Information processing method based on news event and related device

Abstract

The embodiment of the application discloses a news event-based information processing method and related device, the method comprises the following steps: collecting news events pushed by multiple news sources, and extracting feature information in the news events; according to the feature information, excluding repeated news events from different news sources to obtain multiple target news events; tagging the multiple target news events; determining a risk score of each target news event in the multiple target news events; calculating and determining a push priority of each target news event in the multiple target news events; according to the push priority, determining a push order of the multiple target news events, and sending the multiple target news events carrying information of the push order to a terminal device. In this way, the event risk analysis of news events and the dynamic adjustment of news push order according to business needs effectively reduce the load of the server, reduce resource consumption, reduce news push delay, and improve information processing efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data processing, and in particular to an information processing method based on news events and related devices. BACKGROUND

[0002] In the financial market, the analysis of market sentiment and trends of news events is crucial for investment decisions and risk management.

[0003] The current financial news has the following defects: there are a large number of news information sources, in the past, the development context of a certain event was mostly sorted out through manual reading combined with automatic program pushing mode, there was a lack of analysis of the relationship between events and financial risks, there was a large amount of repeated information, the traditional processing method did not effectively remove the repeated information, which led to the need for servers to store and calculate a large amount of redundant data; the traditional news pushing mechanism made the terminal frequently request updates, the interaction times between the server and the terminal were too high on a daily basis, which led to a high network bandwidth occupation; in terms of news presentation, it was often simply sorted by time priority, but in scenarios with business needs, there was a lack of a priority sorting scheme that adapted to the needs. SUMMARY

[0004] Therefore, the embodiments of the present application provide an information processing method based on news events and related devices, so as to analyze the event risk of news events and the like through AI and big model data processing technology, and dynamically adjust the news pushing order according to business needs, effectively reduce the load of the server, reduce resource consumption, reduce news pushing delay, and improve information processing efficiency.

[0005] In a first aspect, the embodiments of the present application provide an information processing method based on news events, applied to a server in a news pushing system, the news pushing system comprising: the server, and a terminal device of a user; the method comprising:

[0006] Collecting news events pushed by multiple news sources, extracting feature information in the news events, the feature information comprising at least one of the following: a person, a place, a time, and an event summary;

[0007] According to the feature information, excluding repeated news events from different news sources to obtain a plurality of target news events;

[0008] Selecting a risk label within a risk label range to label the plurality of target news events, the risk label range being a label range pre-determined by an operation and maintenance personnel of the news pushing system, the risk label being a set of keywords related to financial risks, and the risk label representing the influence degree of the target news event on financial business;

[0009] determine a risk score of each target news event in the plurality of target news events according to the risk label;

[0010] determine a push priority of each target news event in the plurality of target news events according to the risk score, a timeliness index and a topic index, the push priority being used to reflect a degree of importance of the target news event, the timeliness index being used to reflect a degree of timeliness of occurrence of the target news event, and the topic index being used to reflect a degree of spread of the target news event;

[0011] determine a push order of the plurality of target news events according to the push priority, and send the plurality of target news events carrying information of the push order to the terminal device, so as to instruct the terminal device to display the plurality of target news events according to the push order.

[0012] In one possible implementation, the excluding duplicate news events from different news sources according to the feature information comprises: converting text information of a plurality of the feature information into semantic vectors by a pre-trained deep learning model based on a natural language processing algorithm, the semantic vector being a vector with fixed dimensions and used to represent deep semantic meanings of the feature information; comparing cosine similarities of any two vectors in the semantic vectors; if the cosine similarity of the two vectors is greater than or equal to a first similarity threshold, determining that the feature information corresponding to the two vectors is consistent, and the consistent feature information representing that the associated news events are duplicate news events; and determining times of a plurality of the duplicate news events according to the feature information, and retaining a news event with the earliest time.

[0013] In one possible implementation, the determining the risk score of each target news event in the plurality of target news events according to the risk label comprises: determining a type and a number of the risk label in each target news event, the each target news event including a plurality of the risk label, the type of the risk label being classified into a first-level risk, a second-level risk, a third-level risk and a fourth-level risk from low to high according to a degree of influence on the financial business; and calculating the risk score of the each target news event according to the type and the number of the risk label, the risk score being in a positive correlation with the degree of influence of the risk label on the financial business and in a positive correlation with the number of the risk label.

