CN104951434B - The determination method and apparatus of brand mood - Google Patents
The determination method and apparatus of brand mood Download PDFInfo
- Publication number
- CN104951434B CN104951434B CN201510406454.3A CN201510406454A CN104951434B CN 104951434 B CN104951434 B CN 104951434B CN 201510406454 A CN201510406454 A CN 201510406454A CN 104951434 B CN104951434 B CN 104951434B
- Authority
- CN
- China
- Prior art keywords
- brand
- mood
- message
- brand message
- keyword
- 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
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The invention discloses a kind of determination method and apparatus of brand mood.Wherein, this method includes：Obtain the keyword of target brand；Obtain the brand message issued for user with the associated brand message of keyword, brand message；The corresponding brand mood of brand message is determined using the disaggregated model pre-established, wherein, brand mood is mood of the user for target brand, disaggregated model is the training set trained according to preset brand type of emotion, and preset brand type of emotion includes the first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th mood；The quantized value of each mood is calculated separately according to the weight that article number of the corresponding brand message of each mood and brand message are assigned in the first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th mood.The present invention solves the technical issues of tendentiousness mood that can not learn consumer for brand.
Description
Technical field
The present invention relates to brand message process fields, in particular to a kind of determination method and apparatus of brand mood.
Background technology
Currently, consumer product or brand of service when buying product and either servicing are influence its purchase one
Key factor.Therefore, analysis consumer can accurately hold the brand state in which for the mood of brand.However, at present
There are no analysis consumers for the scheme of the mood of brand, therefore, it is impossible to learn tendentiousness mood of the consumer for brand.
For above-mentioned problem, currently no effective solution has been proposed.
Invention content
An embodiment of the present invention provides a kind of determination method and apparatus of brand mood, at least to solve not learning consumption
Person for brand tendency disposition the technical issues of.
One side according to the ... of the embodiment of the present invention provides a kind of determination method of brand mood, including：Obtain target
The keyword of brand；Obtain the brand message issued for user with the associated brand message of the keyword, the brand message；
The corresponding brand mood of the brand message is determined using the disaggregated model pre-established, wherein the brand mood is described
For user for the mood of the target brand, the disaggregated model is the training trained according to preset brand type of emotion
Collection, the preset brand type of emotion include the first mood, the second mood, third mood, the 4th mood, the 5th mood and the
Six moods；According to first mood, second mood, the third mood, the 4th mood, the 5th mood and
The weight that article number of the corresponding brand message of each mood and the brand message are assigned in 6th mood is counted respectively
The quantized value of each mood is calculated, the quantized value is used to indicate the degree of its corresponding Emotion expression.
Further, before determining the corresponding brand mood of the brand message using the disaggregated model pre-established,
The method further includes：Obtain the brand message for establishing the disaggregated model；Extraction is described for establishing the classification mould
Keyword in the brand message of type for showing emotion；According to the preset brand type of emotion to described for expressing feelings
The keyword of sense is trained, and obtains the disaggregated model.
Further, after obtaining for establishing the brand message of the disaggregated model, and extraction is described for building
Before founding the keyword for showing emotion in the brand message of the disaggregated model, the method further includes：Described in judgement
Whether brand message is identifiable text message；If it is judged that the brand message is not identifiable text message, then
It converts the brand message to the identifiable text message, the identifiable text message after conversion is divided
Word；If it is judged that the brand message is identifiable text message, then directly the brand message is segmented.
Further, if it is judged that the brand message is not identifiable text message, then by the brand message
Being converted into the identifiable text message includes：If the brand message is audio-frequency information, utilize speech recognition by institute
It states audio-frequency information and is converted into the identifiable text message；If the brand message is video information, from the video
Audio-frequency information is extracted in information, and the audio-frequency information of extraction is converted to the identifiable text message using speech recognition.
Further, after determining the corresponding brand mood of the brand message using the disaggregated model pre-established,
The method further includes：Obtain the location information when brand message publication；The brand is determined based on the location information
Area where information；The brand mood of each department is determined according to the corresponding brand mood of the brand message.
Further, first mood is satisfaction, and second mood is disappointment, and the third mood is to avoid, institute
It is indignation to state the 4th mood, and the 5th mood is dislike, and the 6th mood is to know.
Another aspect according to the ... of the embodiment of the present invention additionally provides a kind of determining device of brand mood, including：First obtains
Unit is taken, the keyword for obtaining target brand；Second acquisition unit is believed for obtaining with the associated brand of the keyword
Breath, the brand message are the brand message of user's publication；First determination unit, for true using the disaggregated model pre-established
Determine the corresponding brand mood of the brand message, wherein the brand mood is feelings of the user for the target brand
Thread, the disaggregated model are the training set trained according to preset brand type of emotion, the preset brand mood class
Type includes the first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th mood；Computing unit is used for root
According to first mood, second mood, the third mood, the 4th mood, the 5th mood and the described 6th
The weight that the item number of the corresponding brand message of each mood and the brand message are assigned in mood calculates separately each
The quantized value of mood, the quantized value are used to indicate the degree of its corresponding Emotion expression.
