Determination Method, Device, Electronic Device and Storage Medium for First Release Press Release

Abstract

The present application discloses a method, apparatus, electronic device, and storage medium for determining a first-release press release. The method for determining the first-release press release includes: obtaining first news element information of a news event; performing a press release search based on the first news element information to obtain a first press release candidate set, and obtaining second news element information of each candidate press release in the first press release candidate set; determining a feature vector based on the first news element information and the second news element information; inputting the feature vector into a pre-trained classifier to obtain a second press release candidate set that matches the news event; and determining a first-release press release corresponding to the news event based on the release times of the candidate press releases in the second press release candidate set. By using the method for determining the first-release press release provided by the present application, it is possible to achieve the effect of efficiently and accurately determining the first-release press release reporting a certain event.

Description

Technical Field

[0001] The present application relates to the field of news technology, and in particular to a method, device, electronic device and storage medium for determining a first release of a news release. Background Art

[0002] In news media organizations, the timeliness of news reports is very important. The first press release (i.e., the first news report released on the same news event) will also have a greater influence. How to determine the first press release to report a certain event is an important issue facing news media organizations today.

[0003] At present, the identification of the first press release is usually carried out by conventional search + manual sorting, that is, the press release that needs to be analyzed for the first identification is searched through the search engine according to its search term to obtain the search results, and then the release time of the press releases identified as the same event in the search results is manually compared and sorted, so as to determine the first press release reporting the event.

[0004] The above method for determining the first release of a news release is prone to errors due to manual comparison, the determination of the first release of a news release is not accurate enough, and the manual comparison is inefficient. Summary of the invention

[0005] The purpose of the embodiments of the present application is to provide a method, device, electronic device and storage medium for determining the first press release, so as to achieve the effect of efficiently and accurately determining the first press release reporting a certain event.

[0006] The technical solution of this application is as follows:

[0007] In a first aspect, a method for determining a first press release is provided, the method comprising:

[0008] Obtaining the first news element information of a news event;

[0009] Searching for news releases according to the first news element information to obtain a first news release candidate set, wherein the first news release candidate set includes at least one candidate news release;

[0010] Acquire the second news element information of each candidate news release in the first news release candidate set;

[0011] Determining a feature vector based on the first news element information and the second news element information;

[0012] Inputting the feature vector into a pre-trained classifier to obtain a second press release candidate set matching the news event, wherein the classifier is trained using the feature vectors corresponding to historical press releases and the news events corresponding to the historical press releases as training samples;

[0013] Based on the release time of each candidate press release in the second candidate press release set, determine the first released press release corresponding to the news event.

[0014] In a second aspect, a device for determining the first released press release is provided. The device includes:

[0015] A first news element information acquisition module, configured to acquire first news element information of a news event;

[0016] A first press release candidate set determination module, configured to perform a press release search according to the first news element information to obtain a first press release candidate set, where at least one candidate press release is included in the first press release candidate set;

[0017] A second news element information acquisition module, configured to acquire second news element information of each candidate press release in the first press release candidate set;

[0018] A feature vector determination module, configured to determine a feature vector based on the first news element information and the second news element information;

[0019] A second press release candidate set determination module, configured to input the feature vector into a pre-trained classifier to obtain a second press release candidate set that matches the news event, where the classifier is trained with the feature vectors corresponding to historical press releases and the news events corresponding to the historical press releases as training samples;

[0020] A first released press release determination module, configured to determine the first released press release corresponding to the news event based on the release time of each candidate press release in the second press release candidate set.

[0021] In a third aspect, an embodiment of the present application provides an electronic device, which includes a processor, a memory, and a program or instruction stored on the memory and executable on the processor. When the program or instruction is executed by the processor, the steps of the method for determining the first released press release according to the first aspect of the embodiment of the present application are implemented.

[0022] In a fourth aspect, an embodiment of the present application provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the method for determining the first released press release according to the first aspect of the embodiment of the present application are implemented.

[0023] The technical solutions provided by the embodiments of the present application at least bring the following beneficial effects:

[0024] The method for determining the first release press release provided by the embodiments of the present application obtains the first news element information of a news event, searches for press releases based on the first news element information to obtain a first set of press release candidates, obtains the second news element information of each candidate press release in the first set of press release candidates, determines a feature vector based on the first news element information and the second news element information, inputs the feature vector into a pre-trained classifier to obtain a second set of press release candidates that match the news event, and determines the first release press release corresponding to the news event based on the release times of the candidate press releases in the second set of press release candidates. In this way, by using the classifier to screen the candidate press releases in the first set of press release candidates, a second set of press release candidates can be obtained, which can quickly and accurately obtain the second set of press release candidates, and then quickly and accurately determine the first release press release corresponding to the news event based on the release times of the candidate press releases in the second set of press release candidates, without manual screening, improving the efficiency and accuracy of determining the first release press release.

[0025] It should be understood that the above general description and subsequent detailed description are only exemplary and explanatory, and do not limit the present application. Brief Description of the Drawings

[0026] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application, and do not constitute an improper limitation of the present application.

