KR102443629B1 - Solution and system for news positive tendency analysis using deep learning nlp model - Google Patents

Abstract

The present disclosure relates to a news article analysis system for analyzing a news article, and a method for the news article analysis system to analyze a news article. According to an embodiment, a method for a news article analysis system to analyze a news article may include: acquiring, by an electronic device in the news article analysis system, a news article from at least one external device connected to the electronic device; transmitting, by the electronic device, the obtained news article to a server connected to the electronic device; identifying, by the server, a sentence including a pre-stored financial word or a person word in the news article; obtaining, by the server, a first news score from the neural network model by inputting, by the server, some sentences of the identified sentences into a neural network model that is trained to output a score for the news article; obtaining, by the server, a second news score from the pre-learning model by inputting some of the identified sentences into the pre-learning model; and transmitting, by the server, the first news score and the second news score to the electronic device. may include.

Description

News positive analysis solution and system using deep learning NLP model {SOLUTION AND SYSTEM FOR NEWS POSITIVE TENDENCY ANALYSIS USING DEEP LEARNING NLP MODEL}

The present disclosure relates to systems and methods for analyzing news articles. More particularly, it relates to a method and apparatus for analyzing news articles in the financial field.

With the development of Internet technology, media media such as newspapers, broadcasters, and magazines are delivering news articles through the Internet rather than offline print media. In addition, through the Internet, not only domestic news including politics, economy, society and culture in general, but also news in the political, economic, and financial fields of countries around the world are being published.

In particular, with the development of the growing global economy, information on news articles in the financial field is also overflowing on the Internet. These financial news articles cover both the positive and negative aspects of these financial issues.

With the development of the Internet, data processing technologies have been developed to process a large number of news articles produced, but general news data processing technologies still have many limitations in objectively evaluating and analyzing a large number of news articles.

Accordingly, there is a demand for development of a method and device technology for accurately and objectively evaluating a large number of Internet news articles.

Korean Patent Publication No. 10-2009-0047646

According to an embodiment, a method and system for analyzing a news article may be provided.

According to an embodiment, a method of analyzing a news article by interworking between an electronic device and a server may be provided.

According to an embodiment of the present disclosure for achieving the above technical problem, an electronic device in a news article analysis system obtains a news article from at least one external device connected to the electronic device; transmitting, by the electronic device, the obtained news article to a server connected to the electronic device; identifying, by the server, a sentence including a pre-stored financial word or a person word in the news article; obtaining, by the server, a first news score from the neural network model by inputting, by the server, some sentences of the identified sentences into a neural network model that is trained to output a score for the news article; obtaining, by the server, a second news score from the pre-learning model by inputting some of the identified sentences into the pre-learning model; and transmitting, by the server, the first news score and the second news score to the electronic device. A method for analyzing a news article may be provided, including:

In addition, according to another embodiment of the present disclosure for solving the technical problem, a sentence including a financial word or a person word stored in advance in a news article is identified, and some of the identified sentences are used for the news article. A first news score is obtained from the neural network model by input to the neural network model that is trained to output a score, and a second news score is obtained from the pre-trained model by inputting some sentences of the identified sentences into the pre-learning model. server to; and an electronic device for obtaining a news article from an external device, transmitting the obtained news article to the server, and obtaining the first news score and the second news score from the server; A news article analysis system comprising a may be provided.

In addition, according to another embodiment of the present disclosure for solving the technical problem, an electronic device in the news article analysis system may include: acquiring a news article from at least one external device connected to the electronic device; transmitting, by the electronic device, the obtained news article to a server connected to the electronic device; identifying, by the server, a sentence including a pre-stored financial word or a person word in the news article; obtaining, by the server, a first news score from the neural network model by inputting, by the server, some sentences of the identified sentences into a neural network model that is trained to output a score for the news article; obtaining, by the server, a second news score from the pre-learning model by inputting some of the identified sentences into the pre-learning model; and transmitting, by the server, the first news score and the second news score to the electronic device. A computer-readable recording medium recording a program for executing a method of analyzing a news article on a computer, including a computer-readable recording medium, may be provided.

According to an embodiment, since the electronic device and the server interwork with each other, it is possible to effectively evaluate and analyze news articles in the financial field.

According to an embodiment, the news article may be evaluated using at least one news article analysis model.

1 is a diagram schematically illustrating a method of analyzing a news article by a news article analysis system according to an exemplary embodiment.
2 is a flowchart of a method for analyzing a news article by a news article analysis system according to an exemplary embodiment.
3 is a flowchart of a method for analyzing a news article by a news article analysis system according to another exemplary embodiment.
4 is a diagram for explaining a financial word list and a person weight list stored by the news article analysis system according to an exemplary embodiment.
FIG. 5 is a diagram for explaining a process in which a news article analysis system extracts a sentence from a news article and generates summary data using the extracted sentence, according to an exemplary embodiment.
6 is a diagram for explaining a method for a news article analysis system to obtain a first news score by using a neural network model according to an embodiment.
7 is a flowchart of a method for a news article analysis system to obtain a second news score by using a pre-learning model, according to an embodiment.
8 is a flowchart illustrating a process in which a news article analysis system pre-processes a preset number of sentences according to an exemplary embodiment.
9 is a diagram for describing sentence weight elements used by a news article analysis system to determine a sentence weight determined for each sentence according to an embodiment.
10 is a flowchart illustrating a method for a news article analysis system to determine a comprehensive evaluation score for a news article based on a first news score and a second news score according to an exemplary embodiment.
11 is a flowchart illustrating a method for a news article analysis system to determine a comprehensive evaluation score based on a first news score and a second news score according to another exemplary embodiment.
12 is a block diagram of an electronic device according to an embodiment.
13 is a block diagram of a server according to an embodiment.
14 is a flowchart illustrating a method of analyzing a news article by a news article analysis system according to another exemplary embodiment.

Terms used in this specification will be briefly described, and the present disclosure will be described in detail.

The terms used in the present disclosure have been selected as currently widely used general terms as possible while considering the functions in the present disclosure, but these may vary depending on the intention or precedent of a person skilled in the art, the emergence of new technology, and the like. In addition, in a specific case, there is a term arbitrarily selected by the applicant, and in this case, the meaning will be described in detail in the description of the corresponding invention. Therefore, the terms used in the present disclosure should be defined based on the meaning of the term and the contents of the present disclosure, rather than the simple name of the term.

Further, in this specification, when a film (or layer) is referred to as being on another film (or layer) or substrate, it may be formed directly on the other film (or layer) or substrate, or a third layer between them. In addition, in the drawings, the size and thickness of the components are exaggerated for clarity. In various embodiments of the present specification, the terms first, second, third, etc. are used to describe various regions, films (or layers), etc., but these regions and films should not be limited by these terms. do. These terms are only used to distinguish one region or film (or layer) from another region or film (or layer). Accordingly, a film quality referred to as a first film quality in one embodiment may be referred to as a second film quality in another embodiment. Each embodiment described and illustrated herein also includes a complementary embodiment thereof. In this specification, the expression 'and/or' is used in a sense including at least one of the elements listed before and after. Parts indicated with like reference numerals throughout the specification indicate like elements.

In the entire specification, when a part "includes" a certain component, it means that other components may be further included, rather than excluding other components, unless otherwise stated. In addition, terms such as "...unit" and "module" described in the specification mean a unit that processes at least one function or operation, which may be implemented as hardware or software, or a combination of hardware and software. .

Hereinafter, with reference to the accompanying drawings, embodiments of the present disclosure will be described in detail so that those of ordinary skill in the art to which the present disclosure pertains can easily implement them. However, the present disclosure may be implemented in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present disclosure in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

1 is a diagram schematically illustrating a method of analyzing a news article by a news article analysis system according to an exemplary embodiment.

