CN103955452B - Method and equipment for intelligent detection of happiness based on text information - Google Patents
Method and equipment for intelligent detection of happiness based on text information Download PDFInfo
- Publication number
- CN103955452B CN103955452B CN201410215110.XA CN201410215110A CN103955452B CN 103955452 B CN103955452 B CN 103955452B CN 201410215110 A CN201410215110 A CN 201410215110A CN 103955452 B CN103955452 B CN 103955452B
- Authority
- CN
- China
- Prior art keywords
- emotion
- word
- text message
- basic
- dimension
- 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.)
- Active
Links
Abstract
The invention discloses a method and equipment for intelligent detection of happiness based on text information. The method comprises the following steps of carrying out word segmentation processing on the text information, and obtaining at least one segmented word; according to the at least one segmented word, determining all sentiment words contained in the at least one segmented word from a sentiment word base, wherein the sentiment word base is stored with the sentiment words and the component values of the sentiment words on each basic sentiment dimension; determining the word frequency of each sentiment word in the text information; obtaining the component value of each sentiment word on each basic sentiment dimension from the sentiment word base; according to the word frequency of each sentiment word in the text information, the component value of each sentiment word on each basic sentiment dimension, and the weight of each basic sentiment in the happiness, determining the happiness value of the test information. The method and the equipment solve the problem that the happiness of a tested person in the past is difficultly detected and compared due to the reasons of memory, self perception, etc. of the tested person.
Description
Technical field
The present invention relates to areas of information technology, in particular it relates to a kind of happiness Intelligent Measurement side based on text message
Method and equipment.
Background technology
CCTV in 2012 about " happiness " investigation, by " happiness " this target that people constantly pursue always over the past thousands of years
The visual field of the public is introduced back into, is socially caused immediately extensively and heated discussion.In the past for the research of happiness in,
It is in the majority with psychology aspect, as the progress of science and technology has engendered some new measuring methods, such as FMRI (functional MRI into
Picture) technology, PET (positron e mission computed tomography) technology.In Science of Economics, " measurable effectiveness " is made always
It is economists to the basis of happy sensed quantity.The existing research method for happiness has self-report method (self-
Report), insider/observer's report method, physiological measurements method and task mensuration.These methods are to quilt by survey
The happiness of survey person is measured, that is, formulate the scale of happy sensed quantity, then selects measurand, measurand according to scale
After completion problem, effective questionnaire is filtered out, be analyzed by data, draw the happiness of measurand.But this ask
Rolling up the method investigated environment to measured man memory and at that time has dependence very high, can only substantially measure measured
The happiness level of nearest a period of time, it is difficult to the personal happiness in certain special time period of past of quantification ground measurement.And
And, this method is stronger to the dependence that measured perceives self, and this greatly reduces the reliability and convincingness of measurement result,
And the result of mistake may be caused.
The content of the invention
It is an object of the invention to provide a kind of objective, easy happiness quantification Intelligent Measurement side based on text message
Method and equipment.
To achieve these goals, the present invention provides a kind of happiness intelligent detecting method based on text message, the party
Method includes：Word segmentation processing is carried out to the text message, and obtains at least one participle；According at least one participle come from
All emotion words that at least one participle includes are determined in emotion dictionary, be wherein stored with emotion word in the emotion dictionary
With component value of the emotion word in each basic emotion dimension；Determine each emotion word in the text message in the text
Word frequency number in this information；Each emotion word is in described each basic emotion dimension described in being obtained from the emotion dictionary
Component value；And word frequency number, described each emotion word according to described each emotion word in the text message are in each base
Component value and the shared weight in happiness of each basic emotion in this emotion dimension determine the text message
Happy inductance value.
Preferably, the part of speech of the emotion word that is also stored with the emotion dictionary.
Preferably, word segmentation processing is carried out to the text message, also obtains the part of speech of at least one participle；And root
According at least one participle and the part of speech of at least one participle, to determine described at least one from the emotion dictionary
All emotion words that participle includes, wherein, each emotion word carries part-of-speech information.
Preferably, the happy inductance value of the text message can in the following manner be determined：According to described each emotion word
Word frequency number in the text message, it is determined that accounting in all emotion words of described each emotion word in the text message
Than；Accounting and described each emotion word in all emotion words of described each emotion word in the text message
Component value in each basic emotion dimension, determines component of the text message in described each basic emotion dimension
Value；And the component value and each basic emotion according to the text message in described each basic emotion dimension are in good fortune
Shared weight in good fortune sense, determines the happy inductance value of the text message.