[0027] Figure 1 is one of the flowcharts of the method for determining the first release press release shown according to an exemplary embodiment;

[0028] Figure 2 is the second of the flowcharts of the method for determining the first release press release shown according to an exemplary embodiment;

[0029] Figure 3 is the structural block diagram of the device for determining the first release press release shown according to an exemplary embodiment;

[0030] Figure 4 is the structural block diagram of an electronic device provided by the embodiments of the present application. Detailed Description of the Embodiments

[0031] In order to enable those of ordinary skill in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without some of these specific details. The following description of the embodiments is only intended to provide a better understanding of the present application by showing examples of the present application.

[0032] It should be noted that the terms "first", "second", etc. in the description, claims and the above-mentioned drawings of this application are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of this application described here can be implemented in an order other than those illustrated or described here. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. On the contrary, they are only examples consistent with some aspects of this application as detailed in the appended claims.

[0033] As described in the background art section, in the prior art, there is a problem that the determination of the first news release is not accurate enough and the manual comparison efficiency is low because the manual comparison and collation of the release times of news articles determined to be the same event in the search results are used to determine the first news release of the event. To solve the above problems, the embodiments of this application provide a method for determining the first news release. This method obtains the first news element information of the news event, searches for news releases according to the first news element information to obtain a first news release candidate set, obtains the second news element information of each candidate news release in the first news release candidate set, determines a feature vector based on the first news element information and the second news element information, inputs the feature vector into a pre-trained classifier to obtain a second news release candidate set that matches the news event, and determines the first news release corresponding to the news event based on the release times of the candidate news releases in the second news release candidate set. In this way, by using the classifier to screen each candidate news release in the first news release candidate set, a second news release candidate set is obtained, so that the second news release candidate set can be obtained quickly and accurately, and then the first news release corresponding to the news event can be quickly and accurately determined based on the release times of the candidate news releases in the second news release candidate set, without manual screening, improving the efficiency and accuracy of the determination of the first news release.

[0034] It should be noted that the method for determining the first news release provided by the embodiments of this application can be applied to various news releases, such as Chinese news releases, English news releases, financial news releases, entertainment news releases, etc.

[0035] The following combines the drawings to detail the method for determining the first news release provided by the embodiments of this application through specific embodiments and their application scenarios.

[0036] Figure 1 It is a schematic flowchart of a method for determining the first news release provided by the embodiments of this application. The execution subject of this method for determining the first news release can be a server. It should be noted that the above execution subject does not limit this application.

[0037] As Figure 1 shown, the method for determining the first news release provided by the embodiments of the present application may include steps 110 to 160.

[0038] Step 110, obtain the first news element information of the news event.

[0039] Step 120, perform a news release search based on the first news element information to obtain a first news release candidate set.

[0040] Among them, the first news release candidate set may include at least one candidate news release.

[0041] Step 130, obtain the second news element information of each candidate news release in the first news release candidate set.

[0042] Step 140, determine a feature vector based on the first news element information and the second news element information.

[0043] Step 150, input the feature vector into a pre-trained classifier to obtain a second news release candidate set that matches the news event.

[0044] Among them, the classifier can be trained with the feature vectors corresponding to historical news releases and the news events corresponding to the historical news releases as training samples.

[0045] Step 160, determine the first news release corresponding to the news event based on the release times of the candidate news releases in the second news release candidate set.

[0046] In the embodiments of the present application, by obtaining the first news element information of the news event, performing a news release search based on the first news element information to obtain a first news release candidate set, obtaining the second news element information of each candidate news release in the first news release candidate set, determining a feature vector based on the first news element information and the second news element information, inputting the feature vector into a pre-trained classifier to obtain a second news release candidate set that matches the news event, and determining the first news release corresponding to the news event based on the release times of the candidate news releases in the second news release candidate set. In this way, by using the classifier to screen the candidate news releases in the first news release candidate set to obtain the second news release candidate set, the second news release candidate set can be obtained quickly and accurately. Furthermore, based on the release times of the candidate news releases in the second news release candidate set, the first news release corresponding to the news event can be determined quickly and accurately, without manual screening, improving the efficiency and accuracy of determining the first news release.

[0047] The method for determining the first news release provided by the embodiments of the present application will be described in detail below.

[0048] First, introduce step 110, which is to obtain the first news element information of a news event.

[0049] Among them, the news event can be a news that describes a certain event.

[0050] In one example, the news event can be an event that describes the opening of a zoo in a certain area.

[0051] The first news element information can be the news element information of the news event.

[0052] Then introduce step 120, which is to search for news articles based on the first news element information to obtain a first news article candidate set.

[0053] Among them, the first news article candidate set can be a set composed of multiple candidate news articles corresponding to the first news element information obtained after searching for news articles based on the first news element information.

[0054] In some embodiments of the present application, the candidate news article can be a news article corresponding to the first news element information.