According to an embodiment, the news article analysis system 3000 may include the electronic device 1000 and the server 2000 . For example, the news article analysis system 3000 may obtain data on a news article and output a comprehensive evaluation score 112 on the obtained news article. The news article analysis system 3000 may determine a comprehensive evaluation score indicating a financial propensity for a news article by using a plurality of models for analyzing different news articles.

According to an embodiment, the electronic device 1000 and the server 2000 in the news article analysis system 3000 may be connected to each other through a network. According to an embodiment, the electronic device 1000 obtains at least one news article from an external device connected to the electronic device, and uses a plurality of analysis models for directly analyzing the news article, and comprehensively evaluates the news article. The score may be determined, but the news article is transmitted to a server connected to the electronic device, the first news score and the second news score are received from the server, and the news is based on the received first news score and the second news score. You can also print the overall evaluation score for the article.

According to an embodiment, the electronic device 1000 includes a smart phone, a tablet PC, a smart TV, a mobile phone, a media player, a server, and a microphone having an AI program and a voice recognition function for processing data related to news articles. It may be, but is not limited to, a server or other mobile or non-mobile computing device. Also, according to an embodiment, the server 2000 may include other computing devices that can transmit and receive data to and from the electronic device by being connected to the electronic device 1000 through a network. Also, according to an embodiment, the server device 2000 may be a Wearable Business Management Server (W-BMS) for managing the wearable device.

According to an embodiment, the server 2000 includes a local area network (LAN), a wide area network (WAN), a value added network (VAN), a mobile radio communication network, It is a data communication network in a comprehensive sense that includes a satellite communication network and a combination thereof, and allows each network constituent entity shown in FIG. 1 to communicate smoothly with each other, and includes a wired Internet, a wireless Internet, and a mobile wireless communication network. can

According to an embodiment, the news article analysis system 3000 obtains a news article 101 from an external device, and inputs the obtained news article 101 to the neural network model 102 and the pre-training model 104, respectively. , a first news score 106 and a second news score 108 may be obtained. According to an embodiment, the news article 101 may be obtained in the form of text data, image data, or image data. The news article analysis system 3000 may determine the overall evaluation score 112 based on the first news score and the second news score. The news article analysis system may output the determined overall evaluation score 112 .

According to an embodiment, the news article analysis system 3000 may include a plurality of news article analysis models for analyzing news articles. According to an embodiment, the news article analysis model may include a neural network model 102 and a pre-training model 104 . However, according to another embodiment, the news article analysis model may include other artificial intelligence models that can be learned according to an artificial intelligence learning algorithm.

According to an embodiment, the comprehensive evaluation score 110 may be a numerical value representing the financial propensity indicated by the news article based on at least one of a financial word and a person word included in the news article. According to an embodiment, the comprehensive evaluation score 110 is a financial word used by the layers in the neural network model 102 and the weight of the connection strength between the layers is modified and updated, or the pre-learning model 104 is used. Word scores or person weights in the list and person weight list may be changed when they are changed.

According to an embodiment, the neural network model used by the electronic device 1000 may include a deep neural network (DNN), for example, a convolutional neural network (CNN), a deep neural network (DNN), or an RNN. (Recurrent Neural Network), RBM (Restricted Boltzmann Machine), DBN (Deep Belief Network), BRDNN (Bidirectional Recurrent Deep Neural Network), or deep Q-Networks, etc., but are also not limited thereto.

2 is a flowchart of a method for analyzing a news article by a news article analysis system according to an exemplary embodiment.

In S202 , the electronic apparatus 1000 may obtain a news article from an external device connected to the electronic apparatus 1000 . For example, the electronic device 1000 may acquire information about a news article in at least one of a text format, an image format, and a video format. In S204, the electronic device 1000 may transmit the obtained news article to the server.

In S206 , the server 2000 may identify a sentence including a finance word or a person word stored in advance in a news article obtained from the electronic device 1000 . According to an embodiment, the server 2000 may generate a financial word list by matching the financial word and word scores for each financial word, and may store the generated financial word list. Also, the server 2000 may generate a person weight list by matching the person word and the person word weight for each person word, and store the generated person weight list. According to an embodiment, the person word may be predetermined to be related to the field of finance. According to an embodiment, the person word may be determined based on whether the number of publications of the person word in the published financial field article is equal to or greater than a predetermined threshold during a preset period.

However, according to another embodiment, the financial word list and the person weight list may be stored in advance in the electronic device 1000, and in this case, the electronic device 1000 may include a financial word or a person word from a news article. You can also directly identify the sentence.

In S208, the server 2000 may obtain a first news score by inputting some of the identified sentences to the neural network model. According to an embodiment, when summary data for a news article is input, the server 2000 may store a pre-trained neural network model in advance to output a score for the news article, and use the pre-trained neural network model to write to the news article. It is possible to obtain a first news score for the

In S210 , the server 2000 may obtain a second news score by inputting some of the identified sentences to the pre-learning model. According to an embodiment, the dictionary learning model may identify a predetermined word in an article using a dictionary, and output a second news score based on a word score assigned to the identified word.

In S212 , the server 2000 may determine a comprehensive evaluation score regarding the financial propensity of the news article based on the first news score and the second news score. For example, the server 2000 may obtain a neutral index, which is a probability value regarding the degree of confidence in the second news score, in addition to the second news score from the pre-learning model. The server 2000 determines an evaluation weight to be applied to the first news score and the second news score based on the neutral index value, and weights and sums the first news score and the second news score according to the determined evaluation weight to obtain a comprehensive evaluation score. can be decided

In S214 , the server 2000 may transmit information on the determined comprehensive evaluation score to the electronic device 1000 . In S216 , the electronic device 1000 may output a comprehensive evaluation score based on information about the comprehensive evaluation score received from the server. According to an embodiment, the electronic device 1000 may output the comprehensive evaluation score in an audio format or may output it through a display (not shown) in an output unit (not shown) of the electronic device.

3 is a flowchart of a method for analyzing a news article by a news article analysis system according to another exemplary embodiment.

Since S302 to S310 of FIG. 3 may correspond to S202 to S210 of FIG. 2 , a detailed description thereof will be omitted.

In S312 , the server 2000 may transmit information on the first news score and the second news score to the electronic device 1000 . In S314 , the electronic device 1000 may determine a comprehensive evaluation score regarding the financial propensity of a news article based on the first news score and the second news score. S312 may correspond to S212 of FIG. 2 . That is, in the case of FIG. 2 , the server 2000 determines the overall evaluation score using the first news score output from the neural network model and the second news score output from the pre-learning model, but in FIG. 3 , the electronic device ( 1000) may determine the overall evaluation score by using information about the first news score and the second news score obtained from the server 2000 . In S316 , the electronic device 1000 may output a comprehensive evaluation score for the determined news article.

4 is a diagram for explaining a financial word list and a person weight list stored by the news article analysis system according to an exemplary embodiment.

According to an embodiment, the news article analysis system 3000 may generate the financial word list 410 by matching the financial word and the word score for each financial word. According to an embodiment, the news article analysis system 3000 may acquire expert evaluation scores of experts for preset financial words. According to an embodiment, the news article analysis system 3000 may obtain an expert evaluation score for a specific financial word from a plurality of experts, and generate a word score vector based on the obtained expert evaluation score. According to an embodiment, the word score vector may include the evaluation score of each expert as a vector element. The news article analysis system 3000 may identify the mean and standard deviation of the elements in the word score vector. According to an embodiment, the news article analysis system 3000 may generate the financial word list 410 by converting the financial words into a headword form, and matching the financial words converted into the headword form with word scores for each financial word. have.

For example, the news article analysis system 3000 obtains 2 points, 2 points, 3 points, 2 points, and 1 points, respectively, as user evaluation scores for 'competent' from 5 financial experts, and financial experts A word vector score {2,2,3,2,1} may be generated using each expert evaluation score obtained from . The news article analysis system 3000 may identify the mean and standard deviation of the vector elements of the generated word vector score as 2 and 0.632456, respectively. According to an embodiment, the expert evaluation score obtained by the news article analysis system 3000 may include a value between -5 and 5, but is not limited thereto.

