WO2020253043A1

WO2020253043A1 - Intelligent text classification method and apparatus, and computer-readable storage medium

Info

Publication number: WO2020253043A1
Application number: PCT/CN2019/117341
Authority: WO
Inventors: 郑子欧; 刘京华; 汪伟
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-06-20
Filing date: 2019-11-12
Publication date: 2020-12-24
Also published as: CN110413773A; CN110413773B

Abstract

The present application relates to artificial intelligence technology. Disclosed is an intelligent text classification method, comprising: receiving text data and a tag set, and performing part-of-speech tagging on the text data; performing fine-grained word segmentation on the text data according to the part-of-speech tagging in order to obtain a word segmentation sequence set, and performing word vectorization processing on the word segmentation sequence set in order to obtain a word vectorization data set; inputting the word vectorization data set and the tag set into a classification model for training and obtaining a training value, wherein the classification model quits the training when the training value is less than a preset threshold value; and receiving a text input by a user, performing the word vectorization operation on the text to obtain text word vectors, and inputting the text word vectors into the classification model for determination, and outputting a classification result. Further provided are an intelligent text classification apparatus and a computer-readable storage medium. The present application can realize an accurate text classification function.

Description

Intelligent text classification method, device and computer readable storage medium

This application is based on the Paris Convention declaration that it enjoys the priority of the Chinese patent application filed on June 20, 2019 with the application number CN201910540265.3 and titled "Smart text classification method, device and computer readable storage medium". This Chinese patent application The overall content of is incorporated in this application by reference.

Technical field

This application relates to the field of artificial intelligence technology, and in particular to an intelligent text classification method, device and computer-readable storage medium.

Background technique

Text classification is a very important part of text processing, and its applications are also very extensive, such as: spam filtering, news classification, part-of-speech tagging, etc. To classify the content of different texts, currently it is usually classified by labeling keywords. Such a classification method ignores the textual information in the text, and because this classification method lacks part of speech considerations, the classification of the text is not comprehensive and detailed, resulting in low accuracy.

Summary of the invention

This application provides an intelligent text classification method, device, and computer-readable storage medium, the main purpose of which is to provide users with accurate classification results when they input text.

In order to achieve the above purpose, an intelligent text classification method provided by this application includes:

Receive text data and a tag set, and perform part-of-speech tagging on the text data;

Performing fine-grained word segmentation of the text data according to the part-of-speech tagging to obtain a word segmentation sequence set, and performing word vectorization processing on the word segmentation sequence set to obtain a word vectorized data set;

Input the word vectorized data set and the label set into a classification model for training and obtain a training value, and when the training value is less than a preset threshold, the classification model exits the training;

The text input by the user is received, the word vectorization operation is performed on the text to obtain a text word vector, the text word vector is input to the classification model for judgment and the classification result is output.

In addition, in order to achieve the above object, the present application also provides an intelligent text classification device, which includes a memory and a processor. The memory stores a text classification program that can run on the processor. The text classification program When executed by the processor, the following steps are implemented:

In addition, in order to achieve the above-mentioned object, the present application also provides a computer-readable storage medium having a text classification program stored on the computer-readable storage medium, and the text classification program can be executed by one or more processors to achieve The steps of the smart text classification method as described above.

The intelligent text classification method, device and computer readable storage medium proposed in this application. This application performs part-of-speech tagging based on text content, which can effectively convert text data into part-of-speech data. At the same time, word vectorization can further interpret the characteristics of text data to a computer for analysis without loss. Multiple training based on classification models can be effective Improve the robustness and accuracy of text data classification types. Therefore, this application can provide users with accurate classification results.

Description of the drawings

FIG. 1 is a schematic flowchart of an intelligent text classification method provided by an embodiment of this application;

2 is a schematic diagram of the internal structure of an intelligent text classification device provided by an embodiment of the application;

FIG. 3 is a schematic diagram of modules of a text classification program in an intelligent text classification device provided by an embodiment of the application.

The realization, functional characteristics, and advantages of the purpose of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

Detailed ways

It should be understood that the specific embodiments described here are only used to explain the application, and are not used to limit the application.