[0014] In a possible embodiment, before the calculating the push priority of each target news event in the plurality of target news events according to the risk score, the timeliness indicator, and the topic indicator, the method further includes: determining a publishing time of the each target news event; calculating a publishing time difference of the each target news event according to the publishing time and a current time; determining a level of the timeliness indicator according to the publishing time difference, wherein the level of the timeliness indicator includes a first timeliness, a second timeliness, a third timeliness, and a fourth timeliness from high to low in terms of timeliness of the occurrence of the target news event; wherein the publishing time difference is less than or equal to a first time length, and the timeliness indicator is the first timeliness; the publishing time difference is greater than the first time length and less than or equal to a second time length, and the timeliness indicator is the second timeliness; the publishing time difference is greater than the second time length and less than or equal to a third time length, and the timeliness indicator is the third timeliness; the publishing time difference is greater than the third time length, and the timeliness indicator is the fourth timeliness; determining a number of news sources reporting the each target news event within a preset time window; and determining a total number of the plurality of news sources; and calculating the topic indicator according to the number of news sources reporting the each target news event and the total number of the plurality of news sources.

[0015] In a possible embodiment, the calculating the push priority of each target news event in the plurality of target news events includes: performing normalization processing on the risk score, the timeliness indicator, and the topic indicator; configuring a weight parameter for the risk score, the timeliness indicator, and the topic indicator after the normalization processing, wherein the weight parameter corresponding to the risk score is greater than the weight parameter corresponding to the timeliness indicator, and the weight parameter corresponding to the timeliness indicator is greater than the weight parameter corresponding to the topic indicator; and calculating the push priority according to the risk score, the timeliness indicator, and the topic indicator after the normalization processing, and the weight parameter.

[0016] In a possible embodiment, after the cosine similarity of any two vectors in the semantic vectors is compared, the method further comprises: if the cosine similarity of the two vectors is less than the first similarity threshold and greater than or equal to a second similarity threshold, determining that the news events corresponding to the two vectors are related news events of the same event; counting a plurality of related news events and establishing a news event list respectively, the news event list including the plurality of related news events of the same event theme and the target news event; calculating a push priority of the plurality of related news events; if the push priority of at least one event in the plurality of related news events is higher than the push priority of the target news event, replacing the target news event with the event in the news event list; and sending the news event list and prompt information to the terminal device, the prompt information being used to prompt the user to perform a pull instruction at the terminal device to browse the plurality of related news events included in the news event list.

[0017] In a possible embodiment, the method further comprises: determining the time of the plurality of target news events; for the plurality of target news events occurring on the same day, recalculating the push priority of the plurality of target news events according to a first preset period; and for the plurality of target news events occurring before the same day, recalculating the push priority of the plurality of target news events according to a second preset period, wherein the length of the first preset period is less than the length of the second preset period.

[0018] In a second aspect, an embodiment of the present application provides a news event-based information processing apparatus applied to a server in a news pushing system, the news pushing system comprising the server and a terminal device of a user, and the apparatus comprising a collecting unit, a screening unit, a selecting unit, a determining unit, a calculating unit and a pushing unit, wherein the collecting unit is specifically configured to collect news events pushed by multiple news sources, and extract feature information in the news events, the feature information comprising at least one of the following: a person, a place, a time, an event subject and an event summary; the screening unit is specifically configured to exclude duplicate news events from different news sources according to the feature information, and obtain multiple target news events; the selecting unit is specifically configured to select risk labels within a risk label range to mark the multiple target news events, the risk label range being a label range pre-determined by an operation and maintenance personnel of the news pushing system, the risk label being a keyword set related to financial risks, and the risk label representing an influence degree of the target news events on financial businesses; the determining unit is specifically configured to determine a risk score of each target news event in the multiple target news events according to the risk labels; the calculating unit is specifically configured to calculate a pushing priority of each target news event in the multiple target news events according to the risk score, a timeliness index and a topic index, the pushing priority being used to reflect an importance degree of the target news event, the timeliness index being used to reflect a timeliness of occurrence of the target news event, and the topic index being used to reflect a propagation degree of the target news event; and the pushing unit is specifically configured to determine a pushing order of the multiple target news events according to the pushing priority, and send the multiple target news events carrying the pushing order to the terminal device, so as to instruct the terminal device to display the multiple target news events according to the pushing order.

[0019] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor, a memory, a communication interface and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs comprise instructions for performing steps in any method of the first aspect of the embodiments of the present application.

[0020] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program for electronic data exchange, and the computer program causes a computer to perform some or all of the steps described in any method of the first aspect of the embodiments of the present application.

[0021] In a fifth aspect, an embodiment of the present application provides a computer program product, wherein the computer program product includes a non-transitory computer readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute part or all of the steps in any method of the first aspect of the embodiments of the present application. The computer program product can be a software installation package.

[0022] It can be seen that, by the information processing method based on news events and the related apparatus provided in the present application, first, a plurality of news events pushed by a plurality of news sources are collected, and feature information in the news events is extracted; second, according to the feature information, duplicate news events from different news sources are excluded, and a plurality of target news events are obtained; third, risk labels are selected within a risk label range to label the plurality of target news events; fourth, according to the risk labels, risk scores of each target news event in the plurality of target news events are determined; fifth, according to the risk scores, timeliness indexes and topic indexes, push priorities of each target news event in the plurality of target news events are calculated and determined; and finally, according to the push priorities, a push order of the plurality of target news events is determined, and the plurality of target news events carrying information of the push order are sent to the terminal device. In this way, the news events and the like are subjected to event risk analysis, and the news push order is dynamically adjusted according to business requirements, thereby effectively reducing the load of the server, reducing resource consumption, reducing news push delay, and improving information processing efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0024] Figure 1 is an architecture schematic diagram of a news push system provided by an embodiment of the present application;

[0025] Figure 2 is a functional unit composition block diagram of a server provided by an embodiment of the present application;

[0026] Figure 3 is a flow schematic diagram of an information processing method based on news events provided by an embodiment of the present application;

[0027] Figure 4 is a news display interface of a terminal device provided by an embodiment of the present application;

[0028] Figure 5is a flow diagram of another information processing method based on news events provided by an embodiment of the present application;

[0029] Figure 6 is a function unit composition block diagram of an information processing device based on news events provided by an embodiment of the present application;

[0030] Figure 7 is a structure block diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0031] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative labor fall within the scope of protection of the present application.

[0032] The terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish different objects, and are not used to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but can optionally include steps or units not listed, or can optionally include other steps or units inherent to the process, method, product, or device.

[0033] It should be understood that the term "and / or" herein is only used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone. In addition, the character " / " in this paper represents that the front and rear associated objects are a "or" relationship.

[0034] "Multiple" appearing in the embodiments of the present application means two or more. "Connection" appearing in the embodiments of the present application means direct connection or indirect connection and various connection modes to realize communication between devices, which is not limited by the embodiments of the present application.

[0035] Reference to "embodiments" in this document means that the specific features, structures, or characteristics described in conjunction with the embodiments can be included in at least one embodiment of the present application. The phrase appears at various places in the specification does not necessarily mean the same embodiment, nor is it an independent or alternative embodiment to other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0036] The related content, concepts, meanings, technical problems, technical solutions, beneficial effects and the like involved in the embodiments of the present application are described below.

[0037] NLP (Natural Language Processing, natural language processing): a branch of artificial intelligence and computational linguistics, aiming to enable computers to understand and process human language, NLP covers a wide range of tasks from text analysis to text generation, and its goal is to enable computers to understand text meaning like humans.

[0038] In the financial market, the analysis of market sentiment and trends of news events is crucial for investment decisions and risk management.

[0039] The current financial news has the following defects: there are many sources of news information, in the past, to sort out the development context of a certain event, mostly through manual reading combined with automatic program pushing method, lacking of analysis of the relationship between events and financial risks, there are a lot of repeated information, the traditional processing method has not effectively removed the duplication, resulting in the server needs to store and calculate a large amount of redundant data; the traditional news pushing mechanism makes the terminal frequently request update, the interaction times between the server and the terminal are too high per day, resulting in high network bandwidth occupation; in the news presentation, it is often simply sorted by time priority, but in the scene with business needs, there is a lack of a priority sorting scheme that adapts to the needs.

[0040] To solve the above problems, the embodiments of the present application provide a news event-based information processing method and related device, in order to analyze the event risk of news events and the like through AI and large model data processing technology, and dynamically adjust the news pushing order according to business needs, effectively reduce the load of the server, reduce resource consumption, at the same time reduce the news pushing delay, improve the information processing efficiency.