According to an embodiment, the expert evaluation score obtained by the news article analysis system 3000 may be an index value indicating the degree of positive or negative degree of each expert with respect to the corresponding word. According to an embodiment, the expert evaluation score obtained by the news article analysis system 3000 may indicate the following positive or negative degrees. For example, an expert rating score between -5 and 5 may indicate a positive or negative degree of a financial word as follows (eg, -5: extremely bad, -4: very bad, -3: not) good, -2: slightly bad, -1: vaguely bad, 0: average, 1: vaguely good, 2: slightly good, 3: good, 4: very good, 5: extremely good). However, the present invention is not limited thereto, and the range of evaluation scores of experts for financial words may vary.

According to an embodiment, the news article analysis system 3000 may acquire evaluation scores for a predetermined person word from a plurality of financial experts. The news article analysis system 3000 may generate a person word weight based on evaluation scores for the person word obtained from a plurality of financial experts. According to an embodiment, a weight for each person word may be expressed in a vector form. The news article analysis system 3000 may generate the person weight list 420 by matching the person word and the person word weight for each person word.

For example, the news article analysis system 3000 may obtain an evaluation score for 'Warren Edward Buffet', the chairman of Berkshire Hathaway from five financial experts, respectively, and generate a person weight based on the obtained evaluation score. can According to an embodiment, the expert evaluation score for the person word obtained by the news article analysis system 3000 may include a value between -5 and 5. However, the present invention is not limited thereto, and the range of the expert evaluation score for the person word obtained by the news article analysis system 3000 may vary. The news article analysis system 3000 may store the financial word list 410 and the person weight list 420 described above in advance in at least one of an electronic device or a server in the news analysis system 3000 .

FIG. 5 is a diagram for explaining a process in which a news article analysis system extracts a sentence from a news article and generates summary data using the extracted sentence, according to an exemplary embodiment.

According to an embodiment, the news article analysis system 3000 may obtain a news article and identify each sentence in the obtained news article. For example, the news article analysis system 3000 may identify a sentence including a preset financial word and a person word from the news article. For example, the news article analysis system 3000 may obtain a news article 510 and identify sentences 1 to 6 including at least one of a predetermined financial word or a person word in the acquired news article 510 . can

According to an embodiment, the news article analysis system 3000 may identify a sentence included in a news article, and identify position information in each of the identified sentences. The news article analysis system 3000 may identify the number of the identified sentence in the news article. The news article analysis system 3000 may identify the first sentence in the news article and the last sentence in the news article based on the position information of the identified sentence.

According to an embodiment, the news article analysis system 3000 may extract a predetermined sentence from among the sentences identified in the news article. For example, the news article analysis system 3000 may extract sentences 1 to 4 and sentence 6 from among the sentences extracted from the news article 510 . The news article analysis system 3000 may generate a summary news article 520 by extracting some sentences from among sentences including at least one of a predetermined financial word or a person word in the news article, and extracting the extracted some sentences. have.

According to an embodiment, the news article analysis system 3000 selects a predetermined number of sentences to be scored among sentences including the financial word or the person word, the first sentence in the news article, and the last sentence in the news article. It is also possible to extract and generate a summary news article 520 by using the extracted pre-set number of sentences. News article analysis system 3000 may generate summary data 530 by encoding summary news articles.

According to an embodiment, the news article analysis system 3000 may generate text data for a summarized news article as the summary data 530 . According to another embodiment, the news article analysis system 3000 may generate the summary data 530 by converting the summary news article into binarized data by encoding the summary news article.

6 is a diagram for explaining a method for a news article analysis system to obtain a first news score by using a neural network model according to an embodiment.

According to an embodiment, the electronic device 1000 inputs the summary data 602 generated according to the process described above with reference to FIG. 5 into the neural network model 610 , and the first news score output from the neural network model 610 . (612) can be obtained. According to an embodiment, the neural network model 610 may include at least one layer including at least one node and a weight related to the connection strength of the layers.

According to an embodiment, the neural network model used by the news article analysis system 3000 may include a deep neural network (DNN), for example, a Convolutional Neural Network (CNN), a Deep Neural Network (DNN) , RNN (Recurrent Neural Network), RBM (Restricted Boltzmann Machine), DBN (Deep Belief Network), BRDNN (Bidirectional Recurrent Deep Neural Network) or Deep Q-Networks, etc., but are also limited thereto not.

According to an embodiment, the news article analysis system 3000 may obtain a news article from an external device connected to the news article analysis system 3000 according to a preset period, and generate learning data from the obtained news article. According to an embodiment, the news article analysis system 3000 may further generate verification data from the obtained news articles in addition to the learning data. The news article analysis system 3000 trains the neural network model based on the training data, and the output value of the neural network model output by inputting training data from the neural network model, and the neural network model output by inputting verification data from the neural network model The neural network model may be trained so that the difference (eg, loss loss) between the output values of is small.

The operation of the news article analysis system 3000 to train the neural network model may correspond to the operation of modifying and updating the layers in the neural network model and weights related to the connection strength of the layers. The news article analysis system 3000 may acquire a first news score 612 by inputting summary data obtained from a predetermined news article into the neural network model 610 on which the training has been completed.

Also, according to an embodiment, the news article analysis system 3000 identifies a plurality of sentences in a news article received from an external device, pre-processes some extracted sentences from among the identified sentences, and pre-processes some of the pre-processed sentences. By inputting to the learned neural network model 610, it is possible to obtain a first news score from the neural network model.

7 is a flowchart of a method for a news article analysis system to obtain a second news score by using a pre-learning model, according to an embodiment.

In S712 , the news article analysis system 3000 may pre-process some extracted sentences among predetermined sentences identified in the news article. According to an embodiment, the operation of pre-processing the extracted partial sentences by the news article analysis system 3000 may correspond to the operation of pre-processing the sentences in the summary data described above with reference to FIG. 5 . An operation in which the news article analysis system 3000 pre-processes some extracted sentences will be described in detail with reference to FIG. 8, which will be described later.

In S714, the news article analysis system 3000 may determine a sentence weight for each of the pre-processed sentences based on at least one of negative phrases, adverbs, punctuation marks, emphasis phrases, negative words, or person words in the pre-processed sentence. have. For example, the news article analysis system 3000 may determine a person word weight pre-allocated with respect to a person word included in the pre-processed sentence as a weight of a sentence including the person word.

According to an embodiment, when a negative phrase such as not or isn't is included in the pre-processed sentence, the news article analysis system 3000 adds -1 to the sentence weight of the pre-processed sentence including the negative phrase. can be multiplied That is, when a negative phrase such as not or isn't is included in the pre-processed sentence, the news article analysis system 3000 may determine the sentence weight of the pre-processed sentence including the negative phrase as a negative number.

According to an embodiment, the news article analysis system 3000 may determine a sentence weight for each pre-processed sentence based on an adverb included in the pre-processed sentence. For example, when a positive adverb (eg, absolutely, amazingly) is included in the pre-processed sentence, the news article analysis system 3000 determines the sentence weight of the sentence including the positive adverb based on the number of included positive adverbs. A predetermined weight value may be added. According to an embodiment, when a positive adverb is included in the pre-processed sentence, the news article analysis system 3000 may increase the sentence weight of the corresponding sentence by 0.3, but the degree of increase in the sentence weight may vary. Also, according to an embodiment, the news article analysis system 3000 may increase the weight by 0.6 when two positive adverbs are included in the pre-processed sentence, but is not limited thereto.