This application provides an intelligent text classification method. Referring to FIG. 1, it is a schematic flowchart of a smart text classification method provided by an embodiment of this application. The method can be executed by a device, and the device can be implemented by software and/or hardware.

In this embodiment, the intelligent text classification method includes:

S1. Receive text data and a tag set, and perform part-of-speech tagging on the text data.

Preferably, the text data set includes text data of various subjects, such as finance, novels, education, real estate, sports, etc. The label set records the labels of each text data in the text data set, such as Record text data A for sports, text data B for real estate, etc.

In a preferred embodiment of the present application, the part-of-speech tagging first annotates the nouns and verbs in the text data according to a preset part-of-speech tagging template, where the part-of-speech tagging template refers to a recognizer that has marked the features of nouns and verbs. The predicate tag template can identify nouns and verbs by identifying the characteristics of words. For example, [I like to eat apples in particular], [Playing basketball is good for fitness], [The enemy succumbed at the last time], mark [my apple], [basketball fitness], [enemy time] according to the part-of-speech tag template Are nouns, [like to eat], [打益], and [屈服] are verbs;

Search for words with a length greater than a preset length in the text data, such as two characters and containing "的" or "地", and determine that the words with a length greater than two characters and containing "的" or "地" are in the place Whether the words before and after the predicate text data are nouns or verbs. If the preceding and following words are nouns or verbs, the words that are longer than the preset length and contain "的" or "地" are adjectives or adverbs, such as [Angry people fight the hateful thief fiercely], Firstly, according to the part-of-speech tagging template, [people], [厮打], [thief] are identified, and words that are longer than two characters and contain "的" or "地" are recognized as [angry], [給打地] ], [Hateful], it is judged that there are nouns or verbs such as [人], [厮打], [thief] before and after the word, so they are adjectives or adverbs and marked. Preferably, the labeling method may adopt a form including labeling symbols, such as [Angry ^adj people ⁿ fiercely ^adv fight ^v hate ^adj thief ⁿ ].

S2. Perform fine-grained word segmentation of the text data according to the part of speech tagging to obtain a word segmentation sequence set, and perform word vectorization processing on the word segmentation sequence set to obtain a word vectorized data set.

In a preferred embodiment of the present application, the fine-grained word segmentation refers to removing words that are not marked as nouns, verbs, adjectives, or adverbs in the text data, and obtaining a word segmentation sequence set based on the marked symbols. Preferably, the excluded words are called heteromorphic word sets, such as all English letters, Arabic numerals, Chinese numerals, punctuation marks, stop words, etc. The stop words include words such as "了", "于", etc. the ^[adj morning of a boundless ⁿ rain, heavy rain washed wet ⁿ ⁿ ^v regarded land, into a damp ^v ^adj mud ^n] after the fine-grained segmentation obtained ^[adj boundless morning of rain ⁿ Heavy rain ⁿ land ⁿ wetted ^v becomes ^v wet ^adj mud ⁿ ], and then based on the labeling symbols to obtain the word sequence set, such as [majestic heavy rain ^adj morning ⁿ heavy rain ⁿ land ⁿ wet ^v becomes ^v ^adj damp clay ^n] based on the reference symbol to give [boundless morning rain rain wet land red mud becomes damp].

Further, a classification probability model is established based on the word segmentation sequence set, a conditional probability model is constructed based on the classification probability model, and the cumulative sum operation is performed on the conditional probability model to obtain a log likelihood function, which maximizes the log likelihood However, the function solves the optimal solution, and the optimal solution is the word vectorized data set.

Preferably, the classification probability model

for:

Among them, X is the word segmentation sequence set, ω is the nouns, verbs, adjectives, and adverbs of the word segmentation sequence set, which can also be called characteristic words, and e is an infinite non-recurring decimal.

A transposed matrix X _ω, the cumulative summing operation X _ω of the [omega], is the accumulated sum operation:

Wherein, c is the number of data in the word segmentation sequence set, and V(ω _i ) is the word vectorized data set assuming that the word has been vectorized, which can be obtained by subsequently maximizing the log likelihood function.