[0041] First, the method in the embodiments of the present application is applied to the server in the news pushing system, combined with Figure 1 The news event-based information processing method in the embodiments of the present application is described. Figure 1 is a schematic diagram of the architecture of a news pushing system provided by the embodiments of the present application, the news pushing system 100 includes: a server 110, and a terminal device 120 of a user. Wherein, the server 110 and the terminal device 120 are in communication connection.

[0042] First, server 110 collects news events pushed by multiple news sources and extracts feature information from these events. This feature information includes at least one of the following: people, location, time, event overview, and the viewpoint / attitude of the news text content. Second, based on this feature information, server 110 eliminates duplicate news events from different news sources, resulting in multiple target news events. Next, server 110 selects risk tags within a risk tag range to label the multiple target news events. The risk tag range is predetermined by the news push system's maintenance personnel. Risk tags refer to a set of keywords related to financial risk, and these risk tags characterize the degree of impact of the target news events on financial business. Then, server 110... Based on risk labels, a risk score is determined for each of the multiple target news events. Then, server 110 calculates and determines the push priority of each target news event based on the risk score, timeliness index, and topic index. The push priority reflects the importance of the target news event, the timeliness index reflects the immediacy of the event's occurrence, and the topic index reflects the dissemination of the event. Finally, server 110 determines the push order of the multiple target news events based on the push priority and sends multiple target news events carrying push order information to terminal device 120, instructing terminal device 120 to display the multiple target news events according to the push order.

[0043] Specifically, terminal device 120 can be a computer (desktop / laptop) with professional event risk monitoring software installed, supporting large-screen display of multi-dimensional information (such as news event lists, risk tags, event evolution paths, dynamic change curves of scores, etc.); mobile terminals, such as smartphones and tablets, can achieve lightweight access through APPs or mini-programs, allowing users to receive high-risk news pushes anytime, anywhere. The terminal device connects to server 110 via a network interface, receiving structured data pushed by server 110 in real time (such as news text, risk scores, tags, similarity clustering results, etc.); it can also provide feedback on user operation data, such as filtering conditions and tag click counts, to assist the system in optimizing score weights.

[0044] Specifically, Server 110 refers to a computer cluster with high-performance computing, large-capacity storage, and network communication capabilities. It is used to deploy AI models (such as SentenceTransformer and large language models), run data processing algorithms, store news data and intermediate results, and respond to real-time data requests. Server 110 connects to multiple news channels (such as financial websites, social media, and listed company announcement platforms) via network interfaces to receive massive amounts of news data in real time. Server 110 is also the operating platform for core technologies such as the SentenceTransformer algorithm, large language models, and dynamic scoring formulas. Server 110 can also monitor newly pushed news data in real time and trigger a dynamic update mechanism based on preset rules. Please refer to [link / reference]. Figure 2 , Figure 2 This is a block diagram of the functional units of a server provided in an embodiment of this application, such as... Figure 2 As shown, server 110 includes a news preprocessing module 1101, an AI annotation module 1102, a scoring module 1103, and a sorting module 1104. The news preprocessing module 1101 receives real-time news and then extracts key entity information (such as news figures, event details, time, etc.) from the real-time news through the AI ​​model deployed on server 110. Subsequently, it filters out duplicate news and generates a news event list through a news classification engine, that is, it treats the same set of events in real-time news as a news event list. Then, the AI ​​annotation module 1102 annotates the news with risk tags. Then, the scoring module 1103 calculates and determines the importance score (i.e., push priority) of the news by comprehensively considering risk indicators, topic popularity, and timeliness factors. Finally, the sorting module 1104 sorts the news according to the importance score for easy display on the front end (i.e., terminal device 120).

[0045] The following is combined with Figure 3 This application describes an information processing method based on news events in its embodiments. Figure 3 This is a flowchart illustrating an information processing method based on news events provided in an embodiment of this application. The method is applied to... Figure 1 , Figure 2 The server 110 shown includes the following steps:

[0046] Step S310: Collect news events pushed by multiple news sources and extract feature information from the news events.

[0047] The feature information includes at least one of the following: people, places, times, event overview, and the viewpoints and attitudes expressed in the news text.

[0048] Specifically, information from news sources such as financial news websites and social media is crawled in real time through servers, and feature information is extracted through large models.

[0049] Step S320, according to the feature information, excluding the repeated news events from different news sources, obtaining a plurality of target news events.

