Product comparison relation determination method and device and computer storage medium

A technology for comparing relationships and products, applied in computing, relational databases, structured data retrieval, etc., can solve problems such as low efficiency and accuracy, and achieve the effect of improving accuracy and efficiency

Pending Publication Date: 2022-01-07

芜湖汽车前瞻技术研究院有限公司 +1

0 Cites 0 Cited by

AI-Extracted Technical Summary

Problems solved by technology

[0004] The embodiment of the present application provides a method, device and computer storage medium for determining product comparison relation...

Method used

In the embodiment of the present application, the terminal can obtain comment information, and determine the entity corresponding to the product word in the comment information through the product knowledge base built, and integrate the emotional thesaurus at the same time, and add negative words, degree adverbs, etc. to compare the strength Value calculation, and then the product comparison relationship can be determined by comparing the intensity value, so that the comparison relationship between products in the review information can be effectively identified without the need for staff to browse all the...

Abstract

The embodiment of the invention discloses a product comparison relation determination method and device and a computer storage medium, and belongs to the technical field of data mining. The method comprises the steps: obtaining at least one product word, a candidate entity corresponding to the at least one product word and emotion expression information for the at least one product word in comment information, wherein the at least one product word is used for describing products of the same type, and the candidate entity corresponding to the at least one product word has an association relation in attribute; determining a comparison intensity value of the candidate entity corresponding to the at least one product word according to the candidate entity corresponding to the at least one product word and the emotion expression information for the at least one product word; and according to the comparison strength value of the candidate entity corresponding to the at least one product word, determining a product comparison relationship of the candidate entity corresponding to the at least one product word. According to the embodiment of the invention, the product comparison relationship is determined by identifying the emotion expression information in the comment information, so that the accuracy and efficiency of determining the product comparison relationship are improved.

Application Domain

Relational databasesNatural language data processing +2

Technology Topic

Data miningData science

Image

Examples

Experimental program(1)

Example Embodiment

[0059] In order to make the object, technical solutions, and advantages of the present application, the embodiments will be further described in detail below with reference to the accompanying drawings.

[0060] The present application embodiment provides an application scenario before explaining the description of the present application embodiment.

[0061] Since the user will choose the view on the product after purchasing the product, and compare the product with other similar products, and the product company is usually able to excavate the comparison between products from the product view from the user. Thus, potential competitors and their product performance differences, and thus improve the product and obtain competitive advantages. However, due to the current development of the development of the information society, the employees of the product have not had enough time and energy to browse all information, and effectively identify the comparison between products in the browsing information, thereby reducing the determination of product comparison The accuracy and efficiency of the relationship.

[0062] Based on such application scenarios, the application example provides a method of determining a comparative relationship that enhances the accuracy and efficiency of the comparison relationship of the product.

[0063] figure 1 It is a method of determining a method of determining a comparative relationship provided by the embodiment of the present application, and the method of determining the relationship of the product can include the following steps:

[0064] Step 101: Get at least one product word in the comment information, the corresponding candidate entity, and emotional expression information for the at least one product word, the at least one product word is used to describe the same type of product, the at least one The candidate entity corresponding to the product word is associated with an association relationship.

[0065] Step 102: The comparison intensity value of the candidate entity corresponding to the at least one product word is determined based on the candidate entity corresponding to the at least one product word, and the emotional expression information of the at least one product word.

[0066] Step 103: A comparison relationship of the candidate entity corresponding to the candidate entity corresponding to the candidate entity according to the candidate entity corresponding to the at least one product word.

[0067] In the present application embodiment, the emotional expression information corresponding to at least one product word in the comment information can be identified, and the corresponding comparison intensity value can be determined by emotional expression information, and the comparison intensity value can determine the comparison relationship, so that there is no need to browse all. The comment information can effectively identify the comparison relationship between the products in the review information, which in turn improves the accuracy and efficiency of determining the comparative relationship of the product.

[0068] In some embodiments, at least one product word in the comment information is obtained, the at least one product word corresponding candidate entity, and emotional expression information for the at least one product word, including:

[0069] Treatment of this comment information to get multiple words;

[0070] The product word of at least one entity is identified from the plurality of words by the physical class identification tool, and the at least one product word is obtained, and the emotional expression information for the at least one product word is identified from the plurality of words through the emotional word library;

[0071] The candidate entity corresponding to each product word in the at least one product word is obtained from the knowledge base, which includes entity information of the same type of entity as the at least one product word.

[0072] In some embodiments, emotional expression information for this at least one product word is identified from the plurality of words through the emotional word library, including:

[0073] Through the emotional word library, the word geographical class belongs to each of the plurality of words is identified, and the meaning of emotions, degree adverbs, and noble words are included.

[0074] Statistically the number of first words that belong to this negative in the plurality of words and the number of second points belonging to the emotional word;

[0075] According to the number of the first word, the corresponding negative word characteristic of the at least one product word is determined;

[0076] Determine the degree of adverbatic adverbial adverbial reference to the particle belonging to the extent of this degree adverb, and the emotional word intensity value corresponding to the emotional word;

[0077] According to this negative characteristics, the degree adverb feature, the emotional word strength value, and the number of second points, the emotional value of the at least one product word is determined.

