CN117195897A - NLP-based intelligent material description splitting method and system - Google Patents
NLP-based intelligent material description splitting method and system Download PDFInfo
- Publication number
- CN117195897A CN117195897A CN202311160916.9A CN202311160916A CN117195897A CN 117195897 A CN117195897 A CN 117195897A CN 202311160916 A CN202311160916 A CN 202311160916A CN 117195897 A CN117195897 A CN 117195897A
- Authority
- CN
- China
- Prior art keywords
- description
- information
- splitting
- determining
- closest
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 239000000463 material Substances 0.000 title claims abstract description 250
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000002372 labelling Methods 0.000 claims abstract description 57
- 238000012549 training Methods 0.000 claims abstract description 54
- 238000000605 extraction Methods 0.000 claims abstract description 22
- 238000004458 analytical method Methods 0.000 claims description 18
- 238000012986 modification Methods 0.000 claims description 11
- 230000004048 modification Effects 0.000 claims description 11
- 238000010586 diagram Methods 0.000 claims description 8
- 238000012790 confirmation Methods 0.000 claims description 5
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 238000003058 natural language processing Methods 0.000 description 26
- 230000009286 beneficial effect Effects 0.000 description 11
- PXHVJJICTQNCMI-UHFFFAOYSA-N Nickel Chemical compound [Ni] PXHVJJICTQNCMI-UHFFFAOYSA-N 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 239000010963 304 stainless steel Substances 0.000 description 2
- 229910000589 SAE 304 stainless steel Inorganic materials 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 229910052759 nickel Inorganic materials 0.000 description 2
- 238000007747 plating Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- -1 for example Substances 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000013068 supply chain management Methods 0.000 description 1
- 238000004381 surface treatment Methods 0.000 description 1
Landscapes
- Machine Translation (AREA)
Abstract
The invention provides an NLP-based intelligent material description splitting method and system, wherein the method comprises the following steps: acquiring text data of material description; labeling the text data by named entities to obtain labeling data, and forming a training data set by the labeling data; constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set; inputting text data of the material description to be split into a material description splitting model to obtain a splitting result of the material to be split. According to the invention, through intelligent splitting of text data of material description, the accuracy of extraction and application of material information is improved, and a precondition is provided for classification and retrieval of materials in the later period. Meanwhile, the split material description is communicated with other data sources, so that the material utilization efficiency and accuracy are improved.
Description
Technical Field
The invention relates to the technical field of computer data processing, in particular to an NLP-based intelligent material description splitting method and system.
Background
Material management, such as material management for raw materials, parts, and finished products, requires material descriptions to be obtained. The material description contains the characteristics and features of the material, and the material description needs to be split, so that the subsequent supply chain management can accurately understand and process the material. In reality, the material is often split manually, so that a large amount of labor cost is occupied, the splitting efficiency is low, and the error rate is high. Therefore, an intelligent material description splitting method is needed, and the splitting efficiency and accuracy of materials are improved.
In the prior art, CN 111597777A provides a material data processing method, a device and electronic equipment, and through splitting a product order, data integration is performed on material resources among heterogeneous systems, so that the problem that the material resources among the heterogeneous systems are difficult to match is solved, and interconnection and intercommunication among systems on a supply chain are realized. However, the material description is not split through artificial intelligence and machine deep learning, and attribute extraction is completed, so that the requirement of intelligent splitting of the material description cannot be met.
Disclosure of Invention
The material description intelligent splitting method based on NLP provided by the embodiment of the invention comprises the following steps:
acquiring text data of material description;
labeling the text data by named entities to obtain labeling data, and forming a training data set by the labeling data;
constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set;
inputting text data of the material description to be split into a material description splitting model to obtain a splitting result of the material to be split.
Preferably, labeling the text data by named entity to obtain labeling data and forming a training data set, including:
carrying out semantic analysis on the text data and confirming first description information;
performing entity identification on the first description information to determine second description information;
performing attribute analysis on the first description information to determine third description information;
carrying out named entity labeling on the text data by the second descriptive information and the third descriptive information, and determining labeling data;
based on the annotation data, a training dataset is constructed.