[0050] Specifically, in one possible embodiment, according to the feature information, excluding the repeated news events from different news sources, including: based on a natural language processing algorithm, converting the text information of the plurality of feature information into semantic vectors by a pre-trained deep learning model, the semantic vector refers to a vector with fixed dimensions for representing the deep semantic meaning of the feature information; comparing the cosine similarity of any two vectors in the semantic vector; if the cosine similarity of the two vectors is greater than or equal to the first similarity threshold, it is determined that the feature information corresponding to the two vectors is consistent, and the consistent feature information represents that the associated news event is a repeated news event; determining the time in the plurality of repeated news events according to the feature information, and retaining the earliest news event.

[0051] Specifically, the NLP algorithm (pre-trained model is paraphrase-MiniLM-L6-v2) of SentenceTransformer is used to convert the news text into a fixed-dimensional semantic vector. The server calculates the cosine similarity of the semantic vectors of news X and news Y, including: calculating the cosine value of the included angle of the two vectors, the formula is as follows:

[0052] Cosine similarity = (vector X x vector Y) / (||vector X|| x ||vector Y||);

[0053] Wherein, the numerator is the vector dot product, the denominator is the vector modulus product, the result range is [0, 1], and the value is closer to 1, indicating that the semantics are more similar; after calculation, the cosine similarity of news X and news Y is 94%. The system presets the first similarity threshold W1 = 91%, and since 94% ≥ 91%, it is determined that the two news are repeated news events.

[0054] Then, the publishing time of the two news X and Y is extracted: news X is 09:00, and news Y is 09:05, so the earliest news X is retained as the target news event; the server sends an instruction to the terminal device to delete the display permission of news Y, and only retains news X in the news event list of the terminal, avoiding the interference of repeated information to the user.

[0055] It can be seen that in the embodiment, through the natural language processing algorithm, the text information of the plurality of feature information is converted into semantic vectors by a pre-trained deep learning model; the cosine similarity of any two vectors in the semantic vectors is compared; if the cosine similarity of the two vectors is greater than or equal to the first similarity threshold, it is determined that the feature information corresponding to the two vectors is consistent; and the time in the plurality of repeated news events is determined according to the feature information, and the news event with the earliest time is retained. In this way, through the comparison of the semantic vectors, the repeated news with different texts but the same reflected content can be recognized, the noise information is effectively reduced, and the efficiency of focusing on important information is improved.

[0056] Step S330, selecting a risk label in the risk label range to label the plurality of target news events.

[0057] The risk label range is a label range pre-determined by an operation and maintenance personnel of the news pushing system, the risk label is a set of keywords related to financial risk, and the risk label represents the influence degree of the target news event on the financial business.

[0058] Specifically, part of the risk label is given by the operation and maintenance team according to the business scene demand and experience; another part, the risk label range can also be jointly formulated by the large model and the operation and maintenance personnel, the target news event is subjected to keyword extraction by the large model, and after mass data analysis and induction, the operation and maintenance personnel audits and determines; the large model is used to label in the specified risk label range during labeling, and secondary verification is performed to ensure accuracy. The operation and maintenance personnel here refers to a labeling expert in the field of financial risk control.

[0059] Specifically, when the target news event is displayed at the terminal device in subsequent pushing, the risk label is displayed below the title of the target news event, so that the user can understand the monitoring risk points involved in the news and used for classification and screening at the first time.

[0060] Step S340, determining a risk score of each target news event in the plurality of target news events according to the risk label.

[0061] Specifically, one target news event often has a plurality of hierarchical labels, such as 1-4 level labels; the risk level and the number of labels are comprehensively considered as two weights, so that the news with high-level risk labels is ranked more forwardly, and the risk score is determined by the following formula:

[0062] α = 1 / 2 x log(c4 + 1) + 3 x log(c3 + 1) + log(c2 + 1) + 0.5 x log(c1 + 1);

[0063] Wherein, ci is the number of i-level risk labels.

[0064] For example, if a news contains one secondary label (c2=1), then a=log(1+1)=0.6931; if a news contains two quaternary labels, then the risk score a=13.1833.

[0065] In step S350, the push priority of each target news event in the plurality of target news events is calculated and determined according to the risk score, the timeliness index, and the topic index.

[0066] The push priority is used to reflect the importance of the target news event, the timeliness index is used to reflect the timeliness of the occurrence of the target news event, and the topic index is used to reflect the propagation degree of the target news event.