The conditional probability model p(ω|V(ω _i )) is:

Among them, l ^ω represents the number of nodes in the Huffman coding, and the Huffman coding combined with the Huffman binary tree, the tree is the data elements (also called nodes) organized according to the branch relationship A non-linear data structure, a collection of several trees is called a forest. A binary tree is an ordered tree with at most two subtrees per node. The two subtrees are called the left subtree and the right subtree respectively. If there is a binary tree with the smallest path length, it is called a Huffman binary tree, so the ω is a leaf node, and the weight of each leaf node is expressed by Huffman coding. This application uses 0, 1 code Different arrangements to represent words,

Indicates the Huffman code corresponding to the j-th node in the path p ^ω , and the root node has no code,

Is the encoding of the word ω,

Represents the vector corresponding to the j-1th non-leaf node in the path p ^ω . Because the word ω is a leaf node, there is no corresponding vector.

Preferably, the log likelihood function ζ is

among them

It is a thesaurus, which includes all nouns, verbs, adjectives and adverbs in the word segmentation sequence set.

Further, maximizing the log likelihood function is:

among them,

Represents the partial derivative of the log likelihood function to the transposed matrix of the cumulative sum operation. The V(ω _i ) is continuously optimized based on the partial derivative, and the optimization process is:

Among them, η is the set learning rate, and the word vectorized data set V(ω) is obtained based on the above.

S3. Input the word vectorized data set and the label set into a classification model for training and obtain a training value. When the training value is less than a preset threshold, the classification model exits the training.

Preferably, the classification model of the present application includes a convolutional neural network, an activation function and a loss function. The convolutional neural network includes nineteen layers of convolutional layers, nineteen layers of pooling layers, and one layer of fully connected layers.

The inputting the word vectorized data set and the label set into a classification model for training and obtaining a training value, and when the training value is less than a preset threshold, the classification model exiting training includes:

Preferably, after the convolutional neural network receives the word vectorized data set, it inputs the word vectorized data set to the nineteen-layer convolutional layer and nineteen-layer pooling layer for convolution operation and The maximum pooling operation obtains a dimensionality reduction data set, and the dimensionality reduction data set is input to the fully connected layer.

Further, the fully connected layer receives the dimensionality reduction data set, calculates a prediction classification set in combination with the activation function, and inputs the prediction classification set and the label set into the loss function to calculate the loss Value, judging the magnitude relationship between the loss value and a preset threshold, until the loss value is less than the preset threshold, the classification model exits training.

The convolution operation described in the preferred embodiment of this application is:

Where ω'is the output data, ω is the input data, k is the size of the convolution kernel, s is the stride of the convolution operation, and p is the data zero-filling matrix. The pooling operation can select the maximum pooling operation, The maximum pooling operation is to select the largest value in the matrix data in the matrix to replace the entire matrix;

The activation function is:

Where y is the predicted classification set, and e is an infinite non-recurring decimal.

The loss value T in the preferred embodiment of the present application is:

Wherein, n is the data size of the prediction classification set, y _t is the label set, μ _t is the prediction classification set, and the preset threshold is generally set at 0.01.

S4. Receive the text input by the user, perform the word vectorization operation on the text to obtain a text word vector, input the text word vector to the classification model for judgment and output a classification result.

This application also provides an intelligent text classification device. Referring to FIG. 3, it is a schematic diagram of the internal structure of an intelligent text classification device provided by an embodiment of this application.

In this embodiment, the smart text classification device 1 may be a PC (Personal Computer, personal computer), or a terminal device such as a smart phone, a tablet computer, or a portable computer, or a server. The intelligent text classification device 1 at least includes a memory 11, a processor 12, a communication bus 13, and a network interface 14.

Wherein, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of the intelligent text classification device 1, for example, the hard disk of the intelligent text classification device 1. In other embodiments, the memory 11 may also be an external storage device of the smart text classification device 1, such as a plug-in hard disk equipped on the smart text classification device 1, a smart media card (SMC), and a secure digital (Secure Digital). Digital, SD) card, flash card (Flash Card), etc. Further, the memory 11 may also include both an internal storage unit of the intelligent text classification device 1 and an external storage device. The memory 11 can be used not only to store application software and various data installed in the intelligent text classification device 1, such as the code of the text classification program 01, etc., but also to temporarily store data that has been output or will be output.