Preferably, the semantic analysis is performed on the text data, and the confirming of the first description information includes:
extracting words from the text data to determine first word information;
semantic extraction is carried out on the text data, and a first logic relationship is determined;
carrying out grammar extraction on the text data to determine grammar rule relation;
determining a word dendrogram based on the first word information and the first logical relationship;
based on the word dendrogram, first description information is determined.
Preferably, the entity identification is performed on the first description information, and the second description information is determined
Acquiring grammar rule relations and determining part-of-speech relations;
based on the part-of-speech relationship, noun extraction is carried out on the first description information, and second word information is obtained;
extracting the entity of the second word information to determine third word information;
based on the third word information, second description information is determined.
Preferably, performing attribute analysis on the first description information to determine third description information includes:
determining attribute rules;
acquiring context information of the text data, extracting attributes of the first description information based on the context information of the text data, and determining attribute words in the first description information;
determining the repetition times of the attribute words based on the attribute words in the first description information;
sequencing the repetition times of the attribute words, and determining the attribute word with the largest repetition time;
constructing a first description information vocabulary library based on the historical data;
searching a vocabulary with the maximum similarity with the attribute words with the maximum repetition times based on the first description information vocabulary library, and determining fourth word information;
third description information is determined based on the fourth word information.
Preferably, naming entity labeling is performed on the text data by the second description information and the third description information, and labeling data is determined, including:
determining to label the first label and label the second label based on the second description information;
determining a third label and a fourth label based on the third description information;
based on the word tree diagram, fusing the first label, the second label, the third label and the fourth label to determine label information;
and labeling the named entities of the text data based on the labeling information, and determining labeling data.
Preferably, constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by using a training data set, including:
constructing a material description splitting model based on NLP;
training the material description splitting model based on the training data set;
calculating the accuracy of a material description splitting model;
and adjusting and optimizing parameters of the material description splitting model based on the accuracy.
Preferably, the intelligent splitting method for the material description based on NLP further comprises the following steps:
determining the closest material based on the splitting result of the material to be split;
wherein, based on the resolution result of the material to be resolved, determining the closest material comprises:
constructing a material database based on the historical data;
determining name keywords and attribute keywords based on a splitting result of the material to be split;
searching a plurality of search results in a material database based on the name keywords;
performing attribute analysis on the corresponding search results to obtain fifth word information of the search results;
and matching the fifth word information with the attribute keywords, and determining the search result with the largest attribute keyword matching as the closest material.
Preferably, the intelligent splitting method for the material description based on NLP further comprises the following steps:
acquiring the stock of the closest material and judging whether the stock of the closest material is lower than a set threshold value, if so, searching the price based on the closest material and generating a closest material purchase table for modification and confirmation by a user, otherwise, no operation is performed;
the method comprises the steps of acquiring inventory information of the closest materials, judging whether the inventory information of the closest materials is lower than a set threshold value, searching the price based on the closest materials and generating a closest material purchase table for user modification and confirmation if the inventory information of the closest materials is lower than the set threshold value, otherwise, not operating, and comprising the following steps:
determining a material code for the closest material based on the closest material;
determining an inventory of the closest material based on the material codes of the closest material;
judging whether the stock of the closest material is lower than a set threshold value;
if the stock of the closest material is lower than the set threshold value, carrying out online searching based on the closest material;
determining a purchasing order of a number of sellers based on the price, purchasing record, and historical purchasing record of the closest material;
determining a material purchase table based on the purchase order of a plurality of sellers;
sending the material purchase list to a user;
user modification and determination instructions are obtained, a subscription order is generated, and a confirmed seller is sent.
The invention also provides an NLP-based intelligent material description splitting method, which comprises the following steps:
the data acquisition module is used for acquiring text data of material description;
the training data set module is used for labeling the named entities of the text data to obtain labeling data and forming a training data set;
the model training module is used for constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set;
the result splitting module is used for inputting text data of the material description to be split into the material description splitting model to obtain a splitting result of the material to be split.