[0078] In some embodiments, the comparison intensity value of the candidate entity corresponding to the at least one product word is determined according to the candidate entity corresponding to the at least one product word, and the emotional expression information of the at least one product word, the comparison of the candidate entity corresponding to the candidate entity, including:

[0079] The candidate entity corresponding to the at least one product word and the corresponding emotional expression information are input to the classifier to classify the classification decision by the classifier to obtain a comparison intensity value corresponding to the candidate entity corresponding to the at least one product.

[0080] In some embodiments, the comparison of the candidate entity corresponding to the candidate entity corresponding to the candidate entity according to the candidate entity of the at least one product is determined, including:

[0081] According to the comparison intensity value of the candidate entity corresponding to the at least one product word, a comparison relationship of candidate entities corresponding to the at least one product word is plotted, which is used to describe the candidate entity corresponding to the at least one product word. Comparison relationship;

[0082] The comparison relationship between the target entity and the reference entity in the comparative relationship is performed, and the comparative relationship weight between the target entity and the reference entity is obtained.

[0083] In some embodiments, the comparison of candidate entities corresponding to the candidate entity corresponding to the candidate entity of the candidate entity is drawn according to the comparison of the candidate entity corresponding to the at least one product word, including:

[0084] When the comparison intensity value between the target entity and the reference entity is equal to 0, the target entity is connected to the reference entity by the non-harmed segment, which is the comparative entity corresponding to the at least one product word, which is the reference entity Entity for comparison in candidate entities corresponding to at least one product word;

[0085] When the comparison intensity value between the target entity and the reference entity is not equal to 0, the target entity and the reference entity are connected by a direction line segment.

[0086] All of the above alternative techniques can be combined to form an alternative embodiment of the present application, and the present application will not be described in this disclosure.

[0087] figure 2 It is a method of determining a method of determining a product comparison relationship provided by the present application, and the present embodiment is applied to the terminal in the determination method of the compare relationship of the product. The method of determining the compare relationship of the product can include the following steps. :

[0088] Step 201: Terminal acquires comments.

[0089] Since the user usually publish comment information on the relevant website after purchasing the product, the terminal can obtain the product's comment information from the website, or the staff can enter the comment information.

[0090] It should be noted that the terminal can obtain the comment information, or the terminal can acquire the comment information every specified interval every specified interval. This specified time interval can be 3 hours, 6 hours, and more.

[0091] For example, take the automotive field as an example, the terminal can obtain comment information from the automotive-related website, and the obtained comment information can be "better than Y Di's exterior."

[0092] As an example, the relevant website can be a social forum for product related official website, information exchange.

[0093] In some embodiments, the terminal can obtain the review data related to each product from the related website when obtaining the comment information, and then uses the Adaboost algorithm to clean and handle the climbed comment data for some garbage advertisements. information. This cleaning process includes removing null values (empty information), repetition values (repetitive information), and removal words, etc., deactivate words include "this", "" "and other words.

[0094] Step 202: Terminal obtains at least one product word in the comment information, at least one product word candidate entity, and emotional expression information for at least one product word.

[0095] It should be noted that at least one product term is used to describe the same type of product, at least one of the product words corresponds to an association relationship on attributes.

[0096] As an example, since the comparison relationship generally includes a comparison subject, compare object, compare attributes, comparison words, and evaluation words, but the terminal can not only obtain at least one product word in the comment information At least one product word corresponding candidate entity, and emotional expression information for at least one product word, it is also possible to obtain comparison words, compare properties, etc. in the comment information.

[0097] In some embodiments, the at least one product term includes a comparative subject and a comparative object, or may not include a comparison body, such as a comparative object, such as when comment information is similar to "external exterior than Y Di Slash, "This clearly includes comparable principal and comparative objective comment information, can directly determine the" ratio "as the comparative subject," T Sla "is a more object; when comment information is the product A sold offline Or discuss the comment information in the forum, and the information similar to "The shape of this car is not as good as T S-LL", which is implied in this kind of comment information, and can default product A as a comparison entity, "T Sla "For comparative object.

[0098]Since the review information usually appears, therefore, the terminal can obtain at least one product term from review information on the target product as well as other competitive products that compete with the target product, and in order to accurately determine product comparison relations, but also from the review information obtaining at least one candidate word corresponding to the physical product. In addition, the user may occur when the orientation discrimination comments on the product, which typically express the orientation discrimination description through the emotions, Therefore, the terminal can also obtain information for the Qing Gan expressing at least one product review information from word.

[0099] As an example, the terminal obtaining at least one product term review information, the candidate entity to at least one product term corresponding to, and emotional for at least one product term expression information comprises: reviews information word processing, to obtain a plurality of word ; entity class identified by the identification means from the plurality of sub-word product word at least one entity obtaining at least one product term, emotional expression product information for at least one term and identifying from the plurality of sub-word lexicon through the emotions; from the knowledge base entity obtaining at least one candidate word for each product term corresponding to the product, the knowledge base includes entity information entities of the same type with at least one product term describing.

[0100] In some embodiments, the terminal can review the information word to word processing tools open in Chinese, the Chinese word open source tool can include stutter word. The tool can identify the entity class comprises tools Standford-NER.