Further, described device further includes：Third acquiring unit, for being determined using the disaggregated model pre-established
Before the corresponding brand mood of the brand message, the brand message for establishing the disaggregated model is obtained；Extraction unit is used
Keyword in the extraction brand message for establishing the disaggregated model for showing emotion；Training unit is used for
The keyword for showing emotion is trained according to the preset brand type of emotion, obtains the classification mould
Type.
Further, described device further includes：Judging unit, for obtaining the brand for establishing the disaggregated model
The keyword for showing emotion after information, and in the extraction brand message for establishing the disaggregated model it
Before, judge whether the brand message is identifiable text message；Participle unit, for if it is judged that the brand message
It is not identifiable text message, then converts the brand message to the identifiable text message, to the institute after conversion
Identifiable text message is stated to be segmented；If it is judged that the brand message is identifiable text message, then it is directly right
The brand message is segmented.
Further, the participle unit includes：First conversion module, if believed for audio for the brand message
Breath, then convert the audio-frequency information to the identifiable text message using speech recognition；Second conversion module, for such as
Brand message described in fruit is video information, then extracts audio-frequency information from the video information, using speech recognition by extraction
Audio-frequency information is converted into the identifiable text message.
Further, described device further includes：4th acquiring unit, for being determined using the disaggregated model pre-established
After the corresponding brand mood of the brand message, the location information when brand message publication is obtained；Second determination unit,
For determining the area where the brand message based on the location information；Third determination unit, for according to the brand
The corresponding brand mood of information determines the brand mood of each department.
In embodiments of the present invention, it by obtaining the keyword of target brand, obtains and believes with the associated brand of the keyword
Breath, the corresponding brand mood of brand message is determined using the disaggregated model pre-established, according to the first mood, the second mood, the
Article number and brand message of the corresponding brand message of each mood in three moods, the 4th mood, the 5th mood and the 6th mood
The weight assigned calculates separately the quantized value of each mood, to solve the tendency that can not learn consumer for brand
The problem of disposition thread.
Description of the drawings
Attached drawing described herein is used to provide further understanding of the present invention, and is constituted part of this application, this hair
Bright illustrative embodiments and their description are not constituted improper limitations of the present invention for explaining the present invention.In the accompanying drawings：
Fig. 1 is the flow chart of the determination method of brand mood according to the ... of the embodiment of the present invention；
Fig. 2 is the schematic diagram of the determining device of brand mood according to the ... of the embodiment of the present invention.
Specific implementation mode
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
The embodiment of a part of the invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
The every other embodiment that member is obtained without making creative work should all belong to the model that the present invention protects
It encloses.
It should be noted that term " first " in description and claims of this specification and above-mentioned attached drawing, "
Two " etc. be for distinguishing similar object, without being used to describe specific sequence or precedence.It should be appreciated that using in this way
Data can be interchanged in the appropriate case, so as to the embodiment of the present invention described herein can in addition to illustrating herein or
Sequence other than those of description is implemented.In addition, term " comprising " and " having " and their any deformation, it is intended that cover
It includes to be not necessarily limited to for example, containing the process of series of steps or unit, method, system, product or equipment to cover non-exclusive
Those of clearly list step or unit, but may include not listing clearly or for these processes, method, product
Or the other steps or unit that equipment is intrinsic.
According to embodiments of the present invention, a kind of determination embodiment of the method for brand mood is provided, it should be noted that attached
The step of flow of figure illustrates can execute in the computer system of such as a group of computer-executable instructions, though also,
So logical order is shown in flow charts, but in some cases, it can be with different from shown by sequence execution herein
Or the step of description.
Fig. 1 is the flow chart of the determination method of brand mood according to the ... of the embodiment of the present invention, as shown in Figure 1, this method packet
Include following steps：
Step S102 obtains the keyword of target brand.
Step S104 is obtained and the associated brand message of keyword.Brand message is the brand message of user's publication.
Step S106 determines the corresponding brand mood of brand message, wherein brand feelings using the disaggregated model pre-established
Thread be user for the mood of target brand, disaggregated model is the training set trained according to preset brand type of emotion,
Preset brand type of emotion includes the first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th mood.
Step S108, according in the first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th mood
The weight that the item number and brand message of the corresponding brand message of each mood are assigned calculates separately the quantization of each mood
Value.Quantized value is used to indicate the degree of its corresponding Emotion expression.
Based on the weight that the item number and brand message of the brand message of each mood are assigned, to calculate corresponding mood
To the quantized value being in a bad mood.Wherein, quantized value can indicate the degree of the performance of corresponding mood, and the value is bigger, right
The mood answered is stronger；The weight that brand message is assigned then indicates influence power of the brand message to the quantized value of brand mood,
If the weight of brand message is bigger, influence of the brand message to its corresponding brand mood is bigger.