The invention has the beneficial effects that:
according to the invention, through intelligent splitting of text data of material description, the accuracy of extraction and application of material information is improved, and a precondition is provided for classification and retrieval of materials in the later period. Meanwhile, the split material description is communicated with other data sources, so that the material utilization efficiency and accuracy are improved.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a schematic diagram of an NLP-based material description intelligent splitting method in an embodiment of the invention;
fig. 2 is a schematic diagram of an NLP-based material description intelligent splitting system in an embodiment of the invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, it being understood that the preferred embodiments described herein are for illustration and explanation of the present invention only, and are not intended to limit the present invention.
The embodiment of the invention provides an NLP-based intelligent material description splitting method, which comprises the following steps as shown in fig. 1:
step 1: acquiring text data of material description;
step 2: labeling the text data by named entities to obtain labeling data, and forming a training data set by the labeling data;
step 3: constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set;
step 4: inputting text data of the material description to be split into a material description splitting model to obtain a splitting result of the material to be split.
The working principle and the beneficial effects of the technical scheme are as follows:
and step 1, acquiring text data of material description. The text data of the material description obtained in the embodiment includes file numbers, material names (such as nuts), material major types, specification requirements (such as specification model hex, size inside diameter 6 mm, surface treatment nickel plating) and the like, and step 2 performs named entity labeling on the text data to obtain labeling data such as nuts (hex, size inside diameter 6 mm, nickel plating) and forms a training data set from the labeling data. And 3, constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set. Step 4: inputting text data of the material description to be split into a material description splitting model to obtain a splitting result of the material to be split, such as hardware-screw cap.
According to the embodiment of the invention, through intelligent splitting of the text data of the material description, the accuracy of extraction and application of the material information is improved, and a precondition is provided for classification and retrieval of the material in the later period. Meanwhile, the split material description is communicated with other data sources, so that the material utilization efficiency and accuracy are improved.
In one embodiment, step 2 comprises:
step 2.1: carrying out semantic analysis on the text data and confirming first description information;
step 2.2: performing entity identification on the first description information to determine second description information;
step 2.3: performing attribute analysis on the first description information to determine third description information;
step 2.4: carrying out named entity labeling on the text data by the second descriptive information and the third descriptive information, and determining labeling data;
step 2.5: based on the annotation data, a training dataset is constructed.
The working principle and the beneficial effects of the technical scheme are as follows:
and 2.1, carrying out semantic analysis on the text data, and confirming the first description information. For example, the first description information includes entity, attribute, and grammar rule relationships. Step 2.2: and carrying out entity identification on the first descriptive information, and determining second descriptive information, wherein the second descriptive information comprises product names and materials. Step 2.3: and carrying out attribute analysis on the first descriptive information to determine third descriptive information. For example, the third descriptive information includes color, weight, capacity, etc. The attribute analysis in this embodiment may implement attribute extraction using pattern matching, keyword extraction, or a machine learning-based method. Step 2.4: and labeling the text data by using the named entities according to the second descriptive information and the third descriptive information, and determining labeling data. Step 2.5: based on the annotation data, a training dataset is constructed.
According to the embodiment of the invention, the named entity labeling is carried out through entity identification and attribute analysis respectively, the labeling data in the training data set is determined, and the information in the text data is extracted and standardized, so that the model is convenient to train.
In one embodiment, step 2.1 comprises:
step 2.1.1: extracting words from the text data to determine first word information;
step 2.1.2: semantic extraction is carried out on the text data, and a first logic relationship is determined;
step 2.1.3: carrying out grammar extraction on the text data to determine grammar rule relation;
step 2.1.4: determining a word dendrogram based on the first word information and the first logical relationship;
step 2.1.5: based on the word dendrogram, first description information is determined.
The working principle and the beneficial effects of the technical scheme are as follows:
and 2.1.1, extracting words from the text data to determine first word information, namely extracting the text data into words, namely the first word information. Step 2.1.2 performs semantic extraction on the text data to determine a first logical relationship, i.e. a relationship between each word, such as ladder progress, interpretation, etc. Step 2.1.3: grammar extraction is performed on the text data, and grammar rule relations, such as main names, moving guest relations and the like, are determined. Step 2.1.4: a word dendrogram is determined based on the first word information and the first logical relationship. And visually expressing the first word information through a word tree diagram. Step 2.1.5: based on the word dendrogram, first description information is determined.