[0101] For example, when the terminal obtained by the above step 201 to review comments acquired, the terminal word Comment tool by Chinese word, to give "Di ratio Y", "T Slavin", "exterior", "ratio" "pretty" and the like, and the entity class identified by the identification means, can be obtained product word "Di ratio Y" and "T Ezra."

[0102] Since the user Commenting on the product, sometimes referred to will be used, and different entities may correspond to the same short, therefore, in order to accurately determine product comparison relationship, the terminal needs to determine the exact product at least one product in each product term corresponding word, i.e., the terminal needs to obtain at least one product entity candidate word corresponding to the word for each product.

[0103] As an example, the terminal acquires at least one operation from the knowledge base entity candidate words for each product term corresponding to the product comprises at least the following operations: acquiring description information for each product term and described with reference to each reference knowledge base entity information; description information according to the information described with reference to each product and each reference word entities, determining for each text vector product between the word and the similarity of each reference entity; text vector similarity is larger than the reference entity similarity threshold determining candidate entities corresponding to each keyword.

[0104] Incidentally, the knowledge base includes information about the number of entities, the entity located in the same field and the number of entities in the review information appears, and the knowledge base is constructed in advance to obtain the terminal.

[0105] As an example, when building the knowledge base terminal, the terminal can advance in order to crawl from the relevant industry portal product display interface for all products, product attributes and crawling configuration information in the introductory page of each product, and product attributes the configuration information includes two types of information: text and numeric information descriptive information. Thereafter, the terminal can attribute information text descriptive information and products related products from crawling the site similar to Wikipedia supplement described.

[0106] Note that described with reference to the reference information for each entity type information comprises a textual description of each of the reference description.

[0107] In some embodiments, the terminal can be determined with reference to the description of each of the reference description information vector representation of the entity, and each product term in context representation vector, vector description information indicates, for each product and for each word by referring to the reference entity It indicates the description information of the context vector, is determined for each text vector similarity between each reference word and product entity.

[0108] As an example, the terminal can be described with reference to the description information according to information for each product and each reference word entity, determining a similarity between each text vector product of each reference word and the first entity by the following formula.

[0109]

[0110] Incidentally, in the first equation (1), p j Knowledge of the j-th reference entity, C (e i ) Is at least one product in the product word word e i Context text represents the vector, C (p j ) Of the reference entity p j The text representation of information described with reference to a vector, w t P reviews the information to carve the word t is quantified, TFIDF (t, P) represents the word t in P review information in TFIDF weight.

[0111] As an example, the terminal can also be described with reference to the description information according to information for each product and each reference word entity, determining a similarity between each text vector product words each reference entity in other ways. For example, the terminal can determine the same number of words between text description of each product term of each reference entity, according to the same number of words of text, word information of words described with reference to the description of each product and each reference word entity , between the text vector is determined for each product and each reference entity word similarity.

[0112] Incidentally, the similarity threshold can be set in advance according to the requirements, for example, the similarity threshold can be 0.5, 0.6 and so on.

[0113] In order to be able to determine the relationship between the user determines the tendency of the product, the terminal can also be expressed emotion information for at least one product from the plurality of word recognition by word thesaurus emotion.

[0114] In some embodiments, the terminal by emotion recognition vocabulary for emotional expression product of at least one word from the plurality of operation information comprising a sub-word: The meaning of each word belongs classified emotion vocabulary word identifying a plurality of points, meaning the classification including emotional words, the degree of adverbs and negative words; word number and the number of the second word statistics plurality of word belonging to the negative first word belonging emotional words; according to the number of the first word, a negative determination words wherein the at least one product term corresponding to ; determining the degree of adverbs feature word belonging degree adverbs plurality of word corresponding, belonging emotional word word corresponding to the emotional words intensity value; the negative word feature, the degree adverb features, emotion word intensity value and the second partial number of words, determined at least one of the emotional value of the product of the word.

[0115] In some embodiments, the emotional lexicon comprising information related to a large number of words and emotion, the emotion and the terminal Thesaurus Construction obtained beforehand. For example, the terminal can be Hownet emotion dictionary and thesaurus NTSUSD combined emotion, and emotion in the merge process deduplication processing, and also capable of filtering uncommon words, tends to blur the words like emotion. In addition, because during emotional expression, usually there will be negative words, and negative words will change the polarity of the emotional part of speech, therefore, the emotional lexicon also able to collect some common negative words. Also, because the adverbs of degree will have a huge impact on mild emotional words, therefore, the emotional lexicon also be able to include a number of adverbs of degree.

[0116] As the Chinese expression affirmed the double negative, and therefore, the terminal needs more statistical number of the first word in the word belongs to negative words, negative words and determine at least one characteristic of the product according to the first term corresponds to the number of sub-word.

[0117] In some embodiments, the terminal determining at least one operating negative word characteristic of the product words corresponding comprising a first partial number of words: determining a remainder after the number of a first word divided by two; and when the remainder is 1, determining the at least one product term negative word corresponding characteristic polarity inversion emotional words; when the remainder is 0, the polarity of words determined sentiment unchanged.