In the embodiment of the present invention, according to preset brand type of emotion (including the first mood, the second mood, third
Mood, the 4th mood, the 5th mood and the 6th mood) it is trained, training pattern is obtained, can be identified by the training pattern
Brand mood included in the brand message of user's publication, that is to say the corresponding brand mood of brand message, so that it is determined that with
The mood at family is the first mood either the second mood either third mood either the 4th mood or the 5th mood, Huo Zhe
Six moods.
By obtaining the keyword of target brand, acquisition and the associated brand message of the keyword, utilize what is pre-established
Disaggregated model determines the corresponding brand mood of brand message, according to the first mood, the second mood, third mood, the 4th mood,
Article number of the corresponding brand message of each mood and the weight of imparting calculate separately each feelings in five moods and the 6th mood
The quantized value of thread, to solve the problems, such as the tendency disposition that can not learn consumer for brand.Further, due to passing through
The disaggregated model that training obtains can identify that brand mood included in brand message includes a variety of, the brand feelings identified
Thread is more diversified, to accurately reflect brand mood of the user to target brand.All types of mood quantity is carried out
Statistics calculates, and can obtain the tendentiousness of the mood on network to target brand.
Preferably, above-mentioned first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th mood can be with
It is followed successively by satisfied, disappointed, avoidance, indignation, dislikes and knows.Above-mentioned six kinds of moods are corresponding in turn to the quantization obtained by quantization
Value：Satisfaction, disappointment degree, avoidance degree, indignation degree, dislike degree and awareness degree are orderly used to the performance for indicating above-mentioned six kinds of moods
Degree.
Preferably, it can determine its weights according to the source of every information, reuse computation model each is calculated
The quantized value of mood.Wherein, when the information of separate sources calculates, due to its weighted, corresponding weights are different, for example, coming
The weight of the information of automatic network media can be higher than the weight from personal information, and the weights of separate sources can train
It is determined in journey, so as to determine the weights of each information by training pattern when calculating.
For example, in designated time period, the relevant brand message of some brand first does mood with disaggregated model to it
Analysis, determines the corresponding brand mood of every information, the quantized value of each mood is then calculated according to following formula：
Z=a0*b0+a1*b1+a2*b2+a2*b2+a3*b3 ...
Wherein, Z indicates quantized value, the brand message item number in the corresponding source of the expressions such as a0, a1, a2, a3, b0, b1, b2, b3
Deng the weights for indicating separate sources.
By taking the mood of " satisfaction " as an example, can first it be counted by information source.Such as：Official media 100, microblogging is big
V 200, common netizen 300.Again in disaggregated model, the weighted data in each source is read.Such as：Official media 0.8,
The big V 0.5 of microblogging, common netizen 0.2.Then, it for above-mentioned information, is weighted and summarizes.The quantized value of " satisfaction " mood
For：100*0.8+200*0.5+300*0.2=240.Above-mentioned identical mode may be used in other moods, and quantization is calculated
Value.It is finally obtained to be at the appointed time in section, to the various mood quantized values of the brand, such as：240, disappointed 200 are satisfied with,
Avoid 100, indignation 50, dislike 300 knows 400.
It should be noted that the mood in the embodiment of the present invention can also include the 7th mood and/or the 8th mood etc..Separately
Outside, when then calculating the quantized value of each mood, it is also assumed that the weights of all brand messages are all 1, then each feelings
The quantized value of thread can be the item number of the corresponding brand message of the mood.Above-mentioned example is just for the sake of the description embodiment of the present invention
Technical principle, do not have improper restriction to the present invention.
Preferably, in the embodiment of the present invention, the indexs such as area, gender, authenticating identity is can be combined with and be calculated accordingly
Quantized value, in this way, can be analyzed brand mood by these indexs.
Preferably, the brand message in the embodiment of the present invention can be user issued on network blog, microblogging, friend
Enclose message, the model using in community's message (such as QQ space), forum, comment.Brand message can by web crawlers,
The modes such as automatic script, manually input obtain.The concrete form of brand message can be text message, can also be audio letter
Breath, can also be video information etc..
Preferably, before determining the corresponding brand mood of brand message using the disaggregated model pre-established, method is also
Including：Obtain the brand message for establishing disaggregated model；It extracts in the brand message for establishing disaggregated model for expressing
The keyword of emotion；The keyword for showing emotion is trained according to preset brand type of emotion, obtains classification mould
Type.
Brand message for establishing disaggregated model can be obtained by modes such as web crawlers, automatic script, manually inputs
It takes, since the brand message is for establishing disaggregated model, the data volume of the brand message of acquisition is relatively large, so as to
The higher disaggregated model of accuracy is identified in training.
After getting a large amount of brand message, keyword of the extraction for showing emotion from the brand message, or
Person's key message, such as " this brand is very good ", " what brand this is ", " good expensive " etc., then according to preset network feelings
Thread classification (including satisfaction, disappointment, avoidance, indignation, dislike and know) is trained keyword, obtains above-mentioned disaggregated model.