According to another embodiment, when the news article analysis system 3000 includes negative adverbs (eg, scarcely, hardly, etc.) in the pre-processed sentence, based on the number of included negative adverbs, including the negative adverb A predetermined weight value may be subtracted from the sentence weight of the sentence. According to an embodiment, the news article analysis system 3000 may reduce the sentence weight of the corresponding sentence by 0.3 when a negative adverb is included in the pre-processed sentence, but is not limited thereto. Also, according to an embodiment, the news article analysis system 3000 may reduce the weight by 0.6 when two negative adverbs are included in the pre-processed sentence, but is not limited thereto.

According to an embodiment, the news article analysis system 3000 may determine a sentence weight in a corresponding sentence based on the pre-processed punctuation marks in the sentence. For example, when the pre-processed sentence includes an exclamation point punctuation mark for emphasis, the news article analysis system 3000 may increase the sentence weight of the sentence including the exclamation point. According to an embodiment, the operation of the news article analysis system 3000 increasing the sentence weight may include adding a predetermined weight value or multiplying a scale value according to a predetermined scale to scale-up the sentence weight. It can correspond to the action.

However, according to another embodiment, when the pre-processed sentence includes a punctuation mark such as a question mark or a comma, the news article analysis system 3000 may reduce the sentence weight of the corresponding sentence. According to an embodiment, the operation of the news article analysis system 3000 to reduce the sentence weight is to scale-down the sentence weight by subtracting a predetermined weight value or multiplying a scale value according to a predetermined scale. ) can correspond to the action. For example, when a pre-processed sentence having a sentence weight of 0.3 ends with an exclamation mark punctuation mark, the news article analysis system 3000 may scale up the sentence weight 0.3 by 50% to determine it as 0.45.

According to an embodiment, the news article analysis system 3000 may increase the weight of the pre-processed sentence when an emphasis phrase (eg, very) is included in the sentence. According to another embodiment, when a negative word (eg, never, so, without) is included in the pre-processed sentence, the news article analysis system 3000 increases or decreases the sentence weight of the pre-processed sentence including the negative word. can do.

For example, the news article analysis system 3000 sets the weight of the increased sentence due to one positive adverb to 0.3 when the pre-processed sentence includes one positive adverb, ends with an exclamation point, and includes the negative phrase not. , and the sentence weight is scaled up to 0.45 due to the exclamation mark, and since the negative phrase not is included, the scaled-up sentence weight 0.45 can be changed to -0.45. Accordingly, when the pre-processed sentence includes one positive adverb, ends with an exclamation point, and includes the negative phrase not, the news article analysis system 3000 may determine the sentence weight of the corresponding sentence to be −0.45. However, the degree to which the news article analysis system 3000 increases or decreases the sentence weight is not limited thereto.

In S716, the news article analysis system 3000 may determine the sentence score of each pre-processed sentence by applying the sentence weight determined for each pre-processed sentence to the word score of each financial word in the pre-processed sentence. . For example, the news article analysis system 3000 may determine the sentence word score by identifying preset words in the pre-processed sentence and summing the word scores assigned to each of the identified words. The news article analysis system 3000 may finally determine the sentence score by multiplying the sentence word score by the sentence weight.

For example, if the sum of the scores of financial words included in one pre-processed sentence is 0.4, the pre-processed sentence contains one positive adverb, ends with an exclamation mark, and includes the negative phrase not, By applying a sentence weight of -0.45 to the sentence word score of 0.4, the sentence score of the pre-processed sentence may be determined to be -0.18. According to another embodiment, the sum of the scores of financial words included in one pre-processed sentence is 0.4, the pre-processed sentence includes one positive adverb, ends with an exclamation mark, and does not include negative phrases In this case, by applying a sentence weight of 0.45 to the sentence word score of 0.4 (eg, 0.4 * 0.45), the sentence score of the pre-processed sentence may be determined to be 0.18.

According to another embodiment, the news article analysis system 3000 applies the sentence weight determined for each sentence to the word score of each financial word included in each pre-processed sentence, and then weights and sums the word scores to which the sentence weight is applied. It is also possible to determine the sentence score of the pre-processed sentence.

In S718 , the news article analysis system 3000 may adjust the sentence score determined for each of the pre-processed sentences based on the identified location information. Also, according to an embodiment, the news article analysis system 3000 may adjust the sentence score determined for each of the pre-processed sentences based on the location information of the pre-processed sentence and the conjunction included in the pre-processed sentence. have. According to an embodiment, the operation of the news article analysis system 3000 adjusting the determined sentence score for each of the pre-processed sentences based on the identified position information includes the position information identified by the news article analysis system 3000 . may correspond to an operation of adjusting a sentence weight determined for each of the pre-processed sentences based on , and applying the adjusted sentence weight to a sentence score determined for each of the pre-processed sentences.

For example, the news article analysis system 3000 may identify location information regarding the location of the pre-processed sentence in the news article. The news article analysis system 3000 may identify contexts between pre-processed sentences based on location information of the pre-processed sentences, and adjust sentence scores determined for each of the pre-processed sentences based on the identified contexts. According to an embodiment, the context identified by the news article analysis system 3000 based on the location information of the pre-processed sentences may mean a 'context' or a 'context' between the pre-processed sentences.

For example, the news article analysis system 3000 may identify whether the pre-processed sentence is a continuous sentence based on the location information of the pre-processed sentence. When the pre-processed sentences are consecutive sentences, the news article analysis system 3000 may further assign additional weights to all sentence weights determined for the pre-processed sentences. According to an embodiment, when the pre-processed sentences are consecutive sentences, the news article analysis system 3000 may increase the sentence weights determined for each of the pre-processed sentences by 0.1, but is not limited thereto. The news article analysis system 3000 may adjust the sentence scores of each pre-processed sentence by applying the increased sentence weights to each of the pre-processed sentences.

Also, according to an embodiment, the news article analysis system 3000 may adjust the sentence score for each of the pre-processed sentences based on the location information of the pre-processed sentence and the conjunction included in the pre-processed sentence. . For example, the news article analysis system 3000 identifies whether the pre-processed sentence is two consecutive sentences based on the position information of the pre-processed sentence, and the two consecutive sentences are connected by a but conjunction. In this case, the sentence weight of the pre-processed sentence may be decreased, and the sentence weight of the sentence after the but connected after the conjunction may be increased.

According to an embodiment, the news article analysis system 3000 reduces the sentence weight of the previous pre-processed sentence by 0.4 when two consecutive sentences are connected by the but conjunction, and The sentence weight of a sentence can be increased by 0.4. The news article analysis system 3000 may adjust the sentence score of the preprocessed sentence before the but conjunction by applying a sentence weight reduced by 0.4 to the sum of scores of words in the preprocessed sentence before the but conjunction. In addition, the news article analysis system 3000 may adjust the sentence score of the preprocessed sentence after the but conjunction by applying a sentence weight increased by 0.4 to the sum of the scores of the words in the preprocessed sentence after the but conjunction. .

In S720 , the news article analysis system 3000 may obtain a second news score by using the adjusted sentence score. For example, when the sentence scores are adjusted for each of the pre-processed sentences, the news article analysis system 3000 may acquire the second news score by adding the adjusted sentence scores.

8 is a flowchart illustrating a process in which a news article analysis system pre-processes a preset number of sentences according to an exemplary embodiment.

In S810 , the news article analysis system 3000 may identify each word in the sentence by tokenizing some extracted sentences among the identified sentences in the news article. Although not shown in FIG. 7 , the news article analysis system 3000 tokenizes some extracted sentences and then further performs a process of tokenizing words in some tokenized sentences to identify the words in the extracted sentences. may be According to an embodiment, the news article analysis system 3000 may tokenize the sentences by decomposing the extracted sentences into units that can no longer be divided grammatically. Also, the news article analysis system 3000 may tokenize the extracted sentences based on at least one of punctuation marks and spaces in the sentences, and may identify words from the tokenized sentences.