Preferably, the accounting of each emotion word in all emotion words can in the following manner be determined：
Wherein, pkRepresent the accounting in all emotion words of k-th emotion word in the text message；fkRepresent k-th
Word frequency number of the emotion word in the text message；N represents the sum of the emotion word in the text message.
Preferably, component of the text message in described each basic emotion dimension can in the following manner be determined
Value：
Wherein, djRepresent component value of the text message in j-th basic emotion dimension；pkRepresent k-th emotion word
The accounting in all emotion words in the text message；ejkRepresent k-th emotion word in j-th basic emotion dimension
Component value；N represents the sum of the emotion word in the text message.
Preferably, the happy inductance value of the text message can in the following manner be determined：
Wherein, H represents the happy inductance value of the text message；djRepresent that the text message is tieed up in j-th basic emotion
Component value on degree；M represents the sum of the basic emotion dimension；ωjRepresent that j-th basic emotion is shared in happiness
Weight.
The present invention also provides a kind of happiness intelligent detection equipment based on text message, and the equipment includes：For to institute
Stating text message carries out word segmentation processing, and obtains the device of at least one participle；For according at least one participle come from
The device of all emotion words that at least one participle includes is determined in emotion dictionary, is wherein stored with the emotion dictionary
Emotion word and component value of the emotion word in each basic emotion dimension；For determining each emotion in the text message
The device of word frequency number of the word in the text message；For obtaining described each emotion word from the emotion dictionary described
The device of the component value in each basic emotion dimension；And for each emotion word according in the text message
The component value and each basic emotion of word frequency number, described each emotion word in each basic emotion dimension institute in happiness
The weight for accounting for determines the device of the happy inductance value of the text message.
In the above-mentioned technical solutions, based on emotion dictionary to measured's time in the past section in text message (for example, in society
Hand over the text information recorded on network) it is analyzed, it can be deduced that emotion word included in the text message and its each
Component value in individual basic emotion dimension.Afterwards, on the basis of statistical analysis is carried out to the emotion word, can be with quantification ground
Draw the happy inductance value of the measured that the Textual information is reflected.The present invention is solved due to the memory of measured and self felt
The reason such as know, the problem for causing measured to detect and compare with being difficult to quantification in the happiness of time in the past section.The present invention is carried
Environmental factor dependence of the detection method of confession to measured man memory and at that time is smaller, the dependence self perceived to measured
Also it is smaller, such that it is able to it is more objective, easily, quantification detect the happy inductance value of measured, and can greatly improve
The reliability and convincingness of testing result.
Other features and advantages of the present invention will be described in detail in subsequent specific embodiment part.
Brief description of the drawings
Accompanying drawing is, for providing a further understanding of the present invention, and to constitute the part of specification, with following tool
Body implementation method is used to explain the present invention together, but is not construed as limiting the invention.In the accompanying drawings：
Fig. 1 is the flow chart of the happiness intelligent detecting method based on text message according to the embodiment of the present invention；
And
Fig. 2~Fig. 5 is the testing result that the happiness intelligent detecting method provided according to the present invention carries out happiness detection
Schematic diagram.
Specific embodiment
Specific embodiment of the invention is described in detail below in conjunction with accompanying drawing.It should be appreciated that this place is retouched
The specific embodiment stated is merely to illustrate and explain the present invention, and is not intended to limit the invention.
The invention provides a kind of happiness intelligent detecting method based on text message, its flow chart is as shown in Figure 1.Should
Method can include：Step S101, word segmentation processing is carried out to text message, and obtain at least one participle；Step S102, according to
At least one participle determines all emotion words that at least one participle includes from emotion dictionary；Step S103, it is determined that literary
The word frequency number of each emotion word in this information in text message；Step S104 obtains each emotion word from emotion dictionary and exists
Component value in each basic emotion dimension；And step S105, it is word frequency number according to each emotion word in text message, each
Component value and each basic emotion in happiness shared weight of the individual emotion word in each basic emotion dimension, come true
Determine the happy inductance value of text message.Wherein, the emotion word that can be stored with the emotion dictionary and the emotion word are in each basic feelings
Component value in sense dimension.
In the present invention, emotion be divide into several basic emotions, that is, emotion is expressed as a multi-C vector, one
One basic emotion of individual dimension correspondence.In an example of the invention, can using " expectations ", " liking ", " happiness ", " shy
Be surprised ", " anxiety ", " sadness ", " anger " and " hatred " this eight basic emotions.