[0118] In some embodiments, the terminal determines the degree adverb wherein the plurality of sub-word word belongs degree adverbs corresponding to, and an operation sentiment words intensity value belonging to the emotional word word corresponding to at least: from between degree adv degree adverb intensity value corresponding relationship, determine the degree adverb corresponding degree adverbs strength values, the degree adverb intensity values determined is determined as the degree adverbs characterized adverb belonging to the extent the plurality of word in the word corresponding to; from between the emotional words and emotional words intensity value correspondence between the acquisition value corresponding to the intensity of emotion words.

[0119] It should be noted that the correspondence between the degree of adverbs and adverbs of degree intensity values, as well as correspondence between the gun and emotional times the word strength values were able to be set according to demand advance. For example, when the degree of adverbs as "slight", "no" or "relative", the degree adverb intensity corresponding to a value able to 0.25, when the degree adverb is "somewhat" or "somewhat", degree adverb intensity corresponding to a value able to 0.5, when the degree adverb is "more", "no" or "even worse", corresponding to the degree of intensity values can be adverbs as 0.75, when the degree adverb is "absolute", "very" or "very", corresponding to the degree of Adverb intensity value can be 1.

[0120] In some embodiments, a negative terminal according to the feature words, the degree of characteristic adverb, emotional intensity value and a second sub-word number of words, determining at least one value of the product sentiment second word by the following equation.

[0121]

[0122] Incidentally, in the second equation (2), sentiValue i At least one product word feeling value, m is the number of the second sub-word, f 1 (Un% 2) as a negative characteristic word, f 3 (Adv) is characteristic of the degree adverb, f 2 (w ij ) Is the emotional intensity values, intensity values emotion comparison object that is subject to comparison.

[0123] In some embodiments, the at least one terminal not only emotional expression product words corresponding to information can be determined by the above-described manner, it can also be determined in other ways, since the information can include a comment or implied comparison comprises comparing the object body and, therefore, the terminal can be meaning sentiment classification by a thesaurus word identifying each of the plurality of word belongs to, the classification includes meaning emotional words and negative words; third word the number of negative words belonging statistical plurality of word; according to the number of the third word to determine at least wherein a negative word corresponding product term; meaning word determination polar sentiment sentiment words in the plurality of word belong based, i.e. determining positive sentiment words belonging to the class emotion words, the negative word-based emotions, or a neutral sentiment words, word based on the emotion polarity and negative word belongs characterized in determining at least one product term corresponding to the emotion value is greater than 0, equal to 0, or a value less than 0, i.e., determines the emotional value comparison between a comparison subject and the object is greater than 0, equal to 0, or a value less than zero.

[0124] Step 203: the terminal information in accordance with the expression product of at least one candidate word corresponding to the entity, and emotional word for at least one product, the strength is determined comparison value candidate entity corresponding to the at least one product term.

[0125] In order to improve the accuracy of determining the intensity values of the comparison of the at least one product entity candidate word corresponding to the candidate terminal is also capable entity according to at least one product corresponding to the word, and emotional word for at least one expression product information, determining at least one word corresponding to the product comparing intensity values of the candidate entities.

[0126] As an example, the terminal entity expression information according to the candidate word corresponding to the at least one product, and emotional word for at least one product, determining at least one intensity value comparison operation candidate words corresponding to the products of the entity comprises at least: at least one product term emotional expression corresponding candidate entities and input information corresponding to the classifier to carry out classification by the classification decision, compare the intensity value to give at least one product corresponding candidate word corresponding to the entity.

[0127]As can be seen from the above, the emotional expression information can include an emotional value corresponding to at least one product word, an emotional term, and the like, and therefore, the terminal can input at least one product word corresponding to the candidate entity and the corresponding emotional expression information into the classifier. The classification decision is performed by a classifier to obtain a comparison strength value corresponding to the candidate entity corresponding to at least one product term.

[0128] It should be noted that the classifier is used to predict the emotional tendency in the comment information, and can perform results prediction and emotional analysis according to the support vector machine algorithm, and the resulting equation of the calculation result of the support vector can be WX + B = Y, where W is slope, X is the abscissa, y is the ordinate, B is a constant. Further, when the classifier performs emotional analysis, the above equation can be normalized, any of the number of linear can be satisfied with the above equation (N ij Y ij ) Satisfying y ij (WN ij + b).

[0129] It will also be noted that the standard for supporting vector machine SVM is to find the maximum classification interval, and the classification interval can be 2 / || W || in the present application embodiment, and obtain the best classification surface by calculating its maximum classification interval, and finally Transforming into a dual problem by Lagranga optimization method, selecting the internal function function as a nuclear function, and converts nonlinear changes to linear issues. A function of comparing the comparison intensity value corresponding to the candidate entity corresponding to the at least one product word (also a classification decision function) can be the third formula.

[0130]

[0131] It should be noted that in the third formula (3), score i A comparison intensity value corresponding to the candidate entity corresponding to at least one product word. For Laglan Multiplier, B * Classified threshold, k (n ij , N i ) Is the internal function function is in it.

[0132] In some embodiments, since the emotional expression information can include an emotional intensity value corresponding to at least one product word, the terminal can determine the emotional intensity value corresponding to at least one product word included in the emotional expression information to correspond to at least one product word. The comparison intensity value of the candidate entity can be a comparison intensity value between the comparative body and the comparative object.

[0133] Since comparing the comparative body is compared to the comparative object, it may be compared from a plurality of aspects, and therefore, the terminal can determine at least one comparison intensity value between the comparison body and the comparative object in accordance with the above manner.

[0134] Step 204: The terminal determines the product comparison of the candidate entity corresponding to the candidate entity corresponding to the candidate entity according to the candidate entity corresponding to at least one product word.

[0135] As an example, the terminal determines a comparison relationship of the product compare relationship of the candidate entity corresponding to the candidate entity corresponding to the candidate entity according to the at least one product term, at least the comparison of the comparison of candidate entities corresponding to at least one product word Value, draw a comparisonal relationship diagram corresponding to the candidate entity corresponding to at least one product word, which is used to describe the comparison relationship existing in the candidate entity corresponding to at least one product word; for the target entity and reference entity in the comparison relational diagram The comparison relationship is performed to measure the weight metric, resulting in the comparative relationship weight between the target entity and the reference entity.

[0136] As an example, the terminal draws a comparison relationship of the candidate entity corresponding to the candidate entity corresponding to the candidate entity corresponding to the at least one product word, at least the comparison of the comparison intensity of the target entity and the reference entity When the value is equal to 0, the target entity is connected to the reference entity, the target entity is a comparison entity corresponding to at least one product word, and the reference entity is used for the candidate entity corresponding to the at least one product term; When the comparison intensity value between the target entity and the reference entity is not equal to 0, the target entity and the reference entity are connected to the reference entity.

[0137] It should be noted that in the present application embodiment, the target entity is a comparative body, and the reference entity is a comparative object.

[0138] In some embodiments, when the comparison intensity value is 0, it will be regarded between the target entity and the reference entity, and therefore, the target entity and the reference entity can be connected by the non-harmed segment; when the comparison intensity value is not 0 It is to indicate that the target entity is a competitive relationship between the target entity and the reference entity, and therefore, the target entity can be connected to the reference entity by a line segment.

[0139] As an example, when the comparison intensity value is less than 0, the direction line segment is directed by the reference entity to the target entity; when the comparison intensity value is greater than 0, the directional line segment is directed by the target entity to the reference entity.

[0140] It should be noted that, by the above, since the comparative body is compared to the comparative object, it may be compared from a plurality of aspects, and the terminal can determine at least one comparison intensity value between the comparison body and the comparative object in accordance with the above manner. . Therefore, when the target entity is compared with the reference entity, there may be a plurality of comparison intensity values between the target entity and the reference entity, and the terminal is in the comparison line segment of the target entity and the reference entity, the target entity and the reference entity There may be multiple comparison line segments. The comparison line segment includes at least a wireless segment and a line segment.

[0141] In some embodiments, since the target entity is compared to the reference entity, it may include multiple aspects of comparison, and may be better in the comparison process, and some may be better reference entity, so In order to unify the line segment, it is also possible to perform weight metric processing on the comparative relationship between the target entity and the reference entity in the diagram, resulting in the comparative relationship weight between the target entity and the reference entity.

[0142] As an example, the terminal can determine the comparison relationship weight between the target entity and the reference entity, and the comparison relational weight is labeled on a comparison line segment between the target entity and the reference entity.

[0143] As an example, the terminal can determine a comparison relationship weight between the target entity and the reference entity by the following fourth formula.

[0144]

[0145] It should be noted that in the fourth formula (4), the number of emotional words compared between the target entity and the reference entity, N is a comparison intensity value between the target entity and the reference entity. number.

[0146] In the present application embodiment, the terminal can obtain the comment information, and determine the entity corresponding to the product word in the comment information by constructed the product knowledge base, and combine the emotional word library, and add no words, degree adverbs, etc. to compare the calculation of intensity value. After that, the comparative strength value can be determined to determine the comparison relationship, so that there is no need to browse all the commented information, it is possible to effectively identify the comparison relationship between the products in the comment information, thereby improving the accuracy and efficiency of determining the comparison relationship of the product.

[0147] image 3 It is a structural schematic structural diagram of a determination apparatus for comparing a comparative relationship provided by the embodiment of the present application. The determination device for the compare relationship of the product may include: acquisition module 301, the first determination module 302, and the second determination module 303.

[0148] The acquisition module 301 is used to obtain at least one product word in the comment information, the at least one product word candidate entity, and emotional expression information for the at least one product word, the at least one product word is used to describe the same Type Products, the candidate entities corresponding to the at least one product word exist in attributes;

[0149] The first determination module 302 is configured to determine the comparison strength value of the candidate entity corresponding to the at least one product word according to a candidate entity corresponding to the at least one product word, and an emotional expression information for the at least one product.

[0150] The second determination module 303 determines a comparison of the product comparison of candidate entities corresponding to the at least one product term, according to the comparison intensity value of the candidate entity corresponding to the at least one product word.

[0151] In some embodiments, see Figure 4 The acquisition module 301 includes:

[0152] The chronological submodule 3011 is used to distinguish the comment information to obtain a plurality of words;

[0153] The identification submodule 3012 is configured to identify the product word of at least one entity from the plurality of words by the physical class identification tool, and obtain the at least one product word, and identify the plurality of words from the plurality of words through the emotional word library Emotional expression information of at least one product word;

[0154] The submodule 3013 is used to obtain a candidate entity corresponding to each product word in the at least one product word from the knowledge base, including entity information of the same type of entity as the at least one product word.