According to the embodiment of the invention, the Wen Mu data is subjected to word extraction and the first word information is visually expressed through the word tree diagram, so that the first description information is finally formed, and the text data is subjected to primary processing to form
In one embodiment, step 2.2 comprises:
step 2.2.1: acquiring grammar rule relations and determining part-of-speech relations;
step 2.2.2: based on the part-of-speech relationship, noun extraction is carried out on the first description information, and second word information is obtained;
step 2.2.3: extracting the entity of the second word information to determine third word information;
step 2.2.4: based on the third word information, second description information is determined.
The working principle and the beneficial effects of the technical scheme are as follows:
step 2.2.1: and acquiring grammar rule relations and determining part-of-speech relations. For example, a primary predicate relationship may determine that a noun is preceded and a verb is followed, thereby determining the part-of-speech relationship of each term. And 2.2.2, extracting nouns from the first description information according to the part-of-speech relationship to obtain second word information, namely the second word information is mainly nouns. And 2.2.3, extracting the entity of the second word information, and determining that the third word information, namely the third word information, expresses the entity noun. Step 2.2.4 determines the second descriptive information based on the third word information, i.e. the entity noun.
According to the embodiment of the invention, the noun extraction and the entity extraction are carried out on the first description information, the second description information is determined, the entity name of the material is determined, and a precondition is provided for labeling tags of the material in the later period.
In one embodiment, step 2.3 comprises:
step 2.3.1: determining attribute rules;
step 2.3.2: acquiring context information of the text data, extracting attributes of the first description information based on the context information of the text data, and determining attribute words in the first description information;
step 2.3.3: determining the repetition times of the attribute words based on the attribute words in the first description information;
step 2.3.4: sequencing the repetition times of the attribute words, and determining the attribute word with the largest repetition time;
step 2.3.5: constructing a first description information vocabulary library based on the historical data;
step 2.3.6: searching the description information with the maximum similarity with the attribute words with the maximum repetition times based on the first description information vocabulary library, and determining fourth word information;
step 2.3.7: third description information is determined based on the fourth word information.
The working principle and the beneficial effects of the technical scheme are as follows:
step 2.3.1: the attribute rules are determined, for example, according to historical data or common knowledge, such as red, yellow, blue, green, purple, etc., and the specifications are small, medium, large, extra large, etc. Step 2.3.2: the context information of the text data is obtained, attribute extraction is performed on the first description information based on the context information of the text data, attribute words in the first description information are determined, and in the embodiment, the number words of the first description information, such as 10cm, are obtained, and the attribute "big" of the entity nut is determined according to the attribute rule. Step 2.3.3: based on the attribute word "large" in the first description information, the number of repetitions of the attribute word is determined, such as 3 occurrences in the text data. Step 2.3.4: and sequencing the repetition times of the attribute words, and determining the attribute word with the largest repetition time. For example, the "large" of the nut is repeated 10 times more than the other attribute "hexahedral", etc. Step 2.3.5: based on the historical data, a first descriptive information vocabulary library is constructed. In this embodiment, a first description information vocabulary library is constructed according to the history data, and the attribute is normalized in the step 2.3.6 of describing, based on the first description information vocabulary library, the vocabulary having the greatest similarity with the attribute word having the greatest number of repetitions, and the fourth word information is determined. For example, a "large" first description information vocabulary library is searched to obtain specific description information of "large" for the nut, and fourth word information is determined based on the most similar pair of attribute words in the first description information vocabulary library. Step 2.3.7 determines third description information based on the fourth word information.
The embodiment of the invention extracts the first descriptive information attribute and replaces the first descriptive information vocabulary base with standardized fourth part-of-speech information so as to determine the third descriptive information. Providing a precondition for the labeling of the later stage.
In one embodiment, step 2.4 comprises:
step 2.4.1: determining to label the first label and label the second label based on the second description information;
step 2.4.2: determining a third label and a fourth label based on the third description information;
step 2.4.3: based on the word tree diagram, fusing the first label, the second label, the third label and the fourth label to determine label information;
step 2.4.4: and labeling the named entities of the text data based on the labeling information, and determining labeling data.