After obtaining disaggregated model, Emotion identification or classification can be carried out to the brand message newly inputted using the disaggregated model.
Preferably, after obtaining the brand message for establishing disaggregated model, and extraction is for establishing disaggregated model
Brand message in the keyword for showing emotion before, method further includes：Judge whether brand message is identifiable
Text message；If it is judged that brand message is not identifiable text message, then brand message is converted to identifiable text
This information segments the identifiable text message after conversion；If it is judged that brand message is identifiable text envelope
Breath, then directly segment brand message.
In the present embodiment, since the brand message got can be information (including text message, the audio of diversified forms
Information, video information), and the extraction of usually keyword is extracted from identifiable text message, therefore, is used for getting
It establishes after the brand message of disaggregated model, and extracts and be used to show emotion in the brand message for establishing disaggregated model
Keyword before, can first judge whether the brand message got is identifiable text message, if it is, can be straight
It connects and the brand message is segmented, in order to therefrom extract keyword；If it is not, then converting the brand message to recognizable
Text message, then segmented, in order to extract keyword from the text message after conversion.
It should be noted that in the embodiment of the present invention, after obtaining disaggregated model, the brand message newly inputted is carried out
When classification or identification, it can also first judge whether the brand message is identifiable text message, in order to quickly determine product
The corresponding brand mood of board information.
Further, if it is judged that brand message is not identifiable text message, then converting brand message to can
The text message of identification includes：If brand message is audio-frequency information, using speech recognition converts and can know audio-frequency information to
Other text message；If brand message is video information, audio-frequency information is extracted from video information, it will using speech recognition
The audio-frequency information of extraction is converted into identifiable text message.
In the present embodiment, it for audio-frequency information, can identify speech recognition technology, convert voice messaging to text message
Keyword is extracted from text message again；For video information, then the audio-frequency information in the video information can be first extracted, then
It is handled using the transform mode for audio-frequency information.
It should be noted that in the embodiment of the present invention, various brand messages can be the information using various language, example
Such as, Chinese, English, Japanese etc., for different language, can be converted into identical identifiable text message.
Preferably, after determining the corresponding brand mood of brand message using the disaggregated model pre-established, method is also
Including：Obtain location information when brand message publication；The area where brand message is determined based on location information；According to brand
The corresponding brand mood of information determines the brand mood of each department.
The location information where it usually can be all carried in the brand message issued due to user, obtain position letter
Breath, and determine the area where it, this area can be the administrative region divided as unit of province, city etc., then according to determination
The corresponding brand mood of brand message that goes out determines the brand mood of each department.Preferably, if in some regional brand
Mood include it is a variety of in above-mentioned brand mood (including satisfaction, disappointment, avoidance, indignation, dislike and know), can be by institute's accounting
Brand mood of the maximum mood of weight as this area.
For example, for a certain brand, the people in area in all parts of the country can have different moods, by the leading feelings in each area
Mood of the thread as this area may thereby determine that out tendentiousness mood of each department to the brand.
In the embodiment of the present invention, for the associated brand message of keyword, brand message and target brand can be utilized
The degree of association determine that can specifically calculate the degree of association of brand message and the keyword of target brand, the degree of association is more than pre-
If threshold value, it is determined that the brand message be and the associated brand message of keyword.
The embodiment of the present invention additionally provides a kind of determining device of brand mood, which can be used for executing of the invention real
Apply the determination method of the brand mood of example.As shown in Fig. 2, the determining device of the brand mood includes：10 He of first acquisition unit
First determination unit 20, the first determination unit 30 and computing unit 40.
First acquisition unit 10 is used to obtain the keyword of target brand.
Second acquisition unit 20 is for obtaining and the associated brand message of keyword.Brand message is the brand of user's publication
Information.
First determination unit 30 is used to that the disaggregated model pre-established to be utilized to determine the corresponding brand mood of brand message,
In, brand mood is mood of the user for target brand, and disaggregated model is to train to obtain according to preset brand type of emotion
Training set, preset brand type of emotion include the first mood, the second mood, third mood, the 4th mood, the 5th mood and
6th mood.
Computing unit 40 is used for according to the first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th
The weight that the item number of the corresponding brand message of each mood and brand message are assigned in mood calculates separately each mood
Quantized value.Quantized value is used to indicate the degree of its corresponding Emotion expression.
Based on the weight that the item number and brand message of the brand message of each mood are assigned, to calculate corresponding product
The cards one holds thread is to the quantized value be in a bad mood.Wherein, quantized value can indicate the degree of the performance of corresponding brand mood, should
Value is bigger, and corresponding brand mood is stronger；The weight that brand message is assigned then indicates brand message to brand mood
The influence power of quantized value, if the weight of brand message is bigger, influence of the brand message to its corresponding brand mood is bigger.