Specifically, step S101, can carry out word segmentation processing to text message first, by a text message split into
A few participle.The step can select some participle instruments to complete, for example, SCWS (simple Chinese automatic word-cut),
ICTCLAS (Chinese lexical analysis system) etc..
Next, step S102, determines that at least one participle includes according at least one participle from emotion dictionary
All emotion words.Each participle can be matched with the emotion word stored in emotion dictionary.If certain participle is in feelings
There is record in sense dictionary, then the participle can be as text information emotion word., whereas if not recording, then should
Emotion word of the participle not as text information.Carried out one by one by by the emotion word stored in all participles and emotion dictionary
Match somebody with somebody, it may be determined that go out all emotion words that participle includes.
So-called " emotion word " refers to the word per se with emotion, for example, glad, sad, indignation, shocking, going mad.
These emotion words can have one-component value in each basic emotion dimension.That is, these emotion words are reflected
Emotion comprehensive embodiment can be come by the component value in each basic emotion dimension.Just it is stored with emotion dictionary a large amount of such
Emotion word, and component value of each emotion word in each basic emotion dimension.
Because Chinese language implication is extremely complex, thus same emotion word may have different parts of speech, the emotion of its expression
Also differ.So, add part of speech this attribute to help to make a distinction emotion word.
Preferably, the part of speech of the emotion word that can also be stored with the emotion dictionary.Word segmentation processing is carried out to text message, also
The part of speech of at least one participle can be obtained.In such a case, it is possible to according at least one participle and at least one participle
Part of speech, to determine all emotion words that at least one participle includes from emotion dictionary.Wherein, word is carried in each emotion word
Property information.
That is, in the case where emotion word is determined jointly by participle and part of speech, the emotion word is with part-of-speech information
Emotion word.Therefore, even identical word, if part of speech difference, different emotion words are also determined as.For example, right
In " happiness " one word, at least two kinds of different parts of speech of adverbial word and adjective are then determining emotion word jointly by participle and part of speech
In the case of, " happiness " of adverbial word is determined as two different emotion words with adjectival " happiness ".
It should be noted that as long as be stored with a large amount of emotion words and each emotion word are in described each basic emotion dimension
On component value (part of speech of each emotion word that preferably, is also stored with) emotion dictionary, can be used for the present invention.At this
In one example of invention, the emotion dictionary for being used can belong to emotion word and each emotion word to being labelled with which word
The corpus of the component value in each basic emotion dimension arranged obtained from emotion dictionary.Wherein, the corpus can
Being, for example, Ren-CECps (Chinese mood corpus).
It is described more fully below and how the corpus is arranged, is met the emotion of happiness detection demand
The method of dictionary.
It is possible, firstly, to read out all emotion words included in the corpus and each emotion word from corpus each
The record of the component value in individual basic emotion dimension, forms an initial emotion dictionary.For example, result excel can will be read
The form of form is showed.
Preferably, the part of speech of each emotion word is also labeled with the corpus.In such a case, it is possible to together with the emotion
The part of speech of word reads out together (in other words, can also be included corresponding with each emotion word in the initial emotion dictionary
Part of speech).Adding part of speech this attribute helps to make a distinction emotion word, and to the initial emotion dictionary after contributing to
Arranged, to be met the emotion dictionary of happiness detection demand.
The following is with computer to corpus in the example that is read out of " here it is grief of justice " the words.As above
It is described, it is basic with " expectation ", " liking ", " happiness ", " surprised ", " anxiety ", " sadness ", " anger " and " hatred " eight herein
Illustrated as a example by emotion.
The emotion word that computer is read out in the words is " justice " and " grief ", and they are in eight basic emotions
(expect (Expect), like (Love), glad (Joy), surprised (Surprise), anxiety (Anxiety), sad (Sorrow),
Angry (Anger) and hate (Hate)) component value in dimension is respectively：0.5th, 0.7,0,0,0,0,0,0 and 0,0,0,0,0,
1.0th, 0,0, it is shown in Table 1.Additionally, it is all adjective that computer also reads out their part of speech, wherein, " adjective " this part of speech can
To be represented with mark a.
Table 1
| Vocabulary | Part of speech | Expect | Like | It is glad | It is surprised | Anxiety | It is sad | It is angry | Hatred |
| It is fair | a | 0.5 | 0.7 | 0 | 0 | 0 | 0 | 0 | 0 |
| It is grieved | a | 0 | 0 | 0 | 0 | 0 | 1.0 | 0 | 0 |
Corpus is read out according to as above method, it can be deduced that the initial emotion dictionary.In order to eliminate initial feelings
Redundancy in sense dictionary, makes final emotion dictionary more refine, in a preferred embodiment of the invention, can be to institute
State initial emotion dictionary to be arranged, including one or more of following operation：
(1) for same emotion word, can be all identical by its part of speech and component value in each basic emotion dimension
Record merge into a record.That is, only retaining one therein for this record that repeats.Operated more than,
Each record in emotion dictionary is all unduplicated, for same emotion word, or part of speech is different, or some or
Certain several component value is different.