In some embodiments, the processor 12 may be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip, and is used to run the program code or processing stored in the memory 11 Data, for example, execute text classification program 01, etc.

The communication bus 13 is used to realize the connection and communication between these components.

The network interface 14 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface), and is usually used to establish a communication connection between the device 1 and other electronic devices.

Optionally, the device 1 may also include a user interface. The user interface may include a display (Display) and an input unit such as a keyboard (Keyboard). The optional user interface may also include a standard wired interface and a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light emitting diode) touch device, etc. Among them, the display can also be called a display screen or a display unit as appropriate, and is used to display the information processed in the intelligent text classification device 1 and to display a visualized user interface.

FIG. 3 only shows the intelligent text classification device 1 with components 11-14 and the text classification program 01. Those skilled in the art can understand that the structure shown in FIG. 1 does not constitute a limitation on the intelligent text classification device 1. It may include fewer or more components than shown, or a combination of certain components, or a different component arrangement.

In the embodiment of the apparatus 1 shown in FIG. 3, the text classification program 01 is stored in the memory 11; when the processor 12 executes the text classification program 01 stored in the memory 11, the following steps are implemented:

Step 1: Receive text data and a tag set, and perform part-of-speech tagging on the text data.

In a preferred embodiment of the present application, the part-of-speech tagging first annotates nouns and verbs in the text data according to a preset part-of-speech tagging template, where the part-of-speech tagging template refers to a recognizer that has marked the features of nouns and verbs, The predicate tag template can identify nouns and verbs by identifying the characteristics of words. For example, [I like to eat apples in particular], [Playing basketball is good for fitness], [The enemy succumbed at the last time], mark [my apple], [basketball fitness], [enemy time] according to the part-of-speech tag template Is the name, and [likes to eat], [打益], [quit] are verbs;

Search for words with a length greater than a preset length in the text data, such as two characters and containing "的" or "地", and determine that the words with a length greater than two characters and containing "的" or "地" are in the place Whether the preceding and following words in the predicate text data are nouns or verbs. If the preceding and following words are nouns or verbs, the words that are longer than the preset length and contain "的" or "地" are adjectives or adverbs, such as [Angry people fight the hateful thief fiercely], Firstly, according to the part-of-speech tagging template, [people], [厮打], [thief] are identified, and words that are longer than two characters and contain "的" or "地" are recognized as [angry], [給打地] ], [Hateful], it is judged that there are nouns or verbs such as [人], [厮打], [thief] before and after the word, so they are adjectives or adverbs and marked. Preferably, the labeling method may adopt a form including labeling symbols, such as [Angry ^adj people ⁿ fiercely ^adv fight ^v hate ^adj thief ⁿ ].

Step 2: Perform fine-grained word segmentation of the text data according to the part of speech tagging to obtain a word segmentation sequence set, and perform word vectorization processing on the word segmentation sequence set to obtain a word vectorized data set.

Preferably, the classification probability model

for:

The conditional probability model p(ω|V(ω _i )) is:

Is the encoding of the word ω,

Preferably, the log likelihood function ζ is

among them

Further, maximizing the log likelihood function is:

among them,

Step 3: Input the word vectorized data set and the label set into a classification model for training and obtain a training value. When the training value is less than a preset threshold, the classification model exits the training.

The activation function is:

The loss value T in the preferred embodiment of the present application is:

Step 4: Receive the text input by the user, perform the word vectorization operation on the text to obtain a text word vector, input the text word vector to the classification model for judgment and output the classification result.

Optionally, in other embodiments, the text classification program may also be divided into one or more modules, and the one or more modules are stored in the memory 11 and are executed by one or more processors (in this embodiment, the processing The module 12) is executed to complete this application. The module referred to in this application refers to a series of computer program instruction segments that can complete specific functions, and is used to describe the execution process of the text classification program in the intelligent text classification device.