The working principle and the beneficial effects of the technical scheme are as follows:
step 2.4.1, determining to label the first label and label the second label based on the second description information. The first tag set in this embodiment is an entity name, and the second tag is a constituent member of an entity. Step 2.4.2 determines, based on the third description information, to label the third label and label the fourth label, where the first label set in this embodiment is a standardized attribute, and the second label is a standardized specification. And 2.4.3, based on the word tree diagram, fusing the first label, the second label, the third label and the fourth label to determine label information. And 2.4.4, labeling the named entities of the text data based on the labeling information, and determining the labeling data. Namely, corresponding labeling information is labeled on the text data.
The embodiment of the invention carries out named entity labeling through text data, is convenient for carrying out specific data statistics and identification in later period, and provides a premise for machine deep learning.
In one embodiment, step 3 comprises:
step 3.1: constructing a material description splitting model based on NLP;
step 3.2: training the material description splitting model based on the training data set;
step 3.3: calculating the accuracy of a material description splitting model;
step 3.4: and adjusting and optimizing parameters of the material description splitting model based on the accuracy.
The working principle and the beneficial effects of the technical scheme are as follows:
and 3.1, constructing a material description splitting model based on NLP. NLP (Natural Language Processing) is a natural language processing technique, which is to realize the effective communication between natural language and machine. In the embodiment, a material description splitting model is constructed through NLP, and the material description splitting model is realized by adopting machine deep learning and utilizing a neural network. And 3.2, training the material description splitting model based on the training data set. And 3.3, evaluating the training result of the material description splitting model by using the accuracy. And 3.4, based on the accuracy, if the accuracy does not meet the standard, adjusting and optimizing the parameters of the material description splitting model.
In one embodiment, an intelligent splitting method for material description based on NLP further comprises:
step 5: determining the closest material based on the splitting result of the material to be split;
wherein, step 5 includes:
step 5.1: constructing a material database based on the historical data;
step 5.2: determining name keywords and attribute keywords based on a splitting result of the material to be split;
step 5.3: searching a plurality of search results in a material database based on the name keywords;
step 5.4: performing attribute analysis on the corresponding search results to obtain fifth word information of the search results;
step 5.5: and matching the fifth word information with the attribute keywords, and determining the search result with the largest attribute keyword matching as the closest material.
The working principle and the beneficial effects of the technical scheme are as follows:
and 5, determining the closest material according to the splitting result of the material to be split. The user can conveniently inquire the required materials. Step 5.1: based on the historical data, a materials database is constructed. Step 5.2: and determining name keywords and attribute keywords based on the splitting result of the materials to be split. Step 5.3: and searching out a plurality of search results in the material database based on the name keywords. Step 5.4: and carrying out attribute analysis on the corresponding search results to obtain fifth word information of the search results. Step 5.5: the fifth word information is matched with the attribute key words, and the search result with the largest matching attribute key words is determined to be the closest material, for example, 304 stainless steel hexagonal nuts are the closest to the large-size rust-proof nuts.
According to the embodiment of the invention, the name keywords are determined to search the material database, and are matched according to the attribute keywords, so that the closest material is determined, and the user can conveniently select and confirm.
In one embodiment, an intelligent splitting method for material description based on NLP further comprises:
step 6: acquiring the stock of the closest material and judging whether the stock of the closest material is lower than a set threshold value, if so, searching the price based on the closest material and generating a closest material purchase table for modification and confirmation by a user, otherwise, no operation is performed;
wherein, step 6 includes:
step 6.1: determining a material code for the closest material based on the closest material;
step 6.2: determining an inventory of the closest material based on the material codes of the closest material;
step 6.3: judging whether the stock of the closest material is lower than a set threshold value;
step 6.4: if the stock of the closest material is lower than the set threshold value, carrying out online searching based on the closest material;
step 6.5: determining a purchasing order of a number of sellers based on the price, purchasing record, and historical purchasing record of the closest material;
step 6.6: determining a material purchase table based on the purchase order of a plurality of sellers;
step 6.7: sending the material purchase list to a user;
step 6.8: user modification and determination instructions are obtained, a subscription order is generated, and a confirmed seller is sent.