In S820 , the news article analysis system 3000 may remove words that are not used in calculating sentence scores from among the identified words. For example, the news article analysis system 3000 may generate a list of words that are not necessary for calculating the score of the sentence among the words identified by tokenizing the sentence, and when the words included in the generated word list are identified , you can remove those words.

In S830 , the news article analysis system 3000 may remove words not used in calculating sentence scores and convert the remaining words in each sentence into a headword form. For example, the news article analysis system 3000 removes words that are not used for calculating sentence scores, and when the remaining words are not in the form of a headword described in the dictionary (eg, looks), the words remaining in the form of a heading (eg, look) described in the dictionary can be converted The news article analysis system 3000 may determine a weight for each sentence using sentences including words converted into a headword form, and determine a sentence score for each sentence based on the determined weight.

9 is a diagram for describing sentence weight elements used by a news article analysis system to determine a sentence weight determined for each sentence according to an embodiment.

Referring to FIG. 9 , a sentence score element 904 used by the news article analysis system 3000 to determine a sentence weight for each sentence and element details 906 of each sentence score element are illustrated. The sentence score element used by the news article analysis system 3000 may be stored in advance in a memory or database in the electronic device or server according to a predetermined identification number 902 . The news article analysis system 3000 calculates a sentence weight for each pre-processed sentence based on at least one of negative phrases, adverbs, punctuation marks, emphasis phrases, negative words, and person words in the sentences pre-processed according to the pre-processing process of FIG. 8 . can be decided

According to an embodiment, the sentence weighting element 904 may include at least one of negative phrases, adverbs, punctuation marks, conjunctions, emphasis phrases, negative words, and person words. The news article analysis system 3000 may determine a sentence weight for each sentence based on a sentence weight element included in the pre-processed sentence. However, according to another embodiment, the sentence weight element of another sentence adjacent to the sentence including the sentence weight element may be changed. For example, when two consecutive sentences are connected by a but conjunction that is a sentence weighting element 804 (or the last sentence among two consecutive sentences is a sentence weighting element, but the conjunction included), the sentence weight of the preceding pre-processed sentence can be decreased by 0.4, and the sentence weight of the sentence after the but connected after the conjunction can be increased by 0.4.

According to an embodiment, among the sentence weight elements, negative phrases are element details 806 , and may include not and isn't. Also, according to an embodiment, among the sentence weighting factors, the adverb may include positive adverb and negative adverb, positive adverb may include absolutely or amazingly, and negative adverb may include at least one of scarcely or hardly. can However, the present invention is not limited thereto.

According to an embodiment, the punctuation mark may include an exclamation mark, a question mark, and a comma, and may further include other punctuation marks inserted at some positions in other sentences to emphasize or assist the meaning of the sentence.

According to an embodiment, the conjunction may include but, however, and the like, but is not limited thereto. In addition, according to an embodiment, the emphasis phrase may include very, etc., the negative word may include never, so, and without, and the person word may include the person's name shown in the person weight list shown in FIG. 3 . can

10 is a flowchart illustrating a method for a news article analysis system to determine a comprehensive evaluation score for a news article based on a first news score and a second news score according to an exemplary embodiment.

In S1020 , the news article analysis system 3000 may normalize the second news score output from the pre-learning model. For example, the distribution pattern of the score of the second news score that the news article analysis system 3000 can obtain from the pre-learning model is different from the distribution pattern of the score of the first news score that can be obtained from the neural network model. can Accordingly, the news article analysis system 3000 may make the distribution patterns of scores similar by normalizing the second news scores output from the pre-learning model according to the distribution pattern of the scores of the first news scores. According to an embodiment, the news article analysis system 3000 may normalize the second news score to a score between -1 and 1. According to an embodiment, S1020 may be performed by the electronic device 1000 in the news article analysis system 3000 .

In S1040 , the news article analysis system 3000 may determine an evaluation weight to be applied to the first news score and the second news score based on the neutral index value obtained from the prior learning model. For example, the news article analysis system 3000 may further acquire a neutral index from the pre-learning model, in addition to obtaining the second news score for the news article. According to another embodiment, in addition to obtaining the second news score, the news article analysis system 3000 may further obtain a neutral index, a negative index, and a positive index.

According to an embodiment, when the neutral index is less than or equal to the first threshold, the news article analysis system 3000 may determine the first evaluation weight to be small and the second evaluation weight to be large. According to another embodiment, when the neutral index is greater than or equal to the second threshold, the news article analysis system 3000 may determine the first evaluation weight to be large and the second evaluation weight to be small.

In S1060 , the news article analysis system 3000 may determine the overall evaluation score by weighting the first news score and the normalized second news score according to the determined evaluation weight. According to an embodiment, the news article analysis system 3000 determines different evaluation weights to be applied to the first news score and the second news score, based on the range of the neutral index value output from the pre-learning model, Differently determined weights may be applied to the first news score and the second news score. Referring to FIG. 10 to be described later, the electronic device 1000 determines an evaluation weight to be applied to the first news score and the second news score based on the neutral index value, and according to the determined evaluation weight, the first news score and A method of weighting the second news score will be described in more detail.

Hereinafter, the positive index, negative index, and neutral index values obtained by the news article analysis system 3000 from the pre-learning model will be described in detail. According to an embodiment, the neutral index, the negative index, and the positive index obtained by the news article analysis system 3000 from the pre-learning model are distribution patterns of word scores indicated by words included in each of some sentences extracted from the news article. can be determined according to According to an embodiment, the sum of the neutral index, the negative index, and the positive index may be 1, but is not limited thereto.

The positive index can be output large when words (positive finance words) containing a positive word score in the news article are distributed more than words containing a negative word score (negative finance words), The negative index may be output to be large when the words (negative finance words) including the negative word score in the news article are more than the words (positive finance words) including the positive word score. According to another embodiment, when the number of sentences having a positive sentence score in a news article is greater than the number of sentences having a negative sentence score, the positive index may be output to be large, and the negative index may be negative in the news article. When the number of sentences having a sentence score is greater than the number of sentences having a positive sentence score, the output may be large.

According to an embodiment, the neutral index may be a value representing the degree of confidence in the second news score output from the pre-learning model as a probability value. For example, when the neutral index is lower than the preset threshold, the second news score may be reliable with high probability, but when the neutral index is greater than the preset threshold, the confidence probability of the second news score may be low. For example, in the prior learning model used by the news article analysis system 3000, the distribution of sentence scores for each of some sentences (eg, sentences in summary data) extracted from a news article is a negative sentence score and a positive sentence score. If it is distributed so that it can be clearly classified as , the neutral index can be determined to be low. However, the prior learning model used by the news article analysis system 3000 is a neutral index when the distribution of sentence scores for each of some sentences extracted from a news article is not clearly divided into a negative sentence score and a positive sentence score. can be largely determined. According to an embodiment, when the neutral index is large, the reliability of the second news score output from the pre-learning model may be low, and when the neutral index is small, the reliability of the second news score output from the pre-learning model is low. can be high

According to an embodiment, the neutral index may be determined based on some sentence units extracted from the news article. According to an embodiment, the neutral index includes more words having a positive score in a sentence unit than words having a negative score, or more words having a negative score in a sentence unit than words having a positive score in a sentence unit. When the tendency of the sentence score in the sentence is clearly distinguished, such as in the case of inclusion, it may be output low. However, the neutral index may be output large when the number of words having a positive score and a word having a negative score in a sentence unit are similar, or when there is no sentence including a financial word or a person word preset in the corresponding sentence. .

11 is a flowchart illustrating a method for a news article analysis system to determine a comprehensive evaluation score based on a first news score and a second news score according to another exemplary embodiment.

In S1102 , the news article analysis system 3000 may identify a sentence including a pre-stored financial word or a person word in the news article. Since S1010 may correspond to S210 of FIG. 2 , a detailed description thereof will be omitted.