[0055] In some embodiments of the present application, searching for news articles based on the first news element information to obtain the first news article candidate set can be achieved through a search engine, such as an ES search engine, etc.

[0056] In one example, after obtaining the first news element information of the news event, the first news element information can be used as a search term to search for multiple news articles in the database through a search engine to obtain multiple candidate news articles corresponding to the first news element information, forming the first news article candidate set.

[0057] In some embodiments of the present application, for each candidate news article in the first news article candidate set, it can be sorted based on the number of first news element information it contains, that is, the more the number of first news element information it contains, the higher its ranking in the first news article candidate set.

[0058] In some embodiments of the present application, in order to quickly obtain the first news article candidate set, before step 120, the above-mentioned method for determining the first news article can further include:

[0059] Perform word segmentation processing on each news article to be stored to obtain the word segmentation corresponding to each news article;

[0060] Correspondingly store the word segmentation and the news article corresponding to the word segmentation in the database.

[0061] In an embodiment of the present application, by performing word segmentation on each press release to be stored, the word segmentation corresponding to each press release is obtained, and the word segmentation and the press release corresponding to the word segmentation are stored in the database correspondingly. In this way, when using the first news element information as a search term for searching, the first news element information can be compared with the word segmentation, and the press release corresponding to the word segmentation with consistent comparison is used as a candidate press release. In this way, the first press release candidate set can be quickly obtained, and the determination efficiency of the first release press release is improved.

[0062] Next, step 140 is introduced. Based on the first news element information and the second news element information, a feature vector is determined.

[0063] Among them, the feature vector can be a vector used to characterize the similarity between the first news element information and the second news element information.

[0064] In some embodiments of the present application, in order to accurately determine the feature vector, step 140 may include:

[0065] The first news element information and the second news element information are respectively expanded to obtain the first expanded news element information and the second expanded news element information;

[0066] Determine the semantic similarity value between the first press release and each candidate press release in the first press release candidate set;

[0067] Based on the first news element information, the second news element information, the first expanded news element information, the second expanded news element information, and the semantic similarity value, determine the feature vector.

[0068] Among them, the first expanded news element information may be information obtained by expanding the first news element information.

[0069] The second expanded news element information may be information obtained by expanding the second news element information.

[0070] The semantic similarity value may be the value of the semantic similarity between the first press release and each candidate press release in the first press release candidate set.

[0071] In some embodiments of the present application, for any one of the first news element information and the second news element information, when expanding it, verbs and specific noun phrases related to it may be added, etc.

[0072] In one example, the content described in a press release is as follows: During the May Day holiday in 2021 in Beijing, the total number of tourists received was 8.426 million person-times, a year-on-year increase of 81.9% compared to 2020, and restored to 98.4% of that in 2019 calculated according to comparable caliber; the total tourism revenue was 9.3 billion yuan, a year-on-year increase of 1.2 times compared to 2020, and restored to 86% of that in 2019 calculated according to comparable caliber. The news element information of this press release can be: Beijing, 2021, 2020, 8.426 million, 2020, 81.9%, 2019, 98.4%, 9.3 billion, 2020, 1.2 times, and 86%. The news element information can be extended. For example, the "increase" before 81.9%, the "increase" before 1.2 times, the noun phrase "May Day holiday", etc. can also be added to the news element information to form extended news element information.

[0073] In some embodiments of the present application, the semantic similarity value can be obtained by performing semantic analysis on the first press release and each candidate press release in the first candidate press release set based on a semantic similarity algorithm, or by inputting the first press release and each candidate press release in the first candidate press release set into a pre-trained semantic analysis model. Specifically, how to obtain the semantic similarity value between the first press release and each candidate press release in the first candidate press release set is not limited herein, as long as the method can calculate the semantic similarity value between the first press release and each candidate press release in the first candidate press release set falls within the protection scope of the present application.

[0074] In some embodiments of the present application, after obtaining the first news element information, the second news element information, the first extended news element information, the second extended news element information, and the semantic similarity value, an accurate feature vector can be obtained based on the first news element information, the second news element information, the first extended news element information, the second extended news element information, and the semantic similarity value.

[0075] In the embodiments of the present application, by respectively expanding the first news element information and the second news element information to obtain the first extended news element information and the second extended news element information, then determining the semantic similarity value between the first press release and each candidate press release in the first candidate press release set, and determining the feature vector based on the first news element information, the second news element information, the first extended news element information, the second extended news element information, and the semantic similarity value, an accurate feature vector can be obtained in this way, and then an accurate first press release corresponding to the news event can be obtained.