[0155] In some embodiments, the identification sub-module 3012 is for:

[0156] Through the emotional word library identify the word geographical classification belongs to each of the plurality of words, the word wording classification includes emotional words, degree adverbs, and noble words;

[0157] Statistically the number of first words that belong to the noble word in the plurality of words and the number of second wording belonging to the emotion;

[0158] According to the number of first particle numbers, the negative word characteristics corresponding to the at least one product term are determined;

[0159] It is determined that the plurality of words belonging to the level of the degree of the adaptation corresponds to the extent of the emotional word intensity value corresponding to the expression of the emotion.

[0160] According to the negative words, the degree adverb feature, the emotional word intensity value, and the second word number, the emotional value of the at least one product word is determined.

[0161] In some embodiments, the first determining module 302 is for:

[0162] The candidate entity corresponding to the at least one product word and the corresponding emotional expression information are input to the classifier to classify the classification decision through the classifier to obtain a comparison intensity value corresponding to the candidate entity corresponding to the at least one product.

[0163] In some embodiments, see Figure 5 The second determination module 303 includes:

[0164] The submodule 3031 is drawn, and a comparison diagram of a candidate entity corresponding to the candidate entity corresponding to the candidate entity corresponding to the candidate entity corresponding to the at least one product word, the compare relationship diagram for describing the Comparative relationships in at least one product word corresponding to candidate entities;

[0165] Processing submodules 3032 for weight metric processing between the target entities and the reference entity in the compare diagram, to obtain a comparison relationship weight between the target entity and the reference entity.

[0166] In some embodiments, the mapping sub-module 3031 is for:

[0167]When the comparison intensity value between the target entity and the reference entity is equal to 0, the target entity is connected to the reference entity by a non-harmed segment, and the target entity is a comparable entity corresponding to the at least one product word. The reference entity is used for comparing entities for the candidate entities corresponding to the at least one product term;

[0168] When the comparison intensity value between the target entity and the reference entity is not equal to 0, the target entity is connected to the reference entity by a direction line segment.

[0169] In the embodiment of the present application, the terminal can acquire information review, and review information determination entity corresponding product word knowledge by building products, while incorporating the emotional vocabulary, and added negative words, calculates the degree adverb comparing intensity values and then can be determined by comparing the intensity of the relationship between product comparison values, eliminating the need for staff to see all comments information, it can effectively identify the comparison between product review information, thereby improving the accuracy and efficiency of the comparison to determine the relationship of the product.

[0170] Incidentally: determining means for comparing the relationship between the above-described embodiments provide the product in determining the relationship between product comparison, division of the foregoing functional modules is illustrated, in practice, it may assign different from the above-described functions according to the needs the function is performed, the internal structure of the apparatus is divided into different functional modules to complete all or part of the functions described above. Further, the embodiment method determines determining means for comparing the relationship between the above-described embodiments provide products of Comparison Examples relationship with the same conception, the specific embodiments of the method implementation process will not be repeated here.

[0171] Image 6 This application shows a block diagram of a terminal 600 provided in one exemplary embodiment. The terminal 600 may be: smart phones, tablet computers, MP3 player (Moving Picture Experts Group Audio Layer III, motion picture experts compression standard audio level 3), MP4 (Moving Picture Experts Group Audio LayerIV, motion picture experts compression standard audio level 4) players, laptop or desktop computer. Terminal 600 may also be referred to as another name of the user device, a portable terminal, the terminal laptop, desktop terminals.

[0172] Typically, the terminal 600 comprises: a processor 601 and memory 602.

[0173] The processor 601 may include one or more processing cores, such as a processor core 4, 8 core processors. The processor 601 may employ DSP (Digital Signal Processing, digital signal processing), FPGA (Field-Programmable Gate Array, field programmable gate arrays), PLA (Programmable Logic Array, a programmable logic array) at least one hardware form accomplish. The processor 601 may also include a main processor and a coprocessor, the main processor is a data processor for processing the awake state, also called CPU (Central ProcessingUnit, the central processor); coprocessor is to the data in the standby state, low power processors for processing. In some embodiments, processor 601 may be integrated with a GPU (Graphics Processing Unit, an image processor), the GPU is responsible for rendering and display contents displayed drawing required. In some embodiments, processor 601 may further include AI (Artificial Intelligence, AI) processor, a processor for the AI processing for calculating operations related to machine learning.

[0174] The memory 602 may comprise one or more computer-readable storage medium, the computer-readable storage medium may be non-transitory. The memory 602 may also include high-speed random access memory, and a nonvolatile memory, such as one or more magnetic disk storage devices, flash storage device. In some embodiments, the non-transitory computer-readable memory 602 storage medium for storing at least one instruction, at least one instruction to be executed by a processor 601 to implement the present application provides a method of embodiment Compare method of determining relationships.

[0175] In some embodiments, terminal 600 optionally further comprises: a peripheral interface 603, and at least one peripheral device. Processor 601, memory 602, and peripheral interface 603 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 603 via a bus, a signal line or a circuit board. In particular, the peripheral device comprising: a radio frequency circuit 604, a display screen 605, camera assembly 606, an audio circuit 607, and at least one positioning assembly 608 of power supply 609.