In S1104 , the news article analysis system 3000 may obtain a first news score from the neural network model by inputting some of the identified sentences into the neural network model trained in advance to output a score for the news article. Since S1012 may correspond to S220, a detailed description thereof will be omitted.

In S1106 , the news article analysis system 3000 may obtain a second news score and a neutral index from the prior learning model by inputting some of the identified sentences to the prior learning model. According to another embodiment, the news article analysis system 3000 obtains a second news score, a positive index, a negative index, and a neutral index from the pre-learning model by inputting some of the identified sentences into the pre-learning model. You may. In S1106, the operation of the news article analysis system 3000 to obtain the second news score using the pre-learning model may correspond to S712 to S720 of FIG. 7, and thus a detailed description thereof will be omitted.

In S1108 , the news article analysis system 3000 may identify whether the neutral index output from the pre-learning model is equal to or less than the first threshold. In S1110 , the news article analysis system 3000 may identify whether the neutral index output from the pre-learning model is less than the second threshold. In S1112 , the news article analysis system 3000 multiplies the neutral index value and a first evaluation weight smaller than 1 by the first news score when the neutral index output from the pre-learning model is identified as less than or equal to the first threshold value and a value obtained by subtracting the neutral index value from 1 and a value obtained by multiplying a normalized second news score by a value of a second evaluation weight greater than 1; A first evaluation weight*neutral index value*first news score+second evaluation weight*(1-neutral index value)*normalized second news score) may be determined.

In S1114, when the neutral index output from the pre-learning model is greater than the first threshold and less than the second threshold, the news article analysis system 3000 multiplies the neutral index value by the first news score, and the neutral index at 1 By summing the value obtained by subtracting the index value multiplied by the normalized second news score, the composite assessment score (eg, composite assessment score = neutral index value * first news score + (1-neutral index value) * normalized second score) 2 news score).

In S1116, when the neutral index output from the pre-learning model is greater than or equal to the second threshold, the news article analysis system 3000 multiplies the neutral index value and the first evaluation weight greater than 1 by the first news score and , by summing a value obtained by subtracting the neutral index value from 1 and a value obtained by multiplying a normalized second news score by a second evaluation weight less than 1, the overall evaluation score (eg, comprehensive evaluation score = first evaluation weight*neutral index) value*first news score+second evaluation weight*neutral index value*normalized second news score) may be determined.

12 is a block diagram of an electronic device according to an embodiment.

The electronic device 1000 according to an embodiment may include a processor 1400 and a memory 1402 . However, not all illustrated components are essential components, and the electronic device 1000 may be implemented with more components than the illustrated components, and the electronic device 1000 may be implemented with fewer components. may be According to an embodiment, the electronic device 1000 may further include a network interface (not shown) in addition to the processor 1400 and the memory 1402 .

The processor 1400 generally controls the overall operation of the electronic device 1000 . According to an embodiment, the processor 1400 according to the present disclosure may execute at least one function of the news article analysis system 3000 described in FIGS. 1 to 10 by executing programs stored in the memory 1402 . . In addition, the processor 1400 may be composed of one or a plurality of processors, and the one or the plurality of processors is a general-purpose processor such as a CPU, AP, DSP (Digital Signal Processor), etc., a graphics-only processor such as a GPU, or artificial intelligence (AI). ) may include a dedicated processor. However, the present invention is not limited thereto, and the processor 1400 may include other processing units that perform a function for analyzing news articles by executing instructions stored in the memory. According to an embodiment, when the processor 1400 includes a general-purpose processor, an artificial intelligence processor, and a graphics-only processor, the artificial intelligence processor may be implemented as a general-purpose processor or a chip separate from the graphics-only processor.

According to an embodiment, when the processor 1400 is implemented as a plurality of processors or graphics-only processors or artificial intelligence-only processors, at least some of the plurality of processors or graphics-only processors or artificial intelligence-only processors include the electronic device 1000 and It may be mounted on another electronic device connected to the electronic device 1000 or a server.

For example, the processor 1400 may acquire a news article from an external device connected to the electronic apparatus 1000 by executing programs stored in the memory 1402 . Also, the processor 1400 may transmit the obtained news article to a server connected to the electronic device.

Also, according to an embodiment, the processor 1400 determines a comprehensive evaluation score regarding the financial propensity of the news article based on the first news score and the second news score received from the server 2000 , and the determined comprehensive evaluation The score may be output through an output unit (not shown) of the electronic device 1000 .

According to another embodiment, the processor 1400 independently identifies a sentence including a financial word or a person word stored in advance in a news article, and assigns some of the identified sentences to a score for the news article. Obtaining a first news score from the neural network model by input to a neural network model that is trained to output, and obtaining a second news score from the prior learning model by inputting sentences of some of the identified sentences into a pre-learning model, , based on the first news score and the second news score, a comprehensive evaluation score regarding the financial propensity of the news article may be determined.

According to an embodiment, the processor 1400 transmits the modified and updated financial word list and the person weight list to the server when the financial word list including the financial word and the person weight list including the person word are modified and updated. may be However, according to another embodiment, the processor 1400 generates a financial word list by matching the financial word and the word score for each financial word, and a person weight list by matching the person word and the person word weight for each person word may be created, and the generated financial word list and the person weight list may be stored.

According to an embodiment, the processor 1400 identifies the position information of each of the sentences included in the news article, and based on the position information of the identified sentence, the first sentence in the news article and the last sentence in the news article It is possible to identify, and extract a preset number of sentences to be scored among the sentences including the financial word or the person word, the first sentence, and the last sentence.

However, according to another embodiment, the processor 1400 may transmit a news article to the server 2000 and receive information about a preset number of sentences extracted from the news article from the server.

According to an embodiment, the processor 1400 may pre-process the extracted predetermined number of sentences, input the pre-processed sentences into the neural network model, and obtain a first news score output from the neural network model. . However, according to another embodiment, the processor 1400 may receive the first news score output from the server 2000 using the neural network model stored in the server 2000 .

According to an embodiment, the processor 1400 pre-processes the extracted preset number of sentences, and determines the sentence score of each of the pre-processed sentences based on the sentence weight determined for each of the pre-processed sentences. and, based on the identified location information, adjust the sentence score determined for each of the pre-processed sentences, and obtain the second news score using the adjusted sentence score. However, according to an embodiment, the processor 1400 may receive the second news score output by the pre-learning model from the server 2000 . In addition, the processor 1400 may further receive information on the second news score, neutral index, positive index, or negative index output from the pre-learning model in the server 2000 .

According to an embodiment, the processor 1400 identifies each word in the sentence by tokenizing the extracted preset sentence, and among the identified words, a word not used in calculating the sentence score may be removed, words not used in calculating the sentence score may be removed, and words in each remaining sentence may be converted into a headword form.

The memory 1402 may store a program for processing and control of the processor 1400 , and may also store data input to or output from the electronic device 1000 . In addition, the memory 1402 may include a plurality of news article analysis models including a pre-trained model and a neural network model. According to an embodiment, the memory 1402 may store information about the layers constituting the neural network, the nodes included in the layers, and weights related to the connection strength of the layers as a configuration of the neural network model.

According to an embodiment, the memory 1402 may further store the financial word list described above with reference to FIG. 3 and the financial word list including word scores for each financial word. Also, the memory 1402 may further store a person weight list including preset person words and person word weights for each person word.

According to an embodiment, when the pre-learning model and the neural network model previously stored in the memory are modified and updated, the memory 1402 may store the modified and updated pre-learning model and the neural network model. The memory 1402 may include a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (eg, SD or XD memory), and a RAM. (RAM, Random Access Memory) SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, magnetic disk , may include at least one type of storage medium among optical disks. However, the present invention is not limited thereto, and may include instructions for performing a method for analyzing other news articles, and other storage media for storing information about at least one artificial intelligence model for analyzing the news article. have.