[0076] In some embodiments of the present application, in order to further accurately determine the feature vector, determining the feature vector corresponding to the candidate press release based on the first news element information, the second news element information, the first extended news element information, the second extended news element information, and the semantic similarity value may specifically include:

[0077] Calculate respectively the first quantity of the first news element information included in the title of the first press release and the second quantity of the first extended news element information included therein, as well as the third quantity of the first news element information included in the body of the first press release and the fourth quantity of the first extended news element information included therein;

[0078] Calculate respectively the fifth quantity of the second news element information included in the title of each candidate press release and the sixth quantity of the second extended news element information included therein, as well as the seventh quantity of the second news element information included in the body of each candidate press release and the eighth quantity of the second extended news element information included therein;

[0079] Calculate respectively the difference between the first quantity and the fifth quantity to obtain the first difference value, the difference between the second quantity and the sixth quantity to obtain the second difference value, the difference between the third quantity and the seventh quantity to obtain the third difference value, and the difference between the fourth quantity and the eighth quantity to obtain the fourth difference value;

[0080] Calculate respectively the first maximum value of the fifth quantity, the second maximum value of the sixth quantity, the third maximum value of the seventh quantity, and the fourth maximum value of the eighth quantity corresponding to each candidate press release;

[0081] Determine the feature vector based on the first difference value, the second difference value, the third difference value, the fourth difference value, the first maximum value, the second maximum value, the third maximum value, the fourth maximum value, and the semantic similarity value.

[0082] Wherein, the first quantity may be the quantity of the first news element information included in the title of the first press release.

[0083] The second quantity may be the quantity of the first extended news element information included in the title of the first press release.

[0084] The third quantity may be the quantity of the first news element information included in the body of the first press release.

[0085] The fourth quantity may be the quantity of the first extended news element information included in the body of the first press release.

[0086] The fifth quantity may be the quantity of the second news element information included in the title of the candidate press release.

[0087] The sixth quantity may be the quantity of the second extended news element information included in the title of the candidate press release.

[0088] The seventh quantity may be the quantity of the second news element information included in the body of the candidate press release.

[0089] The eighth quantity may be the quantity of the second extended news element information included in the body of the candidate press release.

[0090] The first difference may be the difference between the first quantity and the fifth quantity.

[0091] In one example, the news event is the news of the opening of Zoo B in Beijing. The content of the title of the first press release is: In October 2021, Zoo B in Beijing opened during the National Day. The first news element information is these 4 elements: October 2021, Beijing, Zoo B, and National Day. Then the quantity of the first news element information included in the title of the first press release is 4. The content of the title of a certain candidate press release is: In October 2021, Zoo A in Tianjin opened during the National Day holiday. Then the quantity of the first news element information included in the title of this candidate press release is 2 (i.e., October 2021 and National Day), and the first difference is 2.

[0092] The second difference may be the difference between the second quantity and the sixth quantity.

[0093] The third difference may be the difference between the third quantity and the seventh quantity.

[0094] The fourth difference may be the difference between the fourth quantity and the eighth quantity.

[0095] The first maximum value may be the maximum value of the fifth quantity corresponding to each candidate press release, that is, the maximum value of the quantity of the second news element information included in the title of each candidate press release.

[0096] In one example, there are 3 candidate press releases: the first candidate press release, the second candidate press release, and the third candidate press release. Among them, the quantity of the second news element information included in the title of the first candidate press release is 2, the quantity of the second news element information included in the title of the second candidate press release is 3, and the quantity of the second news element information included in the title of the third candidate press release is 4. Then the first maximum value is 4.

[0097] The second maximum value may be the maximum value of the sixth quantity corresponding to each candidate press release, that is, the maximum value of the quantity of the second news element information included in the body of each candidate press release.

[0098] The third maximum value may be the maximum value of the seventh quantity corresponding to each candidate press release, that is, the maximum value of the quantity of the second extended news element information included in the title of each candidate press release.

[0099] The fourth maximum value may be the maximum value of the eighth quantity corresponding to each candidate press release, that is, the maximum value of the quantity of the second extended news element information included in the text of each candidate press release.

[0100] In some embodiments of the present application, after obtaining the first difference, the second difference, the third difference, the fourth difference, the first maximum value, the second maximum value, the third maximum value, the fourth maximum value, and the semantic similarity value, when determining the feature vector, it may be to form a matrix with the first difference, the second difference, the third difference, the fourth difference, the first maximum value, the second maximum value, the third maximum value, the fourth maximum value, and the semantic similarity value to obtain the feature vector.

[0101] In some embodiments of the present application, since for the same difference, the proportion it occupies in different data is different, and considering the similarity between the candidate press release and the news event in the embodiments of the present application, if only the difference is used to determine the feature vector, it may cause the problem of inaccurate feature vectors.

[0102] In an example, taking the first difference as an example for illustration, the proportion of the first difference in the first quantity and the fifth quantity is taken as the average of the proportion of the first difference in the first quantity and the proportion of the first difference in the fifth quantity for illustration. If the first quantity is 10 and the second quantity is 9, then the first difference is 1, and the average of the proportion of the first difference in the first quantity and the proportion of the first difference in the fifth quantity is 0.105; if the first quantity is 3 and the second quantity is 2, then the first difference is also 1, and the average of the proportion of the first difference in the first quantity and the proportion of the first difference in the fifth quantity is approximately 0.416. It can be seen from this that although the two first differences are both 1, the proportions they occupy are different.