The working principle and the beneficial effects of the technical scheme are as follows:
step 6.1: based on the closest material, a material code for the closest material is determined. For example, a 304 stainless steel hex nut queries for a material code 01304002. Step 6.2: based on the material codes of the closest materials, an inventory of the closest materials is determined. For example, stock of material code 01304002 is 100. Step 6.3: it is determined whether the inventory of the closest material is below a set threshold. For example, the daily material code is 01304002 consumed 50, and the threshold value is 150 in order to ensure 3 days of use. Step 6.4: if the inventory of the closest material is below the set threshold, then web searching is performed based on the closest material. Step 6.5: the purchasing order of the number of sellers is determined based on the price of the closest material, the purchase record, and the historical purchase record. And (3) comprehensively evaluating and arranging the prices, purchase records and historical purchase records of the closest materials to form the purchase sequence of a plurality of sellers. Step 6.6: a material purchase table is determined based on the purchase order of the plurality of sellers. Step 6.7: and sending the material purchase list to the user. Step 6.8: user modification and determination instructions are obtained, a subscription order is generated, and a confirmed seller is sent.
According to the embodiment of the invention, whether to purchase is determined through the closest material and the material purchase table is determined according to the online retrieval information, so that the automation of material purchase is realized, and the material purchase time is saved.
The invention also provides an NLP-based intelligent material description splitting method, which comprises the following steps:
the data acquisition module 1 is used for acquiring text data of material description;
the training data set module 2 is used for labeling the named entities of the text data to obtain labeling data and form a training data set;
the model training module 3 is used for constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set;
and the result splitting module 4 is used for inputting text data of the material description to be split into the material description splitting model to obtain a splitting result of the material to be split.
The working principle and the beneficial effects of the technical scheme are as follows:
the data acquisition module 1 acquires text data of a material description. The training data set module 2 carries out named entity labeling on the text data to obtain labeling data and form a training data set. The model training module 3 builds a material description split model based on NLP, and trains and adjusts the material description split model by utilizing a training data set. And the result splitting module 4 is used for inputting text data of the material description to be split into the material description splitting model to obtain a splitting result of the material to be split.
According to the embodiment of the invention, through intelligent splitting of the text data of the material description, the accuracy of extraction and application of the material information is improved, and a precondition is provided for classification and retrieval of the material in the later period. Meanwhile, the split material description is communicated with other data sources, so that the material utilization efficiency and accuracy are improved.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (10)
1. An intelligent material description splitting method based on NLP is characterized by comprising the following steps:
acquiring text data of material description;
labeling the text data by named entities to obtain labeling data, and forming a training data set by the labeling data;
constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set;
inputting text data of the material description to be split into a material description splitting model to obtain a splitting result of the material to be split.
2. The intelligent splitting method of material description based on NLP as set forth in claim 1, wherein the naming entity labeling of text data to obtain labeling data and constitute training data set comprises:
carrying out semantic analysis on the text data and confirming first description information;
performing entity identification on the first description information to determine second description information;
performing attribute analysis on the first description information to determine third description information;
carrying out named entity labeling on the text data by the second descriptive information and the third descriptive information, and determining labeling data;
based on the annotation data, a training dataset is constructed.
3. The intelligent splitting method of material description based on NLP as claimed in claim 2, wherein the semantic analysis is performed on the text data to confirm the first description information, comprising:
extracting words from the text data to determine first word information;
semantic extraction is carried out on the text data, and a first logic relationship is determined;
carrying out grammar extraction on the text data to determine grammar rule relation;
determining a word dendrogram based on the first word information and the first logical relationship;
based on the word dendrogram, first description information is determined.
4. The intelligent splitting method of material description based on NLP as set forth in claim 2, wherein the first description information is identified by entity to determine the second description information
Acquiring grammar rule relations and determining part-of-speech relations;
based on the part-of-speech relationship, noun extraction is carried out on the first description information, and second word information is obtained;
extracting the entity of the second word information to determine third word information;
based on the third word information, second description information is determined.