A network interface (not shown) may transmit data transmitted/received by the electronic device 1000 to/from an external device or server. For example, the electronic apparatus 1000 may obtain data about a news article from an external device through a network interface. Also, the electronic device 1000 may receive information about the neural network model or the pre-learning model from an external device through a network interface. Also, the electronic device 1000 may output the comprehensive evaluation score as an analysis result of the news article to an external device through the network interface.

13 is a block diagram of a server according to an embodiment.

According to an embodiment, the server 2000 may include a network interface 2100 , a database 2200 , and a processor 2300 .

The network interface 2100 may correspond to the above-described network interface (not shown) of the electronic device 1000 . For example, the network interface 2100 receives information on a neural network model or information on a pre-learning model from the electronic device 1000 , or information on a neural network model stored in a server by the electronic device 1000 or a pre-learning model You can also transmit information about In addition, the network interface 2100 may receive information about a news article from the electronic device, or transmit information on the first news score and the second news score determined by the server 2000 or a comprehensive evaluation score for the news article. .

For example, the network interface 2100 may receive information about layers and nodes included in the layers of the artificial neural network or weight values related to connection strength of layers in the neural network. The network interface 2100 may transmit, as information about the neural network model pre-trained by the server 2000 , to the electronic device 1000 , layers and weight values related to connection strengths of the layers.

The database 2200 may correspond to the memory 1402 of the electronic device shown in FIG. 12 . For example, the database 2200 may store a program for processing and controlling the processor 2300 , and may also store data input to or output from the electronic device 1000 . In addition, the database 2200 may further store information about the layers constituting the artificial neural network model, information about weights related to the nodes included in the layers, and weights related to the connection strength of the layers, and information about the pre-learning model. Also, the database 2200 may further store information on news articles received from an electronic device or an external device connected to the electronic device, a financial word list, and a person word weight list.

The processor 2300 may control the overall operation of the server 2000 . For example, the processor 2300 may perform an operation of analyzing a news article together with the electronic device 1000 by controlling the network interface 2100 and the database 2200 . According to an embodiment, the processor 2300 obtains a news article from the electronic device, identifies a sentence including a financial word or a person word stored in advance in the obtained news article, and selects a part of the identified sentence, A first news score may be obtained from the neural network model by inputting the score to the neural network model trained to output the score for the news article.

Also, according to an embodiment, the processor 2300 may obtain a second news score from the pre-learning model by inputting some of the identified sentences into the pre-learning model, and obtain a first news score and a second news score. may be transmitted to the electronic device.

According to an embodiment, the processor 2300 determines a comprehensive evaluation score regarding financial propensity for a news article received from the electronic device based on the first news score and the second news score, and adds the overall evaluation score to the determined overall evaluation score. information may be transmitted to the electronic device.

Also, according to an embodiment, the processor 2300 generates a financial word list by matching a financial word and a word score for each financial word, and generates a person weight list by matching a person word and a person word weight for each person word and the generated financial word list and the person weight list may be stored.

According to an embodiment, the processor 2300 identifies position information of each of the sentences included in the news article, and identifies the first sentence in the news article and the last sentence in the news article based on the position information of the identified sentence and a predetermined number of sentences to be scored can be extracted from among the financial word or the sentence including the person word, the first sentence, and the last sentence.

According to an embodiment, the processor 2300 may pre-process a preset number of extracted sentences, input the pre-processed sentences to the neural network model, and obtain a first news score output from the neural network model.

In addition, according to an embodiment, the processor 2300 pre-processes a preset number of extracted sentences, and based on a sentence weight determined for each of the pre-processed sentences, the sentence score of each of the pre-processed sentences is calculated. It is determined and based on the identified location information, the determined sentence score for each of the pre-processed sentences may be adjusted, and the second news score may be obtained using the adjusted sentence score.

According to an embodiment, the processor 2300 identifies each word in the sentence by tokenizing the extracted preset sentence, and removes words that are not used in calculating the sentence score from among the identified words. And, words not used in calculating the sentence score may be removed, and words in each remaining sentence may be converted into a headword form.

According to an embodiment, the processor 2300 is configured to calculate a sentence weight for each of the pre-processed sentences based on at least one of a negative phrase, an adverb, a punctuation mark, a emphasized phrase, a negative word, or the person word in the pre-processed sentence. , and applying the determined sentence weight to the word score of each financial word in the pre-processed sentence, the sentence score of each of the pre-processed sentences may be determined.

According to an embodiment, the processor 2300 identifies the contexts between the pre-processed sentences based on the location information of the pre-processed sentences, and based on the identified contexts, determines for each of the pre-processed sentences. You can adjust the sentence score.

According to an embodiment, the processor 2300 obtains, from the pre-learning model, a neutral index regarding the confidence level of the second news score that varies according to a distribution pattern of word scores of words included in some sentences of the identified sentences. can be obtained

According to an embodiment, when the neutral index value is greater than the first threshold and less than the second threshold, the processor 2300 multiplies the neutral index value by the first news score, and the neutral index value is 1 to 1. The overall evaluation score may be determined by adding a value obtained by subtracting the index value multiplied by the normalized second news score.

According to an embodiment, when the neutral index value is less than or equal to the first threshold, the processor 2300 multiplies the neutral index value and a first evaluation weight less than 1 by the first news score and , by adding a value obtained by subtracting the neutral index value from 1 and a value obtained by multiplying the second news score by a second evaluation weight greater than 1 to determine the overall evaluation score.

According to an embodiment, when the neutral index value is greater than or equal to the second threshold, the processor 2300 multiplies the neutral index value and a first evaluation weight greater than 1 by the first news score and , by adding a value obtained by subtracting the neutral index value from 1 and a value obtained by multiplying the second news score by a second evaluation weight less than 1, the overall evaluation score may be determined.

According to an embodiment, when a negative phrase is included in the pre-processed sentence, the processor 2300 determines a weight to be applied to the sentence including the negative phrase as a negative number, and includes a positive adverb in the pre-processed sentence. case, the weight of the sentence including the positive adverb may be increased, and when the negative adverb is included in the pre-processed sentence, the weight of the sentence including the negative adverb may be decreased.

According to an embodiment, when a punctuation mark for emphasis is included in the pre-processed sentence, the processor 2300 increases the weight of the sentence including the punctuation mark, and increases the weight of the sentence including the punctuation mark in the pre-processed sentence. When punctuation marks are included, the weight of the sentences including the punctuation marks may be reduced.

According to an embodiment, the processor 2300 increases the weight of the sentence including the emphasized phrase when an emphasis phrase is included in the pre-processed sentence, and when a negative word is included in the pre-processed sentence, the negative word is You can increase or decrease the weight of the included sentence.

14 is a flowchart illustrating a method of analyzing a news article by a news article analysis system according to another exemplary embodiment.

In S1402 , the electronic device 1000 may store in advance a neural network model and a pre-learning model for analyzing a news article. In S1404, the server 2000 may store a neural network model corresponding to the neural network model stored in the electronic device.

In S1406 , the server 2000 may obtain a news article from an external device according to a preset period. In S1408 , the server 2000 may generate learning data from the obtained news article. According to an embodiment, the server 2000 may generate training data from periodically acquired news articles and validation data for validating a neural network model trained according to the training data. The server 2000 may generate verification data by using a news article published in a period different from the news article used to generate the learning data.

In S1410, the server 2000 may train the neural network model based on the training data. For example, the server 2000 may modify and update weights in the neural network model based on the training data. According to an embodiment, the server 2000 is configured such that the output value of the neural network model output based on the training data decreases the difference between the output values output from the neural network model when the verification data is input to the neural network model. The weights of the model can be modified and updated.