For example, referring to FIG. 4, which is a schematic diagram of the program modules of the text classification program in an embodiment of the intelligent text classification device of this application. In this embodiment, the text classification program can be divided into a part-of-speech tagging module 10 and word vectorization. The conversion module 20, the model training module 30, and the text classification result output module 40 are exemplary:

The part-of-speech tagging module 10 is configured to receive text data and a tag set, and perform part-of-speech tagging on the text data.

The word vectorization conversion module 20 is configured to perform fine-grained word segmentation of the text data according to the part of speech tagging to obtain a word segmentation sequence set, and perform word vectorization processing on the word segmentation sequence set to obtain a word vectorization data set.

The model training module 30 is configured to: input the word vectorized data set and the tag set into a classification model for training and obtain a training value, and when the training value is less than a preset threshold, the classification model exits training .

The text classification result output module 40 is configured to: receive text input by a user, perform the word vectorization operation on the text to obtain a text word vector, input the text word vector to the classification model for judgment and output the classification result .

The above-mentioned part-of-speech tagging module 10, word vectorization conversion module 20, model training module 30, and text classification result output module 40 implement functions or operation steps when executed, which are substantially the same as those in the foregoing embodiment, and will not be repeated here.

In addition, an embodiment of the present application also proposes a computer-readable storage medium having a text classification program stored on the computer-readable storage medium, and the text classification program can be executed by one or more processors to implement the following operations:

Receiving text data, and performing part-of-speech tagging on the text data to obtain text data;

The specific implementation of the computer-readable storage medium of this application is basically the same as the above-mentioned embodiments of the intelligent text classification device and method, and will not be repeated here.

It should be noted that the serial numbers of the above embodiments of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments. And the terms "include", "include" or any other variants thereof in this article are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements not only includes those elements, but also includes The other elements listed may also include elements inherent to the process, device, article, or method. If there are no more restrictions, the element defined by the sentence "including a..." does not exclude the existence of other identical elements in the process, device, article or method that includes the element.

Through the description of the above embodiments, those skilled in the art can clearly understand that the method of the above embodiments can be implemented by means of software plus the necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is better.的实施方式。 Based on this understanding, the technical solution of this application essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as ROM/RAM) as described above. , Magnetic disks, optical disks), including several instructions to make a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) execute the method described in each embodiment of the present application.

The above are only preferred embodiments of this application, and do not limit the scope of this application. Any equivalent structure or equivalent process transformation made using the content of the description and drawings of this application, or directly or indirectly used in other related technical fields , The same reason is included in the scope of patent protection of this application.

Claims

An intelligent text classification method, characterized in that the method includes:

Receive text data and a tag set, and perform part-of-speech tagging on the text data;

Performing fine-grained word segmentation of the text data according to the part-of-speech tagging to obtain a word segmentation sequence set, and performing word vectorization processing on the word segmentation sequence set to obtain a word vectorized data set;

Input the word vectorized data set and the label set into a classification model for training and obtain a training value, and when the training value is less than a preset threshold, the classification model exits the training;

The text input by the user is received, the word vectorization operation is performed on the text to obtain a text word vector, the text word vector is input to the classification model for judgment and the classification result is output.
5. The intelligent text classification method of claim 1, wherein the part-of-speech tagging comprises:

Label the nouns and verbs in the text data according to a preset part of speech tagging template;

Search for words in the text data that are longer than the preset length and contain "的" or "地";

Determine whether the preceding and following words in the text data of the word whose length is greater than the preset length and containing "的" or "地" are nouns or verbs, and if the preceding and following words are nouns or verbs, mark the length Words that are more than two characters and contain "的" or "地" are adjectives or adverbs.
3. The intelligent text classification method of claim 1 or 2, wherein the word vectorization processing includes:

Establishing a classification probability model based on the word segmentation sequence set;

Constructing a conditional probability model based on the classification probability model;

Performing cumulative summation operations on the conditional probability model to obtain a log likelihood function;

Maximizing the log likelihood function to solve an optimal solution, the optimal solution is the word vectorized data set.
The intelligent text classification method of claim 3, wherein the classification probability model
for:

Among them, X is the word segmentation sequence set, ω is the nouns, verbs, adjectives, and adverbs marked by the part of speech, which can also be called characteristic words, and e is the infinite non-recurring decimal.
A transposed matrix X ω, the cumulative summing operation X ω of the [omega], is the accumulated sum operation:

Wherein, c is the number of data in the word segmentation sequence set, and V(ω i ) is the word vectorized data set assuming that the word has been vectorized.
The intelligent text classification method according to claim 4, wherein the classification model includes a convolutional neural network, an activation function, and a loss function, wherein the convolutional neural network includes nineteen convolutional layers, nineteen Layer pooling layer and a fully connected layer; and

The inputting the word vectorized data set and the label set into a classification model for training and obtaining a training value, and when the training value is less than a preset threshold, the classification model exiting training includes:

After the convolutional neural network receives the word vectorized data set, it inputs the word vectorized data set to the nineteen-layer convolutional layer and nineteen-layer pooling layer for convolution operation and maximum pooling operation , Obtain a dimensionality reduction data set, and input the dimensionality reduction data set to the fully connected layer;

The fully connected layer receives the dimensionality reduction data set, calculates a prediction classification set in combination with the activation function, inputs the prediction classification set and the label set into the loss function to calculate a loss value, and judges The magnitude relationship between the loss value and a preset threshold value, until the loss value is less than the preset threshold value, the classification model exits training.
The intelligent text classification method of claim 5, wherein the convolution operation is:

Among them, ω'is output data, ω is input data, k is the size of the convolution kernel, s is the stride of the convolution operation, and p is the data zero-filling matrix.
The intelligent text classification method according to claim 5, wherein the loss function is:

Among them, T is the loss value, n is the data size of the predicted classification set, y t is the label set, and μ t is the predicted classification set.
An intelligent text classification device, characterized in that the device includes a memory and a processor, the memory stores a text classification program that can be run on the processor, and the text classification program is executed by the processor When implementing the following steps:

Receive text data and a tag set, and perform part-of-speech tagging on the text data;

Performing fine-grained word segmentation of the text data according to the part-of-speech tagging to obtain a word segmentation sequence set, and performing word vectorization processing on the word segmentation sequence set to obtain a word vectorized data set;

Input the word vectorized data set and the label set into a classification model for training and obtain a training value, and when the training value is less than a preset threshold, the classification model exits the training;

The text input by the user is received, the word vectorization operation is performed on the text to obtain a text word vector, the text word vector is input to the classification model for judgment and the classification result is output.
8. The intelligent text classification device of claim 8, wherein the part-of-speech tagging comprises:

Label the nouns and verbs in the text data according to a preset part of speech tagging template;

Search for words in the text data that are longer than the preset length and contain "的" or "地";

Determine whether the preceding and following words in the text data of the word whose length is greater than the preset length and containing "的" or "地" are nouns or verbs, and if the preceding and following words are nouns or verbs, mark the length Words that are more than two characters and contain "的" or "地" are adjectives or adverbs.
The intelligent text classification device according to claim 8 or 9, wherein the word vectorization processing includes:

Establishing a classification probability model based on the word segmentation sequence set;

Constructing a conditional probability model based on the classification probability model;

Accumulating and summing the conditional probability model to obtain a log likelihood function;

The optimal solution is solved by maximizing the log likelihood function, and the optimal solution is the word vectorized data set.
The intelligent text classification device of claim 10, wherein the classification probability model
for:

Among them, X is the word segmentation sequence set, ω is the nouns, verbs, adjectives, and adverbs marked by the part of speech, which can also be called characteristic words, and e is the infinite non-recurring decimal.
A transposed matrix X ω, the cumulative summing operation X ω of the [omega], is the accumulated sum operation:

Wherein, c is the number of data in the word segmentation sequence set, and V(ω i ) is the word vectorized data set assuming that the word has been vectorized.
The intelligent text classification device of claim 11, wherein the classification model includes a convolutional neural network, an activation function and a loss function, wherein the convolutional neural network includes nineteen convolutional layers, nineteen Layer pooling layer and a fully connected layer; and

The inputting the word vectorized data set and the label set into a classification model for training and obtaining a training value, and when the training value is less than a preset threshold, the classification model exiting training includes:

After the convolutional neural network receives the word vectorized data set, it inputs the word vectorized data set to the nineteen-layer convolutional layer and nineteen-layer pooling layer for convolution operation and maximum pooling operation , Obtain a dimensionality reduction data set, and input the dimensionality reduction data set to the fully connected layer;

The fully connected layer receives the dimensionality reduction data set, calculates a prediction classification set in combination with the activation function, inputs the prediction classification set and the label set into the loss function to calculate a loss value, and judges The magnitude relationship between the loss value and a preset threshold value, until the loss value is less than the preset threshold value, the classification model exits training.
The intelligent text classification device of claim 12, wherein the convolution operation is:

Among them, ω'is output data, ω is input data, k is the size of the convolution kernel, s is the stride of the convolution operation, and p is the data zero-filling matrix.
The intelligent text classification device of claim 12, wherein the loss function is:

Among them, T is the loss value, n is the data size of the predicted classification set, y t is the label set, and μ t is the predicted classification set.
A computer-readable storage medium, characterized in that a text classification program is stored on the computer-readable storage medium, and the text classification program can be executed by one or more processors to implement the following steps:

Receive text data and a tag set, and perform part-of-speech tagging on the text data;

Performing fine-grained word segmentation of the text data according to the part-of-speech tagging to obtain a word segmentation sequence set, and performing word vectorization processing on the word segmentation sequence set to obtain a word vectorized data set;

Input the word vectorized data set and the label set into a classification model for training and obtain a training value, and when the training value is less than a preset threshold, the classification model exits the training;

The text input by the user is received, the word vectorization operation is performed on the text to obtain a text word vector, the text word vector is input to the classification model for judgment and the classification result is output.
15. The computer-readable storage medium according to claim 15, wherein the part-of-speech tag comprises:

Label the nouns and verbs in the text data according to a preset part of speech tagging template;

Search for words in the text data that are longer than the preset length and contain "的" or "地";

Determine whether the preceding and following words in the text data of the word whose length is greater than the preset length and containing "的" or "地" are nouns or verbs, and if the preceding and following words are nouns or verbs, mark the length Words that are more than two characters and contain "的" or "地" are adjectives or adverbs.
The computer-readable storage medium according to claim 15 or 16, wherein the word vectorization processing includes:

Establishing a classification probability model based on the word segmentation sequence set;

Constructing a conditional probability model based on the classification probability model;

Accumulating and summing the conditional probability model to obtain a log likelihood function;

The optimal solution is solved by maximizing the log likelihood function, and the optimal solution is the word vectorized data set.
The computer-readable storage medium of claim 17, wherein the classification probability model
for:

Among them, X is the word segmentation sequence set, ω is the nouns, verbs, adjectives, and adverbs marked by the part of speech, which can also be called characteristic words, and e is the infinite non-recurring decimal.
A transposed matrix X ω, the cumulative summing operation X ω of the [omega], is the accumulated sum operation:

Wherein, c is the number of data in the word segmentation sequence set, and V(ω i ) is the word vectorized data set assuming that the word has been vectorized.
The computer-readable storage medium according to claim 18, wherein the classification model includes a convolutional neural network, an activation function, and a loss function, wherein the convolutional neural network includes nineteen convolutional layers, ten Nine pooling layers and one fully connected layer; and

The inputting the word vectorized data set and the label set into a classification model for training and obtaining a training value, and when the training value is less than a preset threshold, the classification model exiting training includes:

After the convolutional neural network receives the word vectorized data set, it inputs the word vectorized data set to the nineteen-layer convolutional layer and nineteen-layer pooling layer for convolution operation and maximum pooling operation , Obtain a dimensionality reduction data set, and input the dimensionality reduction data set to the fully connected layer;

The fully connected layer receives the dimensionality reduction data set, calculates a prediction classification set in combination with the activation function, inputs the prediction classification set and the label set into the loss function to calculate a loss value, and judges The magnitude relationship between the loss value and a preset threshold value, until the loss value is less than the preset threshold value, the classification model exits training.
The computer-readable storage medium of claim 19, wherein the convolution operation is:

Among them, ω'is output data, ω is input data, k is the size of the convolution kernel, s is the stride of the convolution operation, and p is the data zero-filling matrix.