(2) for same part of speech under same emotion word, by it in each basic emotion dimension component value correspondence take
Average value, is recorded with a component value for obtaining the emotion word under the part of speech.The purpose of do so is that can retain the feelings
All basic emotions of the sense word under this kind of part of speech.
For example, show that emotion word is identical, part of speech is identical in table 2, but the different situation of basic emotion component value.Such as table 2
It is shown, although the part of speech of vocabulary " patriotic " is adjective, but using in different linguistic context, the emotion expressed by it is endless
Identical, is all to represent " liking " but intensity is different.
Table 2
| Vocabulary | Part of speech | It is surprised | It is sad | Like | It is glad | Hatred | Expect | Anxiety | It is angry |
| It is patriotic | a | 0 | 0 | 0.7 | 0 | 0 | 0 | 0 | 0 |
| It is patriotic | a | 0 | 0 | 0.6 | 0 | 0 | 0 | 0 | 0 |
| It is patriotic | a | 0 | 0 | 0 | 0.6 | 0 | 0 | 0 | 0 |
| It is patriotic | a | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
| It is patriotic | a | 0 | 0 | 0.9 | 0 | 0 | 0 | 0 | 0 |
| It is patriotic | a | 0 | 0 | 0.8 | 0 | 0 | 0 | 0 | 0 |
| It is patriotic | a | 0 | 0 | 0.5 | 0 | 0 | 0 | 0 | 0 |
According to aforesaid operations, to table 2 in component value pair of " patriotic " this emotion word in each basic emotion dimension
Should average, its result is as shown in table 3.
Table 3
| Vocabulary | Part of speech | It is surprised | It is sad | Like | It is glad | Hatred | Expect | Anxiety | It is angry |
| It is patriotic | a | 0 | 0 | 0.643 | 0.086 | 0 | 0 | 0 | 0 |
As shown in table 3, after averaged, average mark value of " patriotic " this emotion word in " liking " dimension is 0.643,
Average mark value in " happiness " dimension is 0.086, and other basic emotions do not have in as adjectival " patriotic "
Reflection.
By aforesaid operations, can be greatly reduced on the basis of all basic emotions that emotion word is reflected are retained
Record in emotion dictionary.Same emotion word under for same part of speech, its component value note in each basic emotion dimension
Record is only one.
(3) component value in each basic emotion dimension is not higher than (be less than or equal to) emotion word of predetermined threshold value
Record deletion.
Due in emotion dictionary may include some without obvious emotion vocabulary, as "Yes", " ", " " etc.
Deng.These vocabulary, although all show certain basic emotion, but it is all fainter.It is less strong in order to reduce these emotions
Strong vocabulary makes the happiness for detecting change more obvious for the influence of testing result, can be by these emotion low intensities
Word record deletion.
If the threshold value is set too high, dictionary coverage rate may be caused relatively low.If the threshold value is set
It is too low, may cause in dictionary still comprising the vocabulary of many emotion low intensities, and then cause the ripple of the happy inductance value for finally giving
Dynamic property is not obvious.Therefore, obtained when the threshold value is set, it is necessary between the fluctuation of happy inductance value and the coverage rate of dictionary
Balance.In practical operation, can be by successively decreasing step by step and the threshold value being determined after repetition test.For example, the threshold value can be by
It is set as 0.4.
After being arranged to initial emotion dictionary by one or more of aforesaid operations, happiness can be met
The emotion dictionary of detection demand, i.e. used in the happiness intelligent detecting method based on text message that the present invention is provided
Emotion dictionary.
Afterwards, in step S103, word frequency number of each emotion word in text message in text message is determined.It is so-called
" word frequency number ", refers to just number of times that the emotion word occurs in whole text message.How an emotion word is counted in text
The method of the word frequency number in information is well known to a person skilled in the art just repeating no more herein.
It should be noted that in above-mentioned preferred embodiment of the invention, emotion word can be with part-of-speech information
Emotion word.Therefore, word frequency number can also be the word frequency number corresponding with the emotion word with part-of-speech information.For example, it is assumed that a text
This information includes " happiness " of adverbial word and adjectival " happiness " the two different emotion words.So, it is determined that each feelings
It is when feeling the word frequency number of word, it should determine the word frequency number of " happiness " of adverbial word in text message respectively and adjectival " high
It is emerging " word frequency number in text message.