In S1412 , the server 2000 may transmit information on the weight of the modified and updated neural network model to the electronic device 1000 . In S1414 , the electronic device 1000 may update the neural network model previously stored in the electronic device based on the weight information received from the server 2000 . For example, the electronic device 1000 may convert the weights of the neural network model stored in advance in the electronic device based on information about the weights received from the server 2000 .

In S1416 , the electronic apparatus 1000 may obtain a new news article from an external device. In S1418 , the electronic device 1000 may identify a sentence including a financial word or a person word stored in advance in the news article. In S1420, the electronic device 1000 may obtain a first news score by inputting some of the identified sentences to the neural network model. Since S1416 to S1420 may correspond to S206 to S210 of FIG. 2 , a detailed description thereof will be omitted.

In S1422 , the electronic device 1000 may acquire a second news score by inputting some of the identified sentences to the pre-learning model. In S1424, the electronic device 1000 may determine a comprehensive evaluation score based on the first news score and the second news score. In S1426, the electronic device 1000 may output the determined overall evaluation score. Since S1422 to S1426 may correspond to S1106 to S1116 of FIG. 11 , a detailed description thereof will be omitted.

The above-described method according to an embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present disclosure, or may be known and available to those skilled in the art of computer software.

In addition, according to the embodiment, a computer program apparatus including a recording medium storing a program for performing another method may be provided. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

Although the embodiments of the present disclosure have been described in detail above, the scope of the present disclosure is not limited thereto, and various modifications and improved forms of the present disclosure are also provided by those skilled in the art using the basic concept of the present disclosure as defined in the following claims. belong to the scope of the right

Claims (15)

A method for a news article analysis system to analyze a news article, the method comprising:
obtaining, by an electronic device in the news article analysis system, a news article from at least one external device connected to the electronic device;
transmitting, by the electronic device, the obtained news article to a server connected to the electronic device;
identifying, by the server, a sentence including a pre-stored financial word or a person word in the news article;
obtaining, by the server, a first news score from the neural network model by inputting, by the server, some sentences of the identified sentences into a neural network model that is trained to output a score for the news article;
obtaining, by the server, a second news score from the pre-learning model by inputting some of the identified sentences into the pre-learning model;
transmitting, by the server, a first news score and a second news score to the electronic device;
determining, by the electronic device, a comprehensive evaluation score regarding the financial propensity of the news article based on the first news score and the second news score; and
outputting the determined overall evaluation score; A method comprising delete The method of claim 1, wherein the method
generating, by the server, a financial word list by matching the financial word and the word score for each financial word;
generating, by the server, a person weight list by matching a person word with a person word weight for each person word; and
storing, by the server, the generated financial word list and the person weight list; A method further comprising: The method of claim 1, wherein the method
identifying, by the server, location information of each of the sentences included in the news article;
identifying, by the server, a first sentence in a news article and a last sentence in the news article based on the location information of the identified sentence; and
extracting, by the server, a preset number of sentences to be score calculation targets from sentences, the first sentences, and the last sentences including the financial word or the person word; A method further comprising: 5. The method of claim 4, wherein obtaining the first news score comprises:
pre-processing, by the server, the extracted preset number of sentences;
inputting, by the server, the pre-processed sentence to the neural network model; and
obtaining, by the server, a first news score output from the neural network model; A method comprising 5. The method of claim 4, wherein obtaining the second news score comprises:
pre-processing, by the server, the extracted preset number of sentences;
determining, by the server, a sentence score of each of the pre-processed sentences based on a sentence weight determined for each of the pre-processed sentences;
adjusting, by the server, a sentence score determined for each of the pre-processed sentences based on the identified location information; and
obtaining, by the server, the second news score using the adjusted sentence score; A method comprising The method of claim 6, wherein the pre-processing of the preset number of sentences comprises:
identifying, by the server, each word in the sentence by tokenizing the extracted preset sentence;
removing, by the server, words not used in calculating sentence scores from among the identified words; and
removing, by the server, words not used in calculating the sentence score and converting the remaining words in each sentence into a headword form; A method comprising The method of claim 6, wherein obtaining the second news score comprises:
determining, by the server, a sentence weight for each of the pre-processed sentences based on at least one of a negative phrase, an adverb, a punctuation mark, a emphasized phrase, a negative word, or the person word in the pre-processed sentence; and
determining, by the server, the sentence score of each of the pre-processed sentences by applying the determined sentence weight to the word scores of each financial word in the pre-processed sentence; A method comprising The method of claim 6, wherein adjusting the sentence score determined for each of the pre-processed sentences comprises:
identifying, by the server, a context between the pre-processed sentences based on the location information of the pre-processed sentences; and
adjusting, by the server, a sentence score determined for each of the pre-processed sentences based on the identified context; A method comprising The method of claim 1, wherein the method
obtaining, by the server, a neutral index with respect to the confidence level of the second news score, which varies according to a distribution pattern of word scores of words included in some sentences of the identified sentences, from the pre-learning model; further comprising,
The step of determining the overall evaluation score is
normalizing, by the electronic device, a second news score obtained from the server;
determining, by the electronic device, an evaluation weight to be applied to the first news score and the second news score, based on the value of the neutral index; and
determining, by the electronic device, the comprehensive evaluation score by weighting the first news score and the normalized second news score according to the determined evaluation weight; A method comprising The method of claim 10, wherein the determining of the overall evaluation score comprises:
When the neutral index value is greater than the first threshold and less than the second threshold, the electronic device calculates a value obtained by multiplying the neutral index value by a first news score and subtracting the neutral index value from 1 as the normalized second value. determining the overall evaluation score by summing a value multiplied by 2 news scores; A method comprising The method of claim 10, wherein the determining of the overall evaluation score comprises:
When the neutral index value is less than or equal to the first threshold, a value obtained by multiplying the first news score by the electronic device by the neutral index value and a first evaluation weight less than 1, and a value obtained by subtracting the neutral index value from 1 and determining the overall evaluation score by summing a value obtained by multiplying the second news score by a second evaluation weight greater than 1; A method comprising The method of claim 10, wherein the determining of the overall evaluation score comprises:
When the neutral index value is greater than or equal to the second threshold, a value obtained by multiplying the first news score by the electronic device by the neutral index value and a first evaluation weight greater than 1, and a value obtained by subtracting the neutral index value from 1 and determining the overall evaluation score by summing a value obtained by multiplying a second news score by a second evaluation weight less than 1; A method comprising The method of claim 8, wherein determining the sentence weight for each of the pre-processed sentences comprises:
determining, by the server, a negative weight to be applied to the sentence including the negative phrase when the negative phrase is included in the pre-processed sentence;
When a positive adverb is included in the pre-processed sentence, the server increases the weight of the sentence including the positive adverb, and when a negative adverb is included in the pre-processed sentence, the sentence including the negative adverb reducing the weight;
When the server includes a punctuation mark for emphasis in the pre-processed sentence, the weight of the sentence including the punctuation mark is increased, and a punctuation mark such as a question mark or a comma is included in the pre-processed sentence, reducing the weight of the sentence including the punctuation mark;
increasing, by the server, a weight of the sentence including the emphasized phrase when the emphasized phrase is included in the pre-processed sentence; and
increasing or decreasing, by the server, a weight of the sentence including the negative word when the negative word is included in the pre-processed sentence; A method comprising In the news article analysis system for analyzing news articles,
First news from the neural network model by identifying a sentence including a financial word or a person word stored in advance in a news article, and inputting a sentence of some of the identified sentences into a neural network model that is trained to output a score for the news article a server for obtaining a score and obtaining a second news score from the pre-learning model by inputting some of the identified sentences into the pre-learning model; and
obtaining a news article from an external device, transmitting the obtained news article to the server, obtaining the first news score and the second news score from the server, and the obtained first news score and the second news score an electronic device for determining a comprehensive evaluation score on the financial propensity of the news article based on the , and outputting the determined comprehensive evaluation score; Including, news article analysis system.

Images