[0103] In some embodiments of the present application, based on the analysis of the above example, when determining the feature vector, in order to avoid the problem of inaccurate feature vectors caused by using the first difference (the second difference, the third difference, or the fourth difference) as a component of the feature vector, the first difference can be replaced by the proportion of the first difference in the first quantity and the fifth quantity, and the corresponding second difference can also be replaced by the proportion of the second difference in the second quantity and the sixth quantity. The third difference and the fourth difference are similar and will not be elaborated here.

[0104] In some embodiments of the present application, taking the first difference as an example for illustration, the proportion of the first difference in the first quantity and the fifth quantity may be the average of the proportion of the first difference in the first quantity and the proportion of the first difference in the fifth quantity, or may be the minimum value of the proportion of the first difference in the first quantity and the proportion of the first difference in the fifth quantity, or may also be the maximum value of the proportion of the first difference in the first quantity and the proportion of the first difference in the fifth quantity. This is not limited herein and can be set according to user requirements.

[0105] In the embodiments of the present application, by separately calculating the first quantity of the first news element information included in the title of the first press release and the second quantity of the first extended news element information included therein, as well as the third quantity of the first news element information included in the body of the first press release and the fourth quantity of the first extended news element information included therein; separately calculating the fifth quantity of the second news element information included in the title of the candidate press release and the sixth quantity of the second extended news element information included therein, as well as the seventh quantity of the second news element information included in the body of the candidate press release and the eighth quantity of the second extended news element information included therein; separately calculating the difference between the first quantity and the fifth quantity to obtain a first difference, the difference between the second quantity and the sixth quantity to obtain a second difference, the difference between the third quantity and the seventh quantity to obtain a third difference, and the difference between the fourth quantity and the eighth quantity to obtain a fourth difference; separately calculating the first maximum value of the fifth quantity, the second maximum value of the sixth quantity, the third maximum value of the seventh quantity, and the fourth maximum value of the eighth quantity corresponding to each candidate press release; determining a feature vector based on the first difference, the second difference, the third difference, the fourth difference, the first maximum value, the second maximum value, the third maximum value, the fourth maximum value, and the semantic similarity value. In this way, an accurate feature vector value can be obtained, and based on this accurate feature vector, an accurate first released press release can be obtained.

[0106] Then, step 150 is introduced: inputting the feature vector into a pre-trained classifier to obtain a second candidate set of press releases that match the news event.

[0107] Among them, the classifier can be trained using the feature vectors corresponding to historical press releases and the news events corresponding to the historical press releases as training samples.

[0108] The second candidate set of press releases can be a set of candidate press releases obtained by inputting the feature vector into a pre-trained classifier and screening each candidate press release in the first candidate set of press releases.

[0109] In some embodiments of the present application, the candidate press releases in the second candidate set of press releases may be some or all of the candidate press releases in the first candidate set of press releases.

[0110] In the embodiments of the present application, by inputting the obtained feature vectors into a pre-trained classifier, the classifier can screen each candidate press release in the first candidate press release set based on the received feature vectors, determine the candidate press releases that are closer to the news event from them, form a set, and obtain the second candidate press release set.

[0111] In some embodiments of the present application, based on the received feature vectors, the classifier may calculate the feature vectors. When the calculation result is greater than or equal to a certain threshold, it is determined that the candidate press release corresponding to the feature vector is closer to the news event, and it is screened out as the candidate press release in the second candidate press release set. How the specific classifier obtains the second candidate press release set based on the feature vectors is not limited herein and can be set according to user needs.

[0112] In the embodiments of the present application, by inputting the obtained feature vectors into a pre-trained classifier, the classifier can screen each candidate press release in the first candidate press release set based on the received feature vectors, determine the candidate press releases that are closer to the news event from them, form a set, and obtain the second candidate press release set. In this way, the effect of quickly determining the second candidate press release set is achieved, without the need for manual collation and confirmation, improving the determination efficiency of the second candidate press release set, and further improving the determination efficiency of the first released press release corresponding to the news event.

[0113] Finally, step 160 is introduced. Based on the release times of the candidate press releases in the second candidate press release set, the first released press release corresponding to the news event is determined.

[0114] In some embodiments of the present application, after the second candidate press release set is determined, if the number of candidate press releases included in the second candidate press release set is small at this time, the first released press release corresponding to the news event can be determined based on the release times of the candidate press releases in the second candidate press release set. Specifically, the press release with the earliest release time among the candidate press releases in the second candidate press release set can be determined as the first released press release corresponding to the news event.

[0115] In some embodiments of the present application, obtaining the release times of the candidate press releases in the second candidate press release set can be achieved through a target text detection algorithm. For example, after crawling the complete text, the release time fields in the candidate press releases in the second candidate press release set can be obtained using the target text detection algorithm, and based on the release time fields, the release times of the candidate press releases in the second candidate press release set can be obtained. In this way, the release times of the candidate press releases in the second candidate press release set can be quickly obtained, improving the acquisition efficiency of the release times of the candidate press releases in the second candidate press release set, and further improving the acquisition efficiency of the first released press release corresponding to the news event.