Next, in step S104, obtained from emotion dictionary each emotion word in each basic emotion dimension point
Value.
As described above, described each emotion word can be the emotion word with part-of-speech information.In this case, carrying out
During step S104, should be according to emotion word in itself and the part-of-speech information that carries, to find the record for matching from emotion dictionary
(so-called matching refers to the same vocabulary of same part of speech).After the record for matching is found, extract with part-of-speech information
The corresponding component value information of emotion word.
Afterwards, it is possible to carry out step S105, determining the happy inductance value of text message.
Be described below in detail in step S105, how word frequency number according to each emotion word in text message, each
Component value of the emotion word in each basic emotion dimension and the shared weight in happiness of each basic emotion determine
The method of the happy inductance value of text message.
It is possible, firstly, to the word frequency number according to each emotion word in text message, determines each emotion word in text message
In all emotion words in accounting.
For example, accounting of each emotion word in all emotion words can be determined in the following manner：
Wherein, pkRepresent the accounting in all emotion words of k-th emotion word in text message；fkRepresent k-th emotion
Word frequency number of the word in text message；N represents the sum of the emotion word in text message.
Afterwards, can be according to the accounting in all emotion words of each emotion word in text message and each emotion
Component value of the word in each basic emotion dimension, determines component value of the text message in each basic emotion dimension.
For example, component value of the text message in each basic emotion dimension can be determined in the following manner：
Wherein, djRepresent component value of the text message in j-th basic emotion dimension.If representing basic emotion with m
The sum of dimension, then j is the natural number less than or equal to m.pkRepresent all emotion words of k-th emotion word in text message
In accounting；ejkRepresent component value of k-th emotion word in j-th basic emotion dimension；N represents the emotion in text message
The sum of word.
Finally, the component value and each basic emotion according to text message in each basic emotion dimension are in happiness
Shared weight in sense, determines the happy inductance value of text message.
For example, the happy inductance value of text message can be determined in the following manner：
Wherein, H represents the happy inductance value of text message；djRepresent text message in j-th basic emotion dimension point
Value；M represents the sum of basic emotion dimension.M values are bigger, and the happy inductance value for obtaining is more accurate, but corresponding amount of calculation also can
Increase.In the present invention, with m=8 as an example.ωjRepresent the shared weight in happiness of j-th basic emotion.
The shared weight in happiness of each basic emotion (that is, ωj) can be set in advance, it is also possible to by asking
At least one of investigation method and expert graded (for example, Delphi expert gradeds) is rolled up to determine.Utilizing survey
In the case that method and expert graded to determine the weight jointly, Questionnaire results can be carried out with expert estimation result
Averagely draw final weight.For example, table 4 show by questionnaire method and expert graded it is described to determine jointly
The result of weight.Wherein, four kinds of positive emotions (expect, like, glad and surprised) are on the occasion of four kinds of Negative Affects (anxiety, compassions
Wound, angry and hatred) it is negative value.
Table 4
It should be appreciated that determining that each basic emotion is shared in happiness using questionnaire method and expert graded
Weight specific method be well known to a person skilled in the art, therefore, repeat no more in the present invention.
By above equation (1)~(3), it is possible to draw the happy inductance value of text message.The happy inductance value can reflect
Go out the happiness intensity that the text message of measured is reflected.Because the present invention is to detect happiness based on text message,
Thus it is possible to prevente effectively from due to the reason such as the memory of measured and self perception, causing measured in the happiness of time in the past section
Sense is difficult to the problem for detecting and comparing.Additionally, the detection method of present invention offer ring to measured man memory and at that time
Border dependence is smaller, to measured self perceive dependence it is also smaller, such that it is able to it is more objective, easily detect measured
Happiness, and the reliability and convincingness of testing result can be greatly improved.
The present invention also provides a kind of happiness intelligent detection equipment based on text message, and the equipment can include：For
Word segmentation processing is carried out to text message, and obtains the device of at least one participle；For according at least one participle come from emotion
The device of all emotion words that at least one participle includes is determined in dictionary, the emotion word that is wherein stored with the emotion dictionary and
Component value of the emotion word in each basic emotion dimension；For determining each emotion word in text message in text message
In word frequency number device；For obtaining component value of each emotion word in each basic emotion dimension from emotion dictionary
Device；And for the word frequency number according to each emotion word in text message, each emotion word in each basic emotion dimension
On component value and the shared weight in happiness of each basic emotion determine the device of the happy inductance value of text message.