[0176] Peripherals interface 603 can be used to connect I / O (Input / Output, Input / Output) at least one peripheral device associated with the processor 601 and memory 602. In some embodiments, processor 601, memory 602, and peripheral device interface 603 is integrated in the same chip or circuit board; In some other embodiments, the processor 601, any peripheral interface 603 and a memory 602 or two separate chips or may be implemented on a circuit board, the present embodiment does not define them.

[0177] RF circuit 604 for receiving and transmitting RF (Radio Frequency, RF) signals, also called electromagnetic signals. A radio frequency circuit 604 communicates with a communication network by means of electromagnetic signals, and other communication devices. RF circuit 604 converts the electric signal to be transmitted electromagnetic signal, or converts the received electromagnetic signal into an electrical signal. Alternatively, the radio frequency circuit 604 comprises: an antenna system, the RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a CODEC chipset, a subscriber identity module card or the like. RF circuitry 604 may communicate with other terminals via at least one wireless communication protocol. The wireless communication protocols include, but are not limited to: MAN, each generation mobile communication network (2G, 3G, 4G and. 5G), wireless LAN and / or WiFi (Wireless Fidelity, WiFi) network. In some embodiments, RF circuit 604 may further include a NFC (Near Field Communication, short-range wireless communication) associated circuitry, which is not to be defined in the present application.

[0178] Display 605 for displaying a UI (User Interface, UI). The UI can include graphics, text, icons, video, and any combination of other combinations. When the display 605 is a touch screen, the touch screen 605 also has the ability to signal collection surface or over a display screen 605. The touch signal may be input as a control signal to the processor 601 for processing. In this case, the display 605 can also be used to provide virtual buttons and / or virtual keyboard, also known as soft buttons and / or keyboard. In some embodiments, the display 605 may be a set terminal of the front panel 600; in other embodiments, the display 605 may be at least two, are disposed on different surfaces of the terminal 600 or a folded design; in in other embodiments, the display 605 may be a flexible display screen, is provided on the curved surface of the terminal 600 or folded surface. Even, the display 605 may also be provided a non-rectangular irregular pattern, i.e. shaped screen. Display 605 may use LCD (LiquidCrystal Display, LCD), was prepared (Organic Light-Emitting Diode, OLED) and other materials OLED.

[0179] Camera assembly 606 for capturing images or video. Alternatively, the camera assembly 606 includes a front camera and a rear camera. Typically, the front camera is disposed at the front panel of the terminal, and the rear camera is disposed on the back of the terminal. In some embodiments, the rear camera is at least two, which are the main camera, a depth of field camera, a wide-angle camera, a telephoto camera, to achieve a main camera and a depth of field camera fusion implementation background blurring function, main camera And wide-angle cameras integrate panoramic shooting and VR (Virtual Reality, virtual reality) shooting function or other fusion shooting function. In some embodiments, the camera assembly 606 can also include a flash. The flash can be a monochromatic thermal sparkle, or a two-color thermostat. The two-color thermostat refers to a combination of warm light flash and cold light flashlights, which can be used for light compensation of different color temperature.

[0180] The audio circuit 607 may include a microphone and a speaker. A microphone for collecting sound waves and the user environment, and sound waves into an electrical signal input to the processor 601 processes an input to the radio frequency circuit 604 to enable voice communication. For the purpose of noise reduction stereo or acquired, it may be a plurality of microphones, disposed at different parts of each terminal 600. The microphone can also be an array microphone or a full-to-pending microphone. A speaker for converting an electrical signal from the processor 601 or the RF circuit 604 into sound waves. The speaker can be a conventional thin film speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, the electrical signal can be converted to the sound waves that the human audible sound can also be converted into the sound wave of human being invisible to ranging. In some embodiments, audio circuitry 607 may further include a headphone jack.

[0181] Positioning assembly 608 for positioning the terminal 600 current location to a navigation or LBS (LocationBased Service, location-based services). Location component 608 may be based on the US GPS (GlobalPositioning System, Global Positioning System), China's Beidou positioning assembly systems, Russia Gray NASDAQ system or the EU's Galileo system.

[0182] Power supply 609 for supplying power to the various components of the terminal 600. Power source 609 may be AC, DC, disposable batteries or rechargeable batteries. When the power supply 609 includes a rechargeable battery, the rechargeable battery can be charged may support wired or wireless charging. This rechargeable battery can also be used to support fast charge technology.

[0183] In some embodiments, terminal 600 further includes one or more sensors 610. The one or more sensors 610 include, but are not limited to: an acceleration sensor 611, a gyro sensor 612, a pressure sensor 613, fingerprint sensor 614, optical sensor 615 and proximity sensor 616.

[0184] The acceleration sensor 611 may detect the magnitude of acceleration in three axes terminal 600 to establish a coordinate system. For example, the acceleration sensor 611 may be used to detect the component of the gravitational acceleration in three axes. The processor 601 may gravitational acceleration signal of the acceleration sensor 611 acquisition, control display 605 to display a user interface view in landscape or portrait view. The acceleration sensor 611 may also be used to collect motion data of the game or the user.

[0185] The gyro sensor 612 may detect the direction of the terminal body 600 and the rotational angle, the gyro sensor 612 may cooperate to collect user operation terminal 600 3D acceleration sensor 611. The data processor 601 collected by the gyro sensor 612, can achieve the following functions: motion sensors (such as the UI are changed according to the user's tilt operation), when the image stabilization shooting game control and inertial navigation.

[0186]The pressure sensor 613 can be disposed on the side frame of the terminal 600 and / or the lower layer of the display screen 605. When the pressure sensor 613 is disposed on the side of the terminal 600, the user's handover signal to the terminal 600 can be detected, and the processor 601 performs left and right hand identification or shortcut to the grip signal collected by the pressure sensor 613. When the pressure sensor 613 is disposed at the lower layer of the display screen 605, the processor 601 performs control of the operabity control on the UI interface according to the pressure of the display screen 605, according to the pressure of the display screen 605. Operable controls include at least one of button controls, scroll strips, icon controls, menu controls.

[0187] The fingerprint sensor 614 is used to collect the user's fingerprint, and the processor 601 identifies the identity of the user according to the fingerprint acquired by the fingerprint sensor 614, or, by the fingerprint sensor 614 identifies the identity of the user according to the collected fingerprint. When identifying the identity of the user is a trusted identity, the processor 601 authorizes the user to perform related sensitive operations, including unlocking screens, viewing encrypted information, downloading software, payment, and changing settings. The fingerprint sensor 614 can be provided with the front, back or side of the terminal 600. When the physical buttons or vendor logo is provided on the terminal 600, the fingerprint sensor 614 can be integrated with the physical buttons or vendor LOGO.

[0188] Optical sensor 615 is used to collect ambient light intensity. In one embodiment, the processor 601 can control the display brightness of the display screen 605 based on the ambient light intensity acquired by the optical sensor 615. Specifically, when the ambient light intensity is high, the display brightness of the display screen 605 is increased; when the ambient light intensity is low, the display brightness of the display screen 605 is lowered. In another embodiment, the processor 601 can also dynamically adjust the shooting parameters of the camera assembly 606 according to the ambient light intensity acquired by the optical sensor 615.

[0189] The proximity sensor 616 is also known as the sensor, typically disposed at the front panel of the terminal 600. The proximity sensor 616 is used to acquire the distance between the user and the front side of the terminal 600. In one embodiment, when the proximity sensor 616 detects that the distance between the user and the front of the terminal 600 is gradually changed, the processor 601 controls the display screen 605 from the bright screen state; when the receiving sensor 616 is detected When the distance between the user and the front of the terminal 600 gradually increases, the processor 601 controls the display screen 605 switching to the bright screen state.

[0190] Those skilled in the art will appreciate that Image 6 The structure shown is not constituied to the limitations of the terminal 600, which may include more or fewer components, or in combination, or different component arrangements.

[0191] The present application embodiment also provides a non-tutaneous computer readable storage medium, and when the instruction in the storage medium is performed by the processor's processor, the terminal can perform a determination method of the product comparison relationship provided by the above-mentioned embodiment.

[0192] The present application embodiment also provides a computer program product comprising instructions, and when it is running on the terminal, the terminal performs a determination method of comparing the product comparison relationship provided by the above embodiment.

[0193] One of ordinary skill in the art will appreciate that all or some of the steps that implement the above embodiment can be accomplished by hardware, and can also be done by a program to instruct the relevant hardware, the program can be stored in a computer readable storage medium, the above The storage medium mentioned can be read-only memory, disk, or disc.

[0194] The preferred embodiment of the present application embodiment is not intended to limit the invention, and any modification, equivalent replacement, improvement, etc. according to the spirit and principles of the present application embodiments. It is included within the scope of protection of this application.

PUM

Description & Claims & Application Information

We can also present the details of the Description, Claims and Application information to help users get a comprehensive understanding of the technical details of the patent, such as background art, summary of invention, brief description of drawings, description of embodiments, and other original content. On the other hand, users can also determine the specific scope of protection of the technology through the list of claims; as well as understand the changes in the life cycle of the technology with the presentation of the patent timeline. Login to view more.

Classification and recommendation of technical efficacy words

Improve efficiency and accuracy

Product comparison relation determination method and device and computer storage medium

A technology for comparing relationships and products, applied in computing, relational databases, structured data retrieval, etc., can solve problems such as low efficiency and accuracy, and achieve the effect of improving accuracy and efficiency

AI-Extracted Technical Summary

Problems solved by technology

Method used

Abstract

Image

Examples

Example Embodiment

PUM

Description & Claims & Application Information

Similar technology patents

Typical operation identification and multi-dimensional analysis method of monitoring equipment

Security protocol vulnerability mining method based on multiple reverse cases

Material demanding method for auxiliary material object disc of ultra-high-speed cigarette making machine

Failure prediction and diagnosis control method applicable to shield tunneling machine

Chinese syntax parsing method with merged semantic information

Classification and recommendation of technical efficacy words

Advertisement playing system and method based on image identification

Realization method of atrial fibrillation detection based on deep convolutional neural network

Android malicious software detecting platform oriented to mobile internet

Bluetooth connecting method and mobile terminal

Multi-robot combined target searching method of animal-simulated space cognition