[0116] In some embodiments of the present application, in order to obtain accurate first news element information, such as Figure 2 As shown, another implementable manner of the method for determining the first news release is further provided in the embodiments of the present application, which may specifically include Step 210 to Step 280.

[0117] Step 210: Obtain a first news release reporting a news event from a database storing multiple news releases.

[0118] Wherein, the first news release may be a news release reporting the above-mentioned news event.

[0119] Step 220: Extract third news element information from the first news release, and determine the third news element information as the first news element information.

[0120] The third news element information may be the news element information of the first news release.

[0121] Step 230: Obtain the first news element information of the news event.

[0122] Step 240: Perform a news release search according to the first news element information to obtain a first news release candidate set.

[0123] Wherein, the first news release candidate set may include at least one candidate news release.

[0124] Step 250: Obtain the second news element information of each candidate news release in the first news release candidate set.

[0125] Step 260: Determine a feature vector based on the first news element information and the second news element information.

[0126] Step 270: Input the feature vector into a pre-trained classifier to obtain a second news release candidate set that matches the news event.

[0127] Step 280: Determine the first news release corresponding to the news event based on the release time of each candidate news release in the second news release candidate set.

[0128] In some embodiments of the present application, Step 230 to Step 270 are the same as Step 110 to Step 150 in the above embodiments, and will not be elaborated here.

[0129] In the embodiments of the present application, by obtaining a first news release reporting a news event from a database storing multiple news releases and extracting third news element information from the first news release, the third news element information can be determined as the first news element information. In this way, more first news element information can be obtained from the first news release, and thus accurate first news element information can be obtained.

[0130] It should be noted that the application scenarios described in the embodiments of the present application above are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. As is known to those of ordinary skill in the art, with the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0131] It should be noted that for the method for determining the first press release provided by the embodiments of the present application, the execution subject may be a device for determining the first press release, or a control module in the device for determining the first press release that is used to execute the method for determining the first press release. In the embodiments of the present application, taking the device for determining the first press release executing the method for determining the first press release as an example, the device for determining the first press release provided by the embodiments of the present application is described.

[0132] Based on the same inventive concept as the above method for determining the first press release, the present application also provides a device for determining the first press release. The following Figure 3 will describe in detail the device for determining the first press release provided by the embodiments of the present application.

[0133] Figure 3 is a structural block diagram of a device for determining the first press release shown according to an exemplary embodiment.

[0134] As Figure 3 shown, the device 300 for determining the first press release may include:

[0135] A first news element information acquisition module 310, configured to acquire first news element information of a news event;

[0136] A first press release candidate set determination module 320, configured to perform a press release search according to the first news element information to obtain a first press release candidate set, where the first press release candidate set includes at least one candidate press release;

[0137] A second news element information acquisition module 330, configured to acquire second news element information of each of the candidate press releases in the first press release candidate set;

[0138] A feature vector determination module 340, configured to determine a feature vector based on the first news element information and the second news element information;

[0139] A second press release candidate set determination module 350, configured to input the feature vector into a pre-trained classifier to obtain a second press release candidate set that matches the news event, where the classifier is trained with the feature vectors corresponding to historical press releases and the news events corresponding to the historical press releases as training samples;

[0140] A first release press release determination module 360 for determining a first release press release corresponding to the news event based on the release times of the candidate press releases in the second candidate press release set.

[0141] In an embodiment of the present application, based on the first news element information acquisition module, the first news element information of the news event is acquired. Based on the first press release candidate set determination module, a press release search is performed according to the first news element information to obtain a first press release candidate set. Based on the second news element information acquisition module, the second news element information of each candidate press release in the first press release candidate set is acquired. Based on the feature vector determination module, a feature vector is determined based on the first news element information and the second news element information. Based on the second press release candidate set determination module, the feature vector is input into a pre-trained classifier to obtain a second press release candidate set that matches the news event. Based on the first release press release determination module, the first release press release corresponding to the news event is determined based on the release times of the candidate press releases in the second press release candidate set. In this way, by using the classifier to screen the candidate press releases in the first press release candidate set, a second press release candidate set is obtained, which can quickly and accurately obtain the second press release candidate set, and then quickly and accurately determine the first release press release corresponding to the news event according to the release times of the candidate press releases in the second press release candidate set, without manual screening, improving the determination efficiency and accuracy of the first release press release.

[0142] In some embodiments of the present application, in order to obtain accurate first news element information, the above-mentioned first release press release determination device may further include:

[0143] A first press release acquisition module for acquiring a first press release reporting the news event from a database storing a plurality of press releases;

[0144] A first news element information determination module for extracting third news element information from the first press release and determining the third news element information as the first news element information.