Additionally, determining the specific method of the happy inductance value of the text message with above with reference to Fig. 1 and combination equation (1)
Method described by~(3) is consistent, and here is omitted.
Using happiness detection method provided by the present invention and equipment, any one text envelope can be quantitatively calculated
The happy inductance value of breath.Detected by many text messages to measured within a period of time, multiple correspondences can be obtained
Happy inductance value.It is compared by the multiple happy inductance value, measured's happiness during this period of time can be analyzed
The variation tendency of inductance value, as described further below.
It is assumed that the writing record using the detection method of present invention offer to a measured on social networks carries out happiness
Sense detection.Can be with an article (for example, blog article of bloger) of measured record for a text message unit.Such as
Fruit will detect the change of this bloger happiness within a period of time, and first intraday all blog articles can be detected respectively, ask
Average, then the average value be designated as the happy inductance value on the same day.Afterwards, it is possible to obtain bloger during this period of time each
It happy inductance value.The change of the bloger during this period of time happiness can be showed with the form of curve, as shown in Figure 2.
Fig. 2 is the detection method provided using the present invention, and the happiness for obtaining detect to the blog article in bloger one week
Value schematic diagram.As seen from Figure 2, the variation tendency of happy inductance value is：Happy inductance value gradually rises, and a peak value occurs,
After small falling, maximum is reached.Closing on weekend is can be seen that from the corresponding time, happy inductance value gradually rises,
There is small falling in Friday, and maximum is reached in weekend happiness inductance value, and this is to meet the normal happiness variation tendency of people
's.
Furthermore, Fig. 3 is one week after before the opening ceremony of the Olympic Games in 2008 that detection method detection provided by the present invention draws
Happy inductance value schematic diagram.As seen from Figure 3, the variation tendency of happy inductance value is：Happy inductance value occurred on the 9th in August in 2008
One peak.This 9th is, at second day of the opening of the Olympics, to add positive place's weekend with August, and the universal happiness of people should be higher
Actual conditions match.Happy inductance value is gradually reduced afterwards, until falling after rise to normal level.Because the opening of the Olympics with
Afterwards, the enthusiasm of people goes down, and along with the end at weekend, workaday beginning, happiness declines rapidly, and shows in one week
Cyclically-varying.
Additionally, Fig. 4 is one week after before " 5.12 " violent earthquakes in 2008 that detection method detection provided by the present invention draws
Happy inductance value schematic diagram.As seen from Figure 4, leading portion, happy inductance value is constantly in higher level, until good fortune after some day
Good fortune inductance value drastically declines, and within subsequent one week, happy inductance value maintains reduced levels always.Comparison time can be seen that good fortune
The good fortune sense peak value same day is the earthquake eve, and from earthquake the previous day to the same day, happy inductance value drastically declines, and people are immersed in always later
Grief.This is consistent with the variation tendency of the happy inductance value curve shown in Fig. 4.
Fig. 5 be detection method provided by the present invention detection draw 2009 around the Spring Festival the happy inductance value of a week illustrate
Figure.As seen from Figure 5, happy inductance value reaches a peak value, and the correspondence time, exactly the weekend before the Spring Festival, this is to meet people
Normal happiness variation tendency.
Therefore, the word that the happiness detection method provided with the present invention is recorded to people can be seen that by Fig. 2~Fig. 5
Information carries out happiness detection, and acquired results are consistent with actual conditions, and can be matched with historical event, fully tests
The reasonability and operability of the happiness detection method of offer of the present invention have been provided.
In sum, in the present invention, based on emotion dictionary to measured's time in the past section in text message (for example,
The text information recorded on social networks) be analyzed, it can be deduced that emotion word included in the text message and its
Component value in each basic emotion dimension.Afterwards, on the basis of statistical analysis is carried out to the emotion word, can quantify
Draw the happy inductance value of the measured that the Textual information is reflected with changing.The present invention solve due to measured memory and from
I such as perceives at the reason, the problem for causing measured to detect and compare with being difficult to quantification in the happiness of time in the past section.This hair
Environmental factor dependence of the detection method of bright offer to measured man memory and at that time is smaller, to measured self perceive according to
Bad property is also smaller, such that it is able to it is more objective, easily, quantification detect the happiness of measured, and can carry significantly
The reliability and convincingness of high detection result.