In the embodiment of the present invention, according to preset brand type of emotion (including the first mood, the second mood, third
Mood, the 4th mood, the 5th mood and the 6th mood) it is trained, training pattern is obtained, can be identified by the training pattern
Brand mood included in the brand message of user's publication, that is to say the corresponding brand mood of brand message, so that it is determined that with
The mood at family is the first mood either the second mood either third mood either the 4th mood or the 5th mood.
By obtaining the keyword of target brand, acquisition and the associated brand message of the keyword, utilize what is pre-established
Disaggregated model determines the corresponding brand mood of brand message, according to the first mood, the second mood, third mood, the 4th mood,
Article number of the corresponding brand message of each mood and the weight of imparting calculate separately each feelings in five moods and the 6th mood
The quantized value of thread, to solve the tendentiousness mood that can not learn consumer for brand.Further, due to by training
Obtained disaggregated model can identify that brand mood included in brand message includes a variety of, and the brand mood identified is more
Add diversification, thus the problem of accurately reflecting brand mood of the user to target brand.To all types of mood quantity into
Row statistics calculates, and can obtain the tendentiousness of the mood on network to target brand.
Preferably, above-mentioned first mood, the second mood, third mood, the 4th mood, the 5th mood and the 6th mood can be with
It is followed successively by satisfied, disappointed, avoidance, indignation, dislikes and knows.
Preferably, it can determine its weights according to the source of every brand message, reuse computation model to be calculated
The quantized value of each mood.Wherein, when the brand message of separate sources calculates, due to its weighted, corresponding weights are different,
For example, the weight of the brand message from the network media can be higher than the weight from personal brand message, separate sources
Weights can determine in the training process, so as to determine the power of each brand message by training pattern when calculating
Value.
For example, in designated time period, the relevant brand message of some brand first does mood with disaggregated model to it
Analysis, determines the corresponding brand mood of every information, the quantized value of each mood is then calculated according to following formula：
Z=a0*b0+a1*b1+a2*b2+a2*b2+a3*b3 ...
Wherein, Z indicates quantized value, the brand message item number in the corresponding source of the expressions such as a0, a1, a2, a3, b0, b1, b2, b3
Deng the weights for indicating separate sources.
By taking the mood of " satisfaction " as an example, can first it be counted by information source.Such as：Official media 100, microblogging is big
V 200, common netizen 300.Again in disaggregated model, the weighted data in each source is read.Such as：Official media 0.8,
The big V 0.5 of microblogging, common netizen 0.2.Then, it for above-mentioned information, is weighted and summarizes.The quantized value of " satisfaction " mood
For：100*0.8+200*0.5+300*0.2=240.Above-mentioned identical mode may be used in other moods, and quantization is calculated
Value.It is finally obtained to be at the appointed time in section, to the various mood quantized values of the brand, such as：240, disappointed 200 are satisfied with,
Avoid 100, indignation 50, dislike 300 knows 400.
It should be noted that the mood in the embodiment of the present invention can also include the 7th mood and/or the 8th mood etc..Separately
Outside, when then calculating the quantized value of each mood, it is also assumed that the weights of all brand messages are all 1, then each feelings
The quantized value of thread can be the item number of the corresponding brand message of the mood.Above-mentioned example is just for the sake of the description embodiment of the present invention
Technical principle, do not have improper restriction to the present invention.
Preferably, in the embodiment of the present invention, the indexs such as area, gender, authenticating identity is can be combined with and be calculated accordingly
Quantized value, in this way, can be analyzed brand mood by these indexs.
Preferably, the brand message in the embodiment of the present invention can be user issued on network blog, microblogging, friend
Enclose message, the model using in community's message (such as QQ space), forum, comment.Brand message can by web crawlers,
The modes such as automatic script, manually input obtain.The concrete form of brand message can be text message, can also be audio letter
Breath, can also be video information etc..
In the embodiment of the present invention, the brand message of user's publication can be the mood for user's current time, also may be used
To be for some event, the mood of some things.Therefore, the scheme of the embodiment of the present invention can be used in network to target product
The brand mood of board either things can specifically obtain with target brand or the relevant brand message of things, utilize classification mould
The brand message got is identified in type, the corresponding brand mood of the brand message is determined, so as to count net
A variety of different moods of the network user to target brand or things.
Preferably, device further includes：Third acquiring unit, for determining that brand is believed using the disaggregated model pre-established
Before ceasing corresponding brand mood, the brand message for establishing disaggregated model is obtained；Extraction unit, for extracting for establishing
Keyword in the brand message of disaggregated model for showing emotion；Training unit, for according to preset brand type of emotion
Keyword for showing emotion is trained, disaggregated model is obtained.
Brand message for establishing disaggregated model can be obtained by modes such as web crawlers, automatic script, manually inputs
It takes, since the brand message is for establishing disaggregated model, the data volume of the brand message of acquisition is relatively large, so as to
The higher disaggregated model of accuracy is identified in training.
After getting a large amount of brand message, keyword of the extraction for showing emotion from the brand message, or
Person's key message, such as " this brand is very good ", " what brand this is ", " good expensive " etc., then according to preset network feelings
Thread classification (including satisfaction, disappointment, avoidance, indignation, dislike and know) is trained keyword, obtains above-mentioned disaggregated model.