5. The intelligent splitting method of material description based on NLP as claimed in claim 2, wherein the attribute analysis is performed on the first description information to determine the third description information, comprising:
determining attribute rules;
acquiring context information of the text data, extracting attributes of the first description information based on the context information of the text data, and determining attribute words in the first description information;
determining the repetition times of the attribute words based on the attribute words in the first description information;
sequencing the repetition times of the attribute words, and determining the attribute word with the largest repetition time;
constructing a first description information vocabulary library based on the historical data;
searching a vocabulary with the maximum similarity with the attribute words with the maximum repetition times based on the first description information vocabulary library, and determining fourth word information;
third description information is determined based on the fourth word information.
6. The intelligent splitting method of material description based on NLP as set forth in claim 2, wherein naming entity labeling is performed on text data for the second description information and the third description information, and determining labeling data includes:
determining to label the first label and label the second label based on the second description information;
determining a third label and a fourth label based on the third description information;
based on the word tree diagram, fusing the first label, the second label, the third label and the fourth label to determine label information;
and labeling the named entities of the text data based on the labeling information, and determining labeling data.
7. The intelligent material description splitting method based on NLP as set forth in claim 1, wherein the building of the material description splitting model based on NLP and the training and tuning of the material description splitting model using the training data set comprises:
constructing a material description splitting model based on NLP;
training the material description splitting model based on the training data set;
calculating the accuracy of a material description splitting model;
and adjusting and optimizing parameters of the material description splitting model based on the accuracy.
8. The intelligent splitting method of material description based on NLP as claimed in claim 1, further comprising:
determining the closest material based on the splitting result of the material to be split;
wherein, based on the resolution result of the material to be resolved, determining the closest material comprises:
constructing a material database based on the historical data;
determining name keywords and attribute keywords based on a splitting result of the material to be split;
searching a plurality of search results in a material database based on the name keywords;
performing attribute analysis on the corresponding search results to obtain fifth word information of the search results;
and matching the fifth word information with the attribute keywords, and determining the search result with the largest attribute keyword matching as the closest material.
9. The intelligent splitting method of material description based on NLP as claimed in claim 8, further comprising:
acquiring the stock of the closest material and judging whether the stock of the closest material is lower than a set threshold value, if so, searching the price based on the closest material and generating a closest material purchase table for modification and confirmation by a user, otherwise, no operation is performed;
the method comprises the steps of acquiring inventory information of the closest materials, judging whether the inventory information of the closest materials is lower than a set threshold value, searching the price based on the closest materials and generating a closest material purchase table for user modification and confirmation if the inventory information of the closest materials is lower than the set threshold value, otherwise, not operating, and comprising the following steps:
determining a material code for the closest material based on the closest material;
determining an inventory of the closest material based on the material codes of the closest material;
judging whether the stock of the closest material is lower than a set threshold value;
if the stock of the closest material is lower than the set threshold value, carrying out online searching based on the closest material;
determining a purchasing order of a number of sellers based on the price, purchasing record, and historical purchasing record of the closest material;
determining a material purchase table based on the purchase order of a plurality of sellers;
sending the material purchase list to a user;
user modification and determination instructions are obtained, a subscription order is generated, and a confirmed seller is sent.