[0145] In some embodiments of the present application, in order to accurately determine the feature vector, the feature vector determination module 340 may specifically include:

[0146] An extended news element information determination unit for respectively expanding the first news element information and the second news element information to obtain first extended news element information and second extended news element information;

[0147] A semantic similarity value determination unit for determining the semantic similarity values between the first press release and the candidate press releases in the first news candidate set;

[0148] A feature vector determination unit, configured to determine a feature vector based on the first news element information, the second news element information, the first extended news element information, the second extended news element information, and the semantic similarity value.

[0149] In some embodiments of the present application, in order to more accurately determine the feature vector, the feature vector determination unit may specifically be configured to:

[0150] Calculate respectively the first quantity of the first news element information included in the title of the first news release and the second quantity of the first extended news element information included therein, as well as the third quantity of the first news element information included in the body of the first news release and the fourth quantity of the first extended news element information included therein;

[0151] Calculate respectively the fifth quantity of the second news element information included in the title of each candidate news release and the sixth quantity of the second extended news element information included therein, as well as the seventh quantity of the second news element information included in the body of each candidate news release and the eighth quantity of the second extended news element information included therein;

[0152] Calculate respectively the difference between the first quantity and the fifth quantity to obtain a first difference value, the difference between the second quantity and the sixth quantity to obtain a second difference value, the difference between the third quantity and the seventh quantity to obtain a third difference value, and the difference between the fourth quantity and the eighth quantity to obtain a fourth difference value;

[0153] Calculate respectively the first maximum value of the fifth quantity, the second maximum value of the sixth quantity, the third maximum value of the seventh quantity, and the fourth maximum value of the eighth quantity corresponding to each candidate news release;

[0154] Determine the feature vector based on the first difference value, the second difference value, the third difference value, the fourth difference value, the first maximum value, the second maximum value, the third maximum value, the fourth maximum value, and the semantic similarity value.

[0155] The device for determining the first news release provided by the embodiments of the present application can be used to execute the method for determining the first news release provided by the above-mentioned method embodiments. The implementation principle and technical effects are similar. For the sake of brevity, they will not be elaborated here.

[0156] Based on the same inventive concept, the embodiments of the present application also provide an electronic device.

[0157] Figure 4 is a schematic structural diagram of an electronic device provided by the embodiments of the present application. As Figure 4 shown, the electronic device may include a processor 401 and a memory 402 storing computer programs or instructions.

[0158] Specifically, the above-mentioned processor 401 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured as one or more integrated circuits for implementing the embodiments of the present invention.

[0159] The memory 402 may include a mass storage for data or instructions. By way of example and not limitation, the memory 402 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disc, a magneto-optical disc, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In a suitable case, the memory 402 may include a removable or non-removable (or fixed) medium. In a suitable case, the memory 402 may be internal or external to the integrated gateway disaster recovery device. In a specific embodiment, the memory 402 is a non-volatile solid state memory. In a specific embodiment, the memory 402 includes a read only memory (ROM). In a suitable case, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or a flash memory or a combination of two or more of these.

[0160] The processor 401 reads and executes the computer program instructions stored in the memory 402 to implement any one of the methods for determining the first press release in the above embodiments.

[0161] In one example, the electronic device may further include a communication interface 403 and a bus 410. As shown, Figure 4 the processor 401, the memory 402, and the communication interface 403 are connected through the bus 410 to complete communication with each other.

[0162] The communication interface 403 is mainly used to implement communication between the various modules, devices, units, and / or devices in the embodiments of the present invention.

[0163] The bus 410 includes hardware, software, or both, and couples components of the electronic device to each other. By way of example and not limitation, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a MicroChannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable bus or a combination of two or more of these. Where appropriate, the bus 410 may include one or more buses. Although embodiments of the present invention describe and illustrate specific buses, the present invention contemplates any suitable bus or interconnect.

[0164] The electronic device can execute the method for determining the first press release in the embodiments of the present invention, so as to implement Figure 1 - Figure 2 any of the described methods for determining the first press release.

[0165] In addition, in combination with the method for determining the first press release in the above embodiments, embodiments of the present invention can provide a readable storage medium to implement. Program instructions are stored on the readable storage medium; when the program instructions are executed by a processor, any of the methods for determining the first press release in the above embodiments is implemented.

[0166] It should be clear that the present invention is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and illustrated as examples. However, the method process of the present invention is not limited to the specific steps described and illustrated, and those skilled in the art can make various changes, modifications, and additions, or change the order between steps after understanding the spirit of the present invention.

[0167] The functional blocks shown in the above-described structural block diagrams can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a functional card, and so on. When implemented in software, the elements of the present invention are programs or code segments for performing the required tasks. The program or code segment can be stored in a machine-readable medium or transmitted via a data signal carried in a carrier wave over a transmission medium or a communication link. A "machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROMs, flash memories, erasable ROMs (EROMs), floppy disks, CD-ROMs, optical discs, hard disks, fiber optic media, radio frequency (RF) links, and so on. The code segment can be downloaded via a computer network such as the Internet, an intranet, and so on.

[0168] It should also be noted that in the exemplary embodiments of the present invention, some methods or systems are described based on a series of steps or devices. However, the present invention is not limited to the order of the above steps, that is, the steps can be executed in the order mentioned in the embodiments, can be different from the order in the embodiments, or several steps can be executed simultaneously.

[0169] As described above, the above is only the specific implementation manner of the present invention. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, modules, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. A method for determining a first release press release, characterized in that, The method includes: Obtaining first news element information of a news event; Performing a news release search based on the first news element information to obtain a first news release candidate set, where at least one candidate news release is included in the first news release candidate set; Obtaining second news element information of each candidate news release in the first news release candidate set; Determining a feature vector based on the first news element information and the second news element information; Inputting the feature vector into a pre-trained classifier to screen each candidate news release in the first news release candidate set, and obtaining a second news release candidate set that matches the news event, where the classifier is trained with the feature vectors corresponding to historical news releases and the news events corresponding to the historical news releases as training samples; Determining a first released news release corresponding to the news event based on the release times of the candidate news releases in the second news release candidate set; The determining the feature vector based on the first news element information and the second news element information includes: Respectively expanding the first news element information and the second news element information according to the first news release and each candidate news release in the first news release candidate set to obtain first expanded news element information and second expanded news element information, where the first news release is a news release reporting the news event obtained from a database storing multiple news releases; Determining the semantic similarity values between the first news release and each candidate news release in the first news release candidate set; Determining a feature vector based on the first news element information, the second news element information, the first expanded news element information, the second expanded news element information, and the semantic similarity values.

2. The method according to claim 1, wherein Before obtaining the first news element information of the news event, the method further includes: Obtaining a first news release reporting the news event from a database storing multiple news releases; Extracting third news element information from the first news release and determining the third news element information as the first news element information.

3. The method according to claim 1, wherein The determining the feature vector based on the first news element information, the second news element information, the first expanded news element information, the second expanded news element information, and the semantic similarity values includes: Respectively calculating a first quantity of the first news element information included in the title of the first news release and a second quantity of the first expanded news element information included therein, and a third quantity of the first news element information included in the text of the first news release and a fourth quantity of the first expanded news element information included therein; Respectively calculating a fifth quantity of the second news element information included in the title of each candidate news release and a sixth quantity of the second expanded news element information included therein, and a seventh quantity of the second news element information included in the text of each candidate news release and an eighth quantity of the second expanded news element information included therein; Calculate the difference between the first quantity and the fifth quantity respectively to obtain a first difference, calculate the difference between the second quantity and the sixth quantity to obtain a second difference, calculate the difference between the third quantity and the seventh quantity to obtain a third difference, and calculate the difference between the fourth quantity and the eighth quantity to obtain a fourth difference; Calculate the first maximum value of the fifth quantity, the second maximum value of the sixth quantity, the third maximum value of the seventh quantity, and the fourth maximum value of the eighth quantity corresponding to each of the candidate press releases respectively; Determine a feature vector based on the first difference, the second difference, the third difference, the fourth difference, the first maximum value, the second maximum value, the third maximum value, the fourth maximum value, and the semantic similarity value.

4. A device for determining a first release press release, characterized in that The device includes: A first news element information acquisition module, configured to acquire first news element information of a news event; A first press release candidate set determination module, configured to perform a press release search according to the first news element information to obtain a first press release candidate set, where the first press release candidate set includes at least one candidate press release; A second news element information acquisition module, configured to acquire second news element information of each of the candidate press releases in the first press release candidate set; A feature vector determination module, configured to determine a feature vector based on the first news element information and the second news element information; A second press release candidate set determination module, configured to input the feature vector into a pre-trained classifier to screen each candidate press release in the first press release candidate set to obtain a second press release candidate set that matches the news event, where the classifier is trained with the feature vectors corresponding to historical press releases and the news events corresponding to the historical press releases as training samples; A first released press release determination module, configured to determine a first released press release corresponding to the news event based on the release times of the candidate press releases in the second press release candidate set; The feature vector determination module is specifically configured to: Respectively expand the first news element information and the second news element information according to the first press release and each of the candidate press releases in the first press release candidate set to obtain first expanded news element information and second expanded news element information, where the first press release is a press release reporting the news event obtained from a database storing a plurality of press releases; Determine the semantic similarity value between the first press release and each of the candidate press releases in the first press release candidate set; Determine a feature vector based on the first news element information, the second news element information, the first expanded news element information, the second expanded news element information, and the semantic similarity value.

5. An electronic device, characterized in that, Includes a processor, a memory, and a program or instruction stored on the memory and running on the processor, and when the program or instruction is executed by the processor, the steps of the method for determining the first released press release as described in any one of claims 1-3 are implemented.

6. A readable storage medium, characterized in that, A program or instruction is stored on the readable storage medium, and when the program or instruction is executed by a processor, the steps of the method for determining the first released press release as described in any one of claims 1-3 are implemented.

Images