After obtaining disaggregated model, Emotion identification or classification can be carried out to the brand message newly inputted using the disaggregated model.
Preferably, device further includes：Judging unit, for after obtaining brand message for establishing disaggregated model,
And before the keyword for showing emotion in brand message of the extraction for establishing disaggregated model, judge that brand message is
No is identifiable text message；Participle unit, for if it is judged that brand message is not identifiable text message, then will
Brand message is converted into identifiable text message, is segmented to the identifiable text message after conversion；If it is judged that
Brand message is identifiable text message, then is directly segmented to brand message.
In the present embodiment, since the brand message got can be information (including text message, the audio of diversified forms
Information, video information), and the extraction of usually keyword is extracted from identifiable text message, therefore, is used for getting
It establishes after the brand message of disaggregated model, and extracts and be used to show emotion in the brand message for establishing disaggregated model
Keyword before, can first judge whether the brand message got is identifiable text message, if it is, can be straight
It connects and the brand message is segmented, in order to therefrom extract keyword；If it is not, then converting the brand message to recognizable
Text message, then segmented, in order to extract keyword from the text message after conversion.
It should be noted that in the embodiment of the present invention, after obtaining disaggregated model, the brand message newly inputted is carried out
When classification or identification, it can also first judge whether the brand message is identifiable text message, in order to quickly determine product
The corresponding brand mood of board information.
Preferably, participle unit includes：First conversion module utilizes language if being audio-frequency information for brand message
Sound identification converts audio-frequency information to identifiable text message；Second conversion module, if believed for video for brand message
Breath, then extract audio-frequency information from video information, the audio-frequency information of extraction converted to identifiable text using speech recognition
Information.
In the present embodiment, it for audio-frequency information, can identify speech recognition technology, convert voice messaging to text message
Keyword is extracted from text message again；For video information, then the audio-frequency information in the video information can be first extracted, then
It is handled using the transform mode for audio-frequency information.
It should be noted that in the embodiment of the present invention, various brand messages can be the information using various language, example
Such as, Chinese, English, Japanese etc., for different language, can be converted into identical identifiable text message.
Preferably, device further includes：4th acquiring unit, for determining that brand is believed using the disaggregated model pre-established
After ceasing corresponding brand mood, location information when brand message publication is obtained；Second determination unit, for being believed based on position
Breath determines the area where brand message；Third determination unit, for determining various regions according to the corresponding brand mood of brand message
The brand mood in area.
The location information where it usually can be all carried in the brand message issued due to user, obtain position letter
Breath, and determine the area where it, this area can be the administrative region divided as unit of province, city etc., then according to determination
The corresponding brand mood of brand message that goes out determines the brand mood of each department.Preferably, if in some regional brand
Mood include it is a variety of in above-mentioned brand mood (including satisfaction, disappointment, avoidance, indignation, dislike and know), can be by institute's accounting
Brand mood of the maximum mood of weight as this area.
For example, for a certain brand, the people in area in all parts of the country can have different moods, by the leading feelings in each area
Mood of the thread as this area may thereby determine that out tendentiousness mood of each department to the brand.
The embodiments of the present invention are for illustration only, can not represent the quality of embodiment.
In the above embodiment of the present invention, all emphasizes particularly on different fields to the description of each embodiment, do not have in some embodiment
The part of detailed description may refer to the associated description of other embodiment.
In several embodiments provided herein, it should be understood that disclosed technology contents can pass through others
Mode is realized.Wherein, the apparatus embodiments described above are merely exemplary, for example, the unit division, Ke Yiwei
A kind of division of logic function, formula that in actual implementation, there may be another division manner, such as multiple units or component can combine or
Person is desirably integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Between coupling, direct-coupling or communication connection can be INDIRECT COUPLING or communication link by some interfaces, unit or module
It connects, can be electrical or other forms.
The unit illustrated as separating component may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, you can be located at a place, or may be distributed over multiple
On unit.Some or all of unit therein can be selected according to the actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, it can also
It is that each unit physically exists alone, it can also be during two or more units be integrated in one unit.Above-mentioned integrated list
The form that hardware had both may be used in member is realized, can also be realized in the form of SFU software functional unit.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can be stored in a computer read/write memory medium.Based on this understanding, technical scheme of the present invention is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
Equipment (can be personal computer, server or network equipment etc.) execute each embodiment the method for the present invention whole or
Part steps.And storage medium above-mentioned includes：USB flash disk, read-only memory (ROM, Read-Only Memory), arbitrary access are deposited
Reservoir (RAM, Random Access Memory), mobile hard disk, magnetic disc or CD etc. are various can to store program code
Medium.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered
It is considered as protection scope of the present invention.