10. An intelligent material description splitting method based on NLP is characterized by comprising the following steps:
the data acquisition module is used for acquiring text data of material description;
the training data set module is used for labeling the named entities of the text data to obtain labeling data and forming a training data set;
the model training module is used for constructing a material description splitting model based on NLP, and training and optimizing the material description splitting model by utilizing a training data set;
the result splitting module is used for inputting text data of the material description to be split into the material description splitting model to obtain a splitting result of the material to be split.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311160916.9A CN117195897A (en) | 2023-09-08 | 2023-09-08 | NLP-based intelligent material description splitting method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311160916.9A CN117195897A (en) | 2023-09-08 | 2023-09-08 | NLP-based intelligent material description splitting method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117195897A true CN117195897A (en) | 2023-12-08 |
Family
ID=88982984
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311160916.9A Pending CN117195897A (en) | 2023-09-08 | 2023-09-08 | NLP-based intelligent material description splitting method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117195897A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117762917A (en) * | 2024-01-16 | 2024-03-26 | 北京三维天地科技股份有限公司 | Medical instrument data cleaning method and system based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678281A (en) * | 2013-12-31 | 2014-03-26 | 北京百度网讯科技有限公司 | Method and device for automatically labeling text |
| CN106777275A (en) * | 2016-12-29 | 2017-05-31 | 北京理工大学 | Entity attribute and property value extracting method based on many granularity semantic chunks |
| CN109614461A (en) * | 2018-11-19 | 2019-04-12 | 广州搜料信息技术有限公司 | Intelligent search method based on physical parameter semantics recognition |
| CN112395881A (en) * | 2020-11-27 | 2021-02-23 | 北京筑龙信息技术有限责任公司 | Material label construction method and device, readable storage medium and electronic equipment |
-
2023
- 2023-09-08 CN CN202311160916.9A patent/CN117195897A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103678281A (en) * | 2013-12-31 | 2014-03-26 | 北京百度网讯科技有限公司 | Method and device for automatically labeling text |
| CN106777275A (en) * | 2016-12-29 | 2017-05-31 | 北京理工大学 | Entity attribute and property value extracting method based on many granularity semantic chunks |
| CN109614461A (en) * | 2018-11-19 | 2019-04-12 | 广州搜料信息技术有限公司 | Intelligent search method based on physical parameter semantics recognition |
| CN112395881A (en) * | 2020-11-27 | 2021-02-23 | 北京筑龙信息技术有限责任公司 | Material label construction method and device, readable storage medium and electronic equipment |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117762917A (en) * | 2024-01-16 | 2024-03-26 | 北京三维天地科技股份有限公司 | Medical instrument data cleaning method and system based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10896212B2 (en) | System and methods for automating trademark and service mark searches | |
| US10565533B2 (en) | Systems and methods for similarity and context measures for trademark and service mark analysis and repository searches | |
| US20200327172A1 (en) | System and method for processing contract documents | |
| CN112632228A (en) | Text mining-based auxiliary bid evaluation method and system | |
| CN109992763A (en) | Language marks processing method, system, electronic equipment and computer-readable medium | |
| CN112507691A (en) | Interpretable financial subject matter generating method and device fusing emotion, industrial chain and case logic | |
| CN112580332B (en) | Enterprise portrait method based on label layering and deepening modeling | |
| CN117195897A (en) | NLP-based intelligent material description splitting method and system | |
| CN114036281A (en) | Citrus control question-answering module construction method based on knowledge graph and question-answering system | |
| CN113570380A (en) | Service complaint processing method, device and equipment based on semantic analysis and computer readable storage medium | |
| CN111782793A (en) | Intelligent customer service processing method, system and equipment | |
| CN110689371A (en) | Intelligent marketing cloud service platform based on AI and big data | |
| CN116187323A (en) | Knowledge graph in field of numerical control machine tool and construction method thereof | |
| CN111563168A (en) | Method for intelligently classifying customs commodities based on AI knowledge graph algorithm | |
| CN113240562A (en) | Method and system for recommending and matching obstetrical and academic research projects based on nlp | |
| CN112131246B (en) | Data center intelligent query statistical method based on natural language semantic analysis | |
| CN113592512A (en) | Online commodity identity uniqueness identification and confirmation system | |
| CN111737498A (en) | Domain knowledge base establishing method applied to discrete manufacturing production process | |
| CN115455196A (en) | Design map construction method based on RFPC (radio frequency personal computer) concept design framework | |
| CN116401338A (en) | Design feature extraction and attention mechanism based on data asset intelligent retrieval input and output requirements and method thereof | |
| Lo et al. | An emperical study on application of big data analytics to automate service desk business process | |
| Khekare et al. | Design of Automatic Key Finder for Search Engine Optimization in Internet of Everything | |
| CN116186293A (en) | Dynamic query method and model of electric power marketing data based on knowledge graph | |
| CN115408506A (en) | NL2SQL method combining semantic parsing and semantic component matching | |
| CN115952282A (en) | Intelligent bank customer complaint diversion handling method and system based on NLP technology |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |