the sum of of the interior corpus participle in addition to default emotion set of words in Negative Affect tendency, then difference is said less than zero
Probability of the bright corpus participle to be identified in positive emotion tendency is less than the probability on Negative Affect, determines corpus to be identified point
The Sentiment orientation of word is Negative Affect tendency；If the corpus in field in addition to default emotion set of words is segmented in positive emotion
It is general in Negative Affect tendency that the sum of probability in tendency is equal to the corpus participle in field in addition to default emotion set of words
The sum of rate illustrates that probability of the corpus participle to be identified in positive emotion tendency is equal in Negative Affect then difference is equal to zero
On probability, determine the Sentiment orientation of corpus to be identified participle for neutral Sentiment orientation.
Further, the specific implementation as Fig. 1 the method, the embodiment of the invention provides a kind of feelings of corpus participle
Identification device is felt, as shown in figure 3, described device includes: acquiring unit 31, screening unit 32, computing unit 33, determination unit
34。
Acquiring unit 31 can be used for obtaining the corpus participle of each data source in field；
Screening unit 32 can be used for filtering out Sentiment orientation in field from the corpus of each data source participle bright
Aobvious default emotion set of words, and the default emotion set of words is merged with the emotion set of words in sentiment dictionary, it obtains
Emotion set of words in field, record has the specific emotion word of Sentiment orientation in field in emotion set of words in the field；
Computing unit 33 can be used for utilizing the interior corpus in addition to default emotion set of words of mutual information calculating field between point
Semantic dependency in emotion set of words in participle and the field between the specific emotion word of Sentiment orientation, obtains in field
Corpus in addition to default emotion set of words segments the probability on clear Sentiment orientation；
Determination unit 34 can be used for according to the corpus participle in the field in addition to default emotion set of words clear
Probability on Sentiment orientation determines the Sentiment orientation of the corpus participle in field in addition to default emotion set of words.
The present invention provides a kind of emotion recognition device of corpus participle, by from the corpus participle for obtaining each data source
The significantly default emotion set of words of Sentiment orientation in field is filtered out, and is merged with the emotion set of words in sentiment dictionary,
The emotion set of words in field is obtained, so that record has emotion known to more fully Sentiment orientation in emotion set of words in field
Word further utilizes the corpus participle between point in mutual information calculating field in addition to default emotion set of words and the emotion in field
Semantic dependency in set of words between the specific emotion word of Sentiment orientation obtains in field in addition to default emotion set of words
Corpus segments the probability on clear Sentiment orientation, so that it is determined that the corpus in field in addition to default emotion set of words segmented
Sentiment orientation can accurately identify the Sentiment orientation of more corpus participles in field.With in the prior art pass through sentiment dictionary in
Basic emotion word compared to carry out corpus participle emotion recognition mode, the embodiment of the present invention is for the corpus point in specific area
Word carry out emotion recognition, by between point mutual information in the way of come calculate corpus participle Sentiment orientation, can effectively extract spy
Determine the emotion word in field, and accurately identify the Sentiment orientation of emotion word, and with the lasting expansion that corpus segments, in field
New emotion vocabulary can be captured effectively.
The further explanation of emotion recognition device as the participle of corpus shown in Fig. 3, Fig. 4 is according to embodiments of the present invention
The structural schematic diagram of the emotion recognition device of another corpus participle, as shown in figure 4, described device further include:
Collector unit 35 can be used for before the corpus participle of each data source in the acquisition field, by counting greatly
The corpus of each data source in multiple known arts is collected according to platform；
Training unit 36 can be used for the corpus of data source each in the multiple known art is defeated as sample data
Enter network model to be trained, obtains corpus field identification model.
Further, the screening unit 32 includes:
Comparison module 321 can be used for scanning the corpus participle of each data source one by one, by corpus participle and basis
Emotion word in sentiment dictionary is compared；
First screening module 322 can be used for filtering out comparison result from the corpus of data source each in field participle
Inconsistent corpus participle obtains the unknown corpus participle of Sentiment orientation in field；
Second screening module 323, can be used for by way of manually marking that Sentiment orientation is unknown out of described field
The apparent corpus participle of Sentiment orientation is filtered out in corpus participle, obtains default emotion set of words.