Claims (9)
1. a kind of determination method of brand mood, which is characterized in that including：
Obtain the keyword of target brand；
Obtain the brand message issued for user with the associated brand message of the keyword, the brand message；
The corresponding brand mood of the brand message is determined using the disaggregated model pre-established, wherein the brand mood is
The user for the mood of the target brand, according to preset brand type of emotion train to obtain by the disaggregated model
Training set, the preset brand type of emotion include the first mood, the second mood, third mood, the 4th mood, the 5th mood
With the 6th mood；
According to first mood, second mood, the third mood, the 4th mood, the 5th mood and institute
The weight that article number of the corresponding brand message of each mood and the brand message are assigned in the 6th mood is stated to calculate separately
The quantized value of each mood, the quantized value are used to indicate the degree of its corresponding Emotion expression, wherein the weight root
It is determined according to the source of the brand message；
Wherein, the method further includes：Obtain the location information when brand message publication；It is determined based on the location information
Area where the brand message；The brand mood of each department is determined according to the corresponding brand mood of the brand message；
Wherein, it obtains with the associated brand message of the keyword and includes：Calculate brand message and the keyword of target brand
The degree of association；When the degree of association of brand message and the keyword of target brand is more than predetermined threshold value, it is determined that the brand message be with
The associated brand message of keyword.
2. according to the method described in claim 1, it is characterized in that, determining the brand using the disaggregated model pre-established
Before the corresponding brand mood of information, the method further includes：
Obtain the brand message for establishing the disaggregated model；
Keyword in the extraction brand message for establishing the disaggregated model for showing emotion；
The keyword for showing emotion is trained according to the preset brand type of emotion, obtains the classification
Model.
3. according to the method described in claim 2, it is characterized in that, obtaining the brand message for establishing the disaggregated model
Later, and before the keyword for showing emotion in the extraction brand message for establishing the disaggregated model,
The method further includes：
Judge whether the brand message is identifiable text message；
If it is judged that the brand message is not identifiable text message, then converting the brand message to described can know
Other text message segments the identifiable text message after conversion；
If it is judged that the brand message is identifiable text message, then directly the brand message is segmented.
4. according to the method described in claim 3, it is characterized in that, if it is judged that the brand message is not identifiable text
This information, then converting the brand message to the identifiable text message includes：
If the brand message is audio-frequency information, converted the audio-frequency information to using speech recognition described identifiable
Text message；
If the brand message is video information, audio-frequency information is extracted from the video information, it will using speech recognition
The audio-frequency information of extraction is converted into the identifiable text message.
5. method according to claim 1 to 4, which is characterized in that first mood is satisfaction, described the
Two moods are disappointment, and the third mood is to avoid, and the 4th mood is indignation, and the 5th mood is dislike, described the
Six moods are to know.
6. a kind of determining device of brand mood, which is characterized in that including：
First acquisition unit, the keyword for obtaining target brand；
Second acquisition unit is issued with the associated brand message of the keyword, the brand message for user for obtaining
Brand message；
First determination unit, for determining the corresponding brand mood of the brand message using the disaggregated model pre-established,
In, the brand mood is mood of the user for the target brand, and the disaggregated model is according to preset brand
The training set that type of emotion is trained, the preset brand type of emotion include the first mood, the second mood, third feelings
Thread, the 4th mood, the 5th mood and the 6th mood；
Computing unit, for according to first mood, second mood, the third mood, the 4th mood, described
Article number of the corresponding brand message of each mood and the brand message are assigned in 5th mood and the 6th mood
Weight calculates separately the quantized value of each mood, and the quantized value is used to indicate the degree of its corresponding Emotion expression,
In, the weight is determined according to the source of the brand message；
Wherein, described device further includes：4th acquiring unit, for determining the brand using the disaggregated model pre-established
After the corresponding brand mood of information, the location information when brand message publication is obtained；Second determination unit, for being based on
The location information determines the area where the brand message；Third determination unit, for being corresponded to according to the brand message
Brand mood determine the brand moods of each department；
Wherein, second acquisition unit is additionally operable to：Calculate the degree of association of brand message and the keyword of target brand；Work as brand message
The degree of association with the keyword of target brand is more than predetermined threshold value, it is determined that the brand message is to believe with the associated brand of keyword
Breath.
7. device according to claim 6, which is characterized in that described device further includes：
Third acquiring unit, for using the disaggregated model that pre-establishes determine the corresponding brand mood of the brand message it
Before, obtain the brand message for establishing the disaggregated model；
Extraction unit, for extracting the key in the brand message for establishing the disaggregated model for showing emotion
Word；
Training unit, for being instructed to the keyword for showing emotion according to the preset brand type of emotion
Practice, obtains the disaggregated model.
8. device according to claim 7, which is characterized in that described device further includes：
Judging unit, for being used to build after obtaining for establishing the brand message of the disaggregated model, and described in extraction
Stand the disaggregated model brand message in the keyword for showing emotion before, judge the brand message whether be can
The text message of identification；
Participle unit, for if it is judged that the brand message is not identifiable text message, then by the brand message
It is converted into the identifiable text message, the identifiable text message after conversion is segmented；If it is judged that
The brand message is identifiable text message, then is directly segmented to the brand message.