[0067] Specifically, the timeliness index is used to measure the timeliness and effectiveness of the occurrence of the news event, and the greater the timeliness index, the closer the news occurs; the topic index is used to represent the heat of a news reported by multiple news media, and the higher the topic index, the wider the propagation of the news.

[0068] In step S360, the push order of the plurality of target news events is determined according to the push priority, and the plurality of target news events carrying information of the push order are sent to the terminal device, for instructing the terminal device to display the plurality of target news events according to the push order.

[0069] Specifically, the display interface of the terminal device is displayed from high to low according to the push order, and the news with the highest push priority value is fixed to the top, and the rest is arranged in descending order of score. For details, please refer to Figure 4 , Figure 4 is a news display interface of a terminal device provided by an embodiment of the present application, as shown in Figure 4 The display interface 40 of the terminal device includes:

[0070] The title bar 41: each news is presented in a card layout, the title bar 41 includes importance: five stars, publishing time: 2025-01-01, and content, the risk label such as “quaternary-reduction ratio” and “secondary-large transaction” is directly displayed below the title, the label can be color-coded, and the specific content “XXX…” is included. Specifically, quaternary red, tertiary orange, secondary yellow, and primary blue are used to intuitively distinguish the risk level; the auxiliary information area 411: the timeliness identifier such as “5 minutes ago” and “1 hour ago” is displayed on the right side of the title, and the topic heat identifier such as “high topic degree and high attention degree” is displayed, to help the user quickly judge the timeliness and propagation degree of the news; the pull-down button 412 of the associated news event list display, the associated news of the same event, such as the subsequent reduction progress report of the same company, is aggregated in the form of the preferentially published news+pull-down list. The ellipsis pattern is displayed on the right side of the main news card, for reference Figure 4The user clicks or pulls down the card to expand the related news sub-list 413 of the news. The sub-list 413 can include A reports: xxxx belonging to the background of the event and B reports: xxxx belonging to the development of the event. The related news in the sub-list 413 can be sorted in chronological order from early to late, such that the news published first is in the front and the follow-up reports are in the back. Alternatively, the related news can be displayed in the order of the development of the event, such that the background of the event is displayed first and the development of the event is displayed subsequently.

[0071] In actual applications, updating the high-priority news can also display a dynamic effect of updating the news. For example, when the server pushes a new high-priority news, the priority of which is higher than that of the current top news or the score is updated, a light-colored prompt bar is popped up at the top of the terminal interface, such as "3 high-priority news updates", and the news card is inserted into the corresponding position with a dynamic effect. If the original top news is replaced, it is automatically sunk to the associated list, accompanied by a downward sliding animation.

[0072] It can be seen that, by using the information processing method based on news events and the related device provided in the present application, first, the news events pushed by multiple news sources are collected, and the feature information in the news events is extracted. Second, according to the feature information, the duplicate news events from different news sources are excluded, and multiple target news events are obtained. Third, the risk labels within the risk label range are selected to label the multiple target news events. Fourth, according to the risk labels, the risk scores of each target news event in the multiple target news events are determined. Fifth, according to the risk scores, the timeliness indicators, and the topic indicators, the push priorities of each target news event in the multiple target news events are calculated and determined. Finally, according to the push priorities, the push order of the multiple target news events is determined, and the multiple target news events carrying the information of the push order are sent to the terminal device. In this way, the event risk analysis of the news events and the like is performed, and the news push order is dynamically adjusted according to the business requirements, which effectively reduces the load of the server, reduces the resource consumption, reduces the news push delay, and improves the information processing efficiency.

[0073] In one possible embodiment, according to the risk labels, the risk scores of each target news event in the multiple target news events are determined, which includes: determining the types and quantities of the risk labels in each target news event. Each target news event includes multiple risk labels. The types of the risk labels are divided into first-level risk, second-level risk, third-level risk, and fourth-level risk according to the influence degree on the financial business from low to high. According to the types and quantities of the risk labels, the risk scores of each target news event are calculated. The risk scores are positively correlated with the influence degree of the risk labels on the financial business, and the risk scores are positively correlated with the quantities of the risk labels.

[0074] Among them, according to the influence degree of financial business, the risk label is divided into four levels from low to high; exemplary: first-level risk, slight influence on financial business, such as postponing the shareholders' meeting, etc.; second-level risk, certain influence on financial business, such as information disclosure delay, etc.; third-level risk, greater influence on financial business, such as abnormal related party transactions, etc.; fourth-level risk, serious influence on financial business, such as financial fraud, etc.