It further, include the emotion word of positive Sentiment orientation and passive feelings in field in emotion set in the field
Feel the emotion word of tendency, the computing unit 33 includes:
First computing module 331 can be used for interior in addition to default emotion set using mutual information calculating field between point
Semantic dependency between corpus participle and the emotion word of positive emotion tendency, obtains in field in addition to default emotion set
Probability of the corpus participle in positive emotion tendency；
Second computing module 332 can be used for interior in addition to default emotion set using mutual information calculating field between point
Semantic dependency between corpus participle and the emotion word of Negative Affect tendency, obtains in field in addition to default emotion set
Probability of the corpus participle in Negative Affect tendency.
Further, the determination unit 34 includes:
Third computing module 341 can be used for calculating separately the corpus in the field in addition to default emotion set of words
The sum of probability on positive emotion is inclined to is segmented to disappear with the corpus participle in the field in addition to default emotion set of words
The sum of probability on the Sentiment orientation of pole；
Determining module 342 can be used for segmenting the corpus in the field in addition to default emotion set of words positive
The sum of probability on Sentiment orientation subtracts the participle of the corpus in the field in addition to default emotion set of words and inclines in Negative Affect
The sum of upward probability determines the Sentiment orientation of the participle of the corpus in field in addition to default emotion set of words according to difference.
Further, the determining module 342, if specifically can be used for the difference greater than zero, it is determined that in field
The Sentiment orientation of corpus participle in addition to default emotion set of words is positive emotion tendency；
The determining module 342, if specifically can be also used for the difference less than zero, it is determined that except default feelings in field
The Sentiment orientation of the corpus participle except set of words is felt for Negative Affect tendency；
The determining module 342, if specifically can be also used for the difference equal to zero, it is determined that except default feelings in field
The Sentiment orientation for feeling the corpus participle except set of words is neutral Sentiment orientation.
Further, the acquiring unit 31 includes:
Module 311 is obtained, can be used for for the corpus data of each data source of tera incognita being input to the corpus neck
Domain identification model obtains the corresponding field of corpus of each data source；
Summarizing module 312 obtains each data in field for summarizing the corpus for belonging to each data source in same area
The corpus in source；
Word segmentation module 313 can be used for carrying out word segmentation processing to the corpus of data source each in the field, obtain field
The corpus of interior each data source segments.
It should be noted that each functional unit involved by a kind of emotion recognition device of corpus participle provided in this embodiment
Other it is corresponding describe, can be with reference to the corresponding description in Fig. 1 and Fig. 2, details are not described herein.
It is deposited thereon based on above-mentioned method as depicted in figs. 1 and 2 correspondingly, the present embodiment additionally provides a kind of storage medium
Computer program is contained, which realizes the emotion recognition of above-mentioned corpus participle as depicted in figs. 1 and 2 when being executed by processor
Method.
Based on this understanding, the technical solution of the application can be embodied in the form of software products, which produces
Product can store in a non-volatile memory medium (can be CD-ROM, USB flash disk, mobile hard disk etc.), including some instructions
With so that computer equipment (can be personal computer, server or the network equipment an etc.) execution the application is each
Method described in implement scene.
Based on above-mentioned method as shown in Figure 1 and Figure 2 and Fig. 3, virtual bench embodiment shown in Fig. 4, in order to realize
Above-mentioned purpose, the embodiment of the present application also provides a kind of computer equipments, are specifically as follows personal computer, server, network
Equipment etc., the entity device include storage medium and processor；Storage medium, for storing computer program；Processor is used for
Computer program is executed to realize the emotion identification method of above-mentioned corpus participle as depicted in figs. 1 and 2.
Optionally, which can also include user interface, network interface, camera, radio frequency (Radio
Frequency, RF) circuit, sensor, voicefrequency circuit, WI-FI module etc..User interface may include display screen
(Display), input unit such as keyboard (Keyboard) etc., optional user interface can also connect including USB interface, card reader
Mouthful etc..Network interface optionally may include standard wireline interface and wireless interface (such as blue tooth interface, WI-FI interface).
It will be understood by those skilled in the art that the entity device structure of the emotion recognition of corpus participle provided in this embodiment
The restriction to the entity device is not constituted, may include more or fewer components, perhaps combines certain components or difference
Component layout.
It can also include operating system, network communication module in storage medium.Operating system is that the above-mentioned computer of management is set
The program of standby hardware and software resource, supports the operation of message handling program and other softwares and/or program.Network communication mould
Block leads to for realizing the communication between each component in storage medium inside, and between other hardware and softwares in the entity device
Letter.
Through the above description of the embodiments, those skilled in the art can be understood that the application can borrow
It helps software that the mode of necessary general hardware platform is added to realize, hardware realization can also be passed through.Pass through the skill of application the application
Art scheme, compared with currently available technology, the embodiment of the present invention carries out emotion recognition, benefit for the corpus participle in specific area
The Sentiment orientation that corpus participle is calculated with the mode of mutual information between point, can effectively extract the emotion word of specific area, and
Accurately identify the Sentiment orientation of emotion word, and with the lasting expansion that corpus segments, the new emotion vocabulary in field can be by
Effectively capture.
It will be appreciated by those skilled in the art that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, module in attached drawing or
Process is not necessarily implemented necessary to the application.It will be appreciated by those skilled in the art that the mould in device in implement scene
Block can according to implement scene describe be distributed in the device of implement scene, can also carry out corresponding change be located at be different from
In one or more devices of this implement scene.The module of above-mentioned implement scene can be merged into a module, can also be into one
Step splits into multiple submodule.
Above-mentioned the application serial number is for illustration only, does not represent the superiority and inferiority of implement scene.Disclosed above is only the application
Several specific implementation scenes, still, the application is not limited to this, and the changes that any person skilled in the art can think of is all
The protection scope of the application should be fallen into.
Claims (10)
1. a kind of emotion identification method of corpus participle, which is characterized in that the described method includes:
The corpus participle of each data source in acquisition field；
The significantly default emotion set of words of Sentiment orientation in field is filtered out from the corpus of each data source participle, and will
The default emotion set of words is merged with the emotion set of words in sentiment dictionary, obtains the emotion set of words in field, institute
It states to record in emotion set of words in field and has the specific emotion word of Sentiment orientation in field；
Utilize the corpus participle between point in mutual information calculating field in addition to default emotion set of words and the emotion in the field
Semantic dependency in set of words between the specific emotion word of Sentiment orientation obtains in field in addition to default emotion set of words
Corpus segments the probability on clear Sentiment orientation；
The probability on clear Sentiment orientation is segmented according to the corpus in the field in addition to default emotion set of words, determines neck
The Sentiment orientation of corpus participle in domain in addition to default emotion set of words.
2. the method according to claim 1, wherein described screen from the corpus of each data source participle
The significantly default emotion set of words of Sentiment orientation includes: in field out
The corpus participle for scanning each data source one by one, corpus participle and the emotion word in basic sentiment dictionary are compared
It is right；
The inconsistent corpus participle of comparison result is filtered out from the corpus of data source each in field participle, obtains field inside information
Unknown corpus participle is inclined in sense；
It is obvious that Sentiment orientation is filtered out from the unknown corpus participle of Sentiment orientation in the field by way of manually marking
Corpus participle, obtain default emotion set of words.
3. the method according to claim 1, wherein including positive feelings in field in emotion set in the field
The emotion word of sense tendency and the emotion word of Negative Affect tendency, it is described interior except default emotion using mutual information calculating field between point
Semanteme in corpus participle except set of words and the emotion set of words in the field between the specific emotion word of Sentiment orientation
Correlation, obtaining the probability that the corpus in field in addition to default emotion set of words segments on clear Sentiment orientation includes:
Utilize the corpus participle and the emotion of positive emotion tendency between point in mutual information calculating field in addition to default emotion set
Semantic dependency between word, the corpus participle obtained in field in addition to default emotion set are general in positive emotion tendency
Rate；
Utilize the corpus participle and the emotion of Negative Affect tendency between point in mutual information calculating field in addition to default emotion set
Semantic dependency between word, the corpus participle obtained in field in addition to default emotion set are general in Negative Affect tendency
Rate.
4. according to the method described in claim 3, it is characterized in that, it is described according in the field except default emotion set of words it
Outer corpus segments the probability on clear Sentiment orientation, determines the corpus participle in field in addition to default emotion set of words
Sentiment orientation includes:
Calculate separately in the field in addition to default emotion set of words corpus participle positive emotion tendency on probability it
With the sum of the probability with the corpus participle in the field in addition to default emotion set of words in Negative Affect tendency；
Probability the sum of of the corpus participle in the field in addition to default emotion set of words in positive emotion tendency is subtracted
Probability the sum of of the corpus participle in Negative Affect tendency in the field in addition to default emotion set of words, it is true according to difference
Determine the Sentiment orientation of the corpus participle in field in addition to default emotion set of words.
5. according to the method described in claim 4, it is characterized in that, it is described by the field in addition to default emotion set of words
Probability the sum of of the corpus participle in positive emotion tendency subtract the corpus in the field in addition to default emotion set of words
The sum of the probability in Negative Affect tendency is segmented, the corpus in field in addition to default emotion set of words point is determined according to difference
The Sentiment orientation of word includes:
If the difference is greater than zero, it is determined that the Sentiment orientation that corpus in addition to default emotion set of words segments in the field in is
Positive emotion tendency；
If the difference is less than zero, it is determined that the Sentiment orientation of corpus in field in addition to default emotion set of words participle is
Negative Affect tendency；
If the difference is equal to zero, it is determined that the Sentiment orientation that corpus in addition to default emotion set of words segments in the field in is
Neutral Sentiment orientation.
6. method according to any one of claims 1-5, which is characterized in that each data source in the acquisition field
Corpus participle before, the method also includes:
The corpus of each data source in multiple known arts is collected by big data platform；
It is trained, obtains using the corpus of each data source in the multiple known art as sample data input network model
Corpus field identification model.
7. according to the method described in claim 6, it is characterized in that, the corpus of each data source segments packet in the acquisition field
It includes:
The corpus data of each data source of tera incognita is input to corpus field identification model, obtains each data source
The corresponding field of corpus；
Summarize the corpus for belonging to each data source in same area, obtains the corpus of each data source in field；
Word segmentation processing is carried out to the corpus of each data source in the field, obtains the corpus participle of each data source in field.
8. a kind of emotion recognition device of corpus participle, which is characterized in that described device includes:
Acquiring unit, the corpus for obtaining each data source in field segment；
Screening unit, for filtering out the significantly default feelings of Sentiment orientation in field from the corpus of each data source participle
Feel set of words, and the default emotion set of words is merged with the emotion set of words in sentiment dictionary, obtains in field
Emotion set of words, record has the specific emotion word of Sentiment orientation in field in emotion set of words in the field；
Computing unit, for using between point in mutual information calculating field in addition to default emotion set of words corpus participle with it is described
Semantic dependency in emotion set of words in field between the specific emotion word of Sentiment orientation obtains in field except default emotion
Corpus except set of words segments the probability on clear Sentiment orientation；
Determination unit, for being segmented according to the corpus in the field in addition to default emotion set of words in clear Sentiment orientation
Probability, determine in field in addition to default emotion set of words corpus participle Sentiment orientation.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 7 the method when executing the computer program.
10. a kind of computer storage medium, is stored thereon with computer program, which is characterized in that the computer program is located
The step of reason device realizes method described in any one of claims 1 to 7 when executing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910119320.1A CN109977396A (en) | 2019-02-18 | 2019-02-18 | Emotion identification method, device, computer equipment and the computer storage medium of corpus participle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910119320.1A CN109977396A (en) | 2019-02-18 | 2019-02-18 | Emotion identification method, device, computer equipment and the computer storage medium of corpus participle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109977396A true CN109977396A (en) | 2019-07-05 |
Family
ID=67077092
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910119320.1A Pending CN109977396A (en) | 2019-02-18 | 2019-02-18 | Emotion identification method, device, computer equipment and the computer storage medium of corpus participle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109977396A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112149428A (en) * | 2020-10-12 | 2020-12-29 | 珍岛信息技术（上海）股份有限公司 | Intelligent writing auxiliary system based on semantic analysis and deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090125371A1 (en) * | 2007-08-23 | 2009-05-14 | Google Inc. | Domain-Specific Sentiment Classification |
| CN101782898A (en) * | 2010-03-25 | 2010-07-21 | 中国科学院计算技术研究所 | Method for analyzing tendentiousness of affective words |
| CN102663139A (en) * | 2012-05-07 | 2012-09-12 | 苏州大学 | Method and system for constructing emotional dictionary |
| CN105069021A (en) * | 2015-07-15 | 2015-11-18 | 广东石油化工学院 | Chinese short text sentiment classification method based on fields |
-
2019
- 2019-02-18 CN CN201910119320.1A patent/CN109977396A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090125371A1 (en) * | 2007-08-23 | 2009-05-14 | Google Inc. | Domain-Specific Sentiment Classification |
| CN101782898A (en) * | 2010-03-25 | 2010-07-21 | 中国科学院计算技术研究所 | Method for analyzing tendentiousness of affective words |
| CN102663139A (en) * | 2012-05-07 | 2012-09-12 | 苏州大学 | Method and system for constructing emotional dictionary |
| CN105069021A (en) * | 2015-07-15 | 2015-11-18 | 广东石油化工学院 | Chinese short text sentiment classification method based on fields |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112149428A (en) * | 2020-10-12 | 2020-12-29 | 珍岛信息技术（上海）股份有限公司 | Intelligent writing auxiliary system based on semantic analysis and deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109284729B (en) | Method, device and medium for acquiring face recognition model training data based on video | |
| Wang et al. | Stereoscopic thumbnail creation via efficient stereo saliency detection | |
| Sabino et al. | A texture approach to leukocyte recognition | |
| Zeleny et al. | Box clustering segmentation: A new method for vision-based web page preprocessing | |
| CN109800731A (en) | Fingerprint input method and relevant apparatus | |
| CN107958230A (en) | Facial expression recognizing method and device | |
| CN108205580A (en) | A kind of image search method, device and computer readable storage medium | |
| CN110059156A (en) | Coordinate retrieval method, apparatus, equipment and readable storage medium storing program for executing based on conjunctive word | |
| CN113963303A (en) | Image processing method, video recognition method, device, equipment and storage medium | |
| CN106649276A (en) | Identification method and device for core product word in title | |
| CN110232331A (en) | A kind of method and system of online face cluster | |
| CN106815253B (en) | Mining method based on mixed data type data | |
| CN113204636A (en) | Knowledge graph-based user dynamic personalized image drawing method | |
| CN111144215A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN106997350A (en) | A kind of method and device of data processing | |
| Jeong et al. | Automatic detection of slide transitions in lecture videos | |
| CN109977396A (en) | Emotion identification method, device, computer equipment and the computer storage medium of corpus participle | |
| CN110472018A (en) | Information processing method, device and computer storage medium based on deep learning | |
| CN110990834A (en) | Static detection method, system and medium for android malicious software | |
| KR102444172B1 (en) | Method and System for Intelligent Mining of Digital Image Big-Data | |
| Xu et al. | Estimating similarity of rich internet pages using visual information | |
| CN110460711A (en) | A kind of method and device that picture is shown | |
| Ahmad et al. | A hierarchical approach to event discovery from single images using MIL framework | |
| Capdevila et al. | Recognizing warblers: a probabilistic model for event detection in Twitter | |
| David et al. | Authentication of Vincent van Gogh’s work |
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 |
Application publication date: 20190705 |
|
| RJ01 | Rejection of invention patent application after publication |