The preferred embodiment of the present invention is described in detail above in association with accompanying drawing, but, the present invention is not limited to above-mentioned reality
The detail in mode is applied, in range of the technology design of the invention, various letters can be carried out to technical scheme
Monotropic type, these simple variants belong to protection scope of the present invention.
It is further to note that each particular technique feature described in above-mentioned specific embodiment, in not lance
In the case of shield, can be combined by any suitable means, in order to avoid unnecessary repetition, the present invention to it is various can
The combination of energy is no longer separately illustrated.
Additionally, can also be combined between a variety of implementation methods of the invention, as long as it is without prejudice to originally
The thought of invention, it should equally be considered as content disclosed in this invention.
Claims (10)
1. a kind of happiness intelligent detecting method based on text message, the method includes：
Word segmentation processing is carried out to the text message, and obtains at least one participle；
All emotion words that at least one participle includes are determined from emotion dictionary according at least one participle,
Wherein, the emotion word that is stored with the emotion dictionary and component value of the emotion word in each basic emotion dimension；
Determine word frequency number of each emotion word in the text message in the text message；
The component value of each emotion word in described each basic emotion dimension is obtained from the emotion dictionary；
According to word frequency number of described each emotion word in the text message, it is determined that described each emotion word is in the text envelope
The accounting in all emotion words in breath；
Accounting and described each emotion word in all emotion words of described each emotion word in the text message
Component value in each basic emotion dimension, determines that the text message is tieed up in described each basic emotion in the following manner
Component value on degree：
Wherein, djRepresent component value of the text message in j-th basic emotion dimension；
pkRepresent the accounting in all emotion words of k-th emotion word in the text message；
ejkRepresent component value of k-th emotion word in j-th basic emotion dimension；
N represents the sum of the emotion word in the text message；And
Component value and each basic emotion according to the text message in described each basic emotion dimension is in happiness
In shared weight, determine the happy inductance value of the text message.
2. method according to claim 1, it is characterised in that the part of speech of the emotion word that is also stored with the emotion dictionary.
3. method according to claim 2, it is characterised in that word segmentation processing is carried out to the text message, institute is also obtained
State the part of speech of at least one participle；And
According at least one participle and the part of speech of at least one participle, determine from the emotion dictionary it is described to
All emotion words that a few participle includes, wherein, each emotion word carries part-of-speech information.
4. method according to claim 1, it is characterised in that determine described each emotion word all in the following manner
Accounting in emotion word：
Wherein, pkRepresent the accounting in all emotion words of k-th emotion word in the text message；
fkRepresent word frequency number of k-th emotion word in the text message；
N represents the sum of the emotion word in the text message.
5. method according to claim 1, it is characterised in that determine the happiness of the text message in the following manner
Value：
Wherein, H represents the happy inductance value of the text message；
djRepresent component value of the text message in j-th basic emotion dimension；
M represents the sum of the basic emotion dimension；
ωjRepresent the shared weight in happiness of j-th basic emotion.
6. a kind of happiness intelligent detection equipment based on text message, the equipment includes：
For carrying out word segmentation processing to the text message, and obtain the device of at least one participle；
It is in love for determining the institute that at least one participle includes from emotion dictionary according at least one participle
Feel the device of word, the emotion word that is wherein stored with the emotion dictionary and component of the emotion word in each basic emotion dimension
Value；
Device for determining word frequency number of each emotion word in the text message in the text message；
For obtaining the component value of each emotion word in described each basic emotion dimension from the emotion dictionary
Device；And
For word frequency number, described each emotion word of each emotion word in the text message according in each basic feelings
Feel component value and the shared weight in happiness of each basic emotion in dimension to determine the happiness of the text message
The device of inductance value；
The happy inductance value of the text message is determined in the following manner：
According to word frequency number of described each emotion word in the text message, it is determined that described each emotion word is in the text envelope
The accounting in all emotion words in breath；
Accounting and described each emotion word in all emotion words of described each emotion word in the text message
Component value in each basic emotion dimension, determines that the text message is tieed up in described each basic emotion in the following manner
Component value on degree：
Wherein, djRepresent component value of the text message in j-th basic emotion dimension；
pkRepresent the accounting in all emotion words of k-th emotion word in the text message；
ejkRepresent component value of k-th emotion word in j-th basic emotion dimension；
N represents the sum of the emotion word in the text message；And
Component value and each basic emotion according to the text message in described each basic emotion dimension is in happiness
In shared weight, determine the happy inductance value of the text message.
7. equipment according to claim 6, it is characterised in that the part of speech of the emotion word that is also stored with the emotion dictionary.