9. device according to claim 8, which is characterized in that the participle unit includes：
First conversion module utilizes speech recognition by the audio-frequency information if being audio-frequency information for the brand message
It is converted into the identifiable text message；
Second conversion module extracts audio letter if being video information for the brand message from the video information
Breath, the audio-frequency information of extraction is converted to the identifiable text message using speech recognition.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510406454.3A CN104951434B (en) | 2015-07-10 | 2015-07-10 | The determination method and apparatus of brand mood |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510406454.3A CN104951434B (en) | 2015-07-10 | 2015-07-10 | The determination method and apparatus of brand mood |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104951434A CN104951434A (en) | 2015-09-30 |
| CN104951434B true CN104951434B (en) | 2018-10-23 |
Family
ID=54166097
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510406454.3A Active CN104951434B (en) | 2015-07-10 | 2015-07-10 | The determination method and apparatus of brand mood |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104951434B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106294779B (en) * | 2016-08-12 | 2020-03-17 | 杭州一来二去广告有限公司 | Personal brand label generation method and system |
| CN109118626B (en) * | 2018-08-08 | 2022-09-13 | 腾讯科技（深圳）有限公司 | Lock control method and device, storage medium and electronic device |
| CN109359246A (en) * | 2018-12-07 | 2019-02-19 | 上海宏原信息科技有限公司 | A kind of brand cohesion calculation method based on forum user speech |
| CN109697472B (en) * | 2018-12-28 | 2021-05-04 | 泰州市津达电子科技有限公司 | Sub-emotion marking-in method |
| CN110458615A (en) * | 2019-07-31 | 2019-11-15 | 沃民高新科技（北京）股份有限公司 | Pass through the method and device of internet information assessment customer satisfaction |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101763401B (en) * | 2009-12-30 | 2012-05-30 | 暨南大学 | Network public sentiment hotspot prediction and analysis method |
| CN104765733B (en) * | 2014-01-02 | 2018-06-15 | 华为技术有限公司 | A kind of method and apparatus of social networks event analysis |
| CN104750674B (en) * | 2015-02-17 | 2018-12-21 | 北京京东尚科信息技术有限公司 | A kind of man-machine conversation's satisfaction degree estimation method and system |
-
2015
- 2015-07-10 CN CN201510406454.3A patent/CN104951434B/en active Active
Non-Patent Citations (2)
| Title |
|---|
| 基于Web文本挖掘的企业口碑情感分类模型研究;孙莹;《中国优秀硕士学位论文全文数据库信息科技辑》;20140215(第2期);参见第4-6章 * |
| 突发事件网络舆情分析与威胁估计方法研究;王铁套;《中国优秀硕士学位论文全文数据库信息科技辑》;20130615(第6期);第I139-219页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104951434A (en) | 2015-09-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104951807B (en) | The determination method and apparatus of stock market's mood | |
| CN104951434B (en) | The determination method and apparatus of brand mood | |
| CN106919661B (en) | Emotion type identification method and related device | |
| CN108446813A (en) | A kind of method of electric business service quality overall merit | |
| CN106815192B (en) | Model training method and device and sentence emotion recognition method and device | |
| CN109918560A (en) | A kind of answering method and device based on search engine | |
| CN105095415A (en) | Method and apparatus for confirming network emotion | |
| CN110598070B (en) | Application type identification method and device, server and storage medium | |
| CN110008984B (en) | Target fraud transaction model training method and device based on multitasking samples | |
| CN111680147A (en) | Data processing method, device, equipment and readable storage medium | |
| CN103729474A (en) | Method and system for identifying vest account numbers of forum users | |
| CN106951571A (en) | A kind of method and apparatus for giving application mark label | |
| CN116861258B (en) | Model processing method, device, equipment and storage medium | |
| CN110138794A (en) | A kind of counterfeit website identification method, device, equipment and readable storage medium storing program for executing | |
| CN110136748A (en) | A kind of rhythm identification bearing calibration, device, equipment and storage medium | |
| CN108304387A (en) | The recognition methods of noise word, device, server group and storage medium in text | |
| CN109857861A (en) | File classification method, device, server and medium based on convolutional neural networks | |
| CN107330009A (en) | Descriptor disaggregated model creation method, creating device and storage medium | |
| CN108090201A (en) | A kind of method, apparatus and electronic equipment of article content classification | |
| CN107665202B (en) | Method and device for constructing interest model and electronic equipment | |
| CN110879832A (en) | Target text detection method, model training method, device and equipment | |
| CN111159377A (en) | Attribute recall model training method and device, electronic equipment and storage medium | |
| CN114548119A (en) | Test set generation method, test method, device, equipment and medium | |
| CN113609865A (en) | Text emotion recognition method and device, electronic equipment and readable storage medium | |
| CN113591881A (en) | Intention recognition method and device based on model fusion, electronic equipment and medium |
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 |