8. equipment according to claim 7, it is characterised in that word segmentation processing is carried out to the text message, institute is also obtained
State the part of speech of at least one participle；And
According at least one participle and the part of speech of at least one participle, determine from the emotion dictionary it is described to
All emotion words that a few participle includes, wherein, each emotion word carries part-of-speech information.
9. equipment according to claim 6, it is characterised in that determine described each emotion word all in the following manner
Accounting in emotion word：
Wherein, pkRepresent the accounting in all emotion words of k-th emotion word in the text message；
fkRepresent word frequency number of k-th emotion word in the text message；
N represents the sum of the emotion word in the text message.
10. equipment according to claim 6, it is characterised in that determine the happiness of the text message in the following manner
Inductance value：
Wherein, H represents the happy inductance value of the text message；
djRepresent component value of the text message in j-th basic emotion dimension；
M represents the sum of the basic emotion dimension；
ωjRepresent the shared weight in happiness of j-th basic emotion.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410215110.XA CN103955452B (en) | 2014-05-21 | 2014-05-21 | Method and equipment for intelligent detection of happiness based on text information |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410215110.XA CN103955452B (en) | 2014-05-21 | 2014-05-21 | Method and equipment for intelligent detection of happiness based on text information |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103955452A CN103955452A (en) | 2014-07-30 |
| CN103955452B true CN103955452B (en) | 2017-05-24 |
Family
ID=51332727
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410215110.XA Active CN103955452B (en) | 2014-05-21 | 2014-05-21 | Method and equipment for intelligent detection of happiness based on text information |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103955452B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109254993B (en) * | 2017-07-07 | 2021-06-01 | 掌沃云科技(北京)有限公司 | Text-based character data analysis method and system |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8463594B2 (en) * | 2008-03-21 | 2013-06-11 | Sauriel Llc | System and method for analyzing text using emotional intelligence factors |
| CN102122297A (en) * | 2011-03-04 | 2011-07-13 | 北京航空航天大学 | Semantic-based Chinese network text emotion extracting method |
| CN102163191A (en) * | 2011-05-11 | 2011-08-24 | 北京航空航天大学 | Short text emotion recognition method based on HowNet |
-
2014
- 2014-05-21 CN CN201410215110.XA patent/CN103955452B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN103955452A (en) | 2014-07-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Neuendorf | Content analysis and thematic analysis | |
| Topolewska et al. | The short IPIP-BFM-20 questionnaire for measuring the Big Five | |
| Suhr | Step your way through path analysis | |
| McNamara et al. | Coh-Metrix: An automated tool for theoretical and applied natural language processing | |
| Marjoribanks | Family background, individual and environmental influences, aspirations and young adults' educational attainment: A follow-up study | |
| Li et al. | Influence of entrepreneurial experience, alertness, and prior knowledge on opportunity recognition | |
| US9443193B2 (en) | Systems and methods for generating automated evaluation models | |
| Brady et al. | Validation of the Emotion Regulation Questionnaire in older community‐dwelling adults | |
| CN113871015B (en) | Man-machine interaction scheme pushing method and system for improving cognition | |
| CN106063699A (en) | A kind of medical apparatus and instruments description usability evaluation method based on eye movement technique | |
| Nanda et al. | Diagnostics for pretesting questionnaires: a comparative analysis | |
| Ferrara et al. | Contextual characteristics of locally dependent open-ended item clusters in a large-scale performance | |
| CN103955452B (en) | Method and equipment for intelligent detection of happiness based on text information | |
| Sakaluk et al. | Measurement memo I: Updated practices in psychological measurement for sexual scientists | |
| Sugara et al. | Factorial structure and psychometric properties of the quality of life inventory in an Indonesian college sample | |
| Connell | Survival analysis in prevention and intervention programs. | |
| Zanon et al. | Adaptation of the Steen Happiness Index (SHI) to Brazil: A comparison of the psychometric properties of the SHI and the Subjective Happiness Scale | |
| Zhang et al. | The defining features of emotions in online stories | |
| CN113808709A (en) | Text analysis-based psychoelasticity prediction method and system | |
| Setijono | DisPMO and DePMO as Six Sigma‐based forward‐looking quality performance measures | |
| Plank et al. | Exploring the CETSCALE in Soviet Armenia | |
| Najafi et al. | Long range dependence in texts: A method for quantifying coherence of text | |
| Tončić et al. | Effects of Momentary Affect on Satisfaction Judgments | |
| Pritoni | How to measure interest group influence: evidence from Italy | |
| Tian et al. | A study on the method of satisfaction measurement based on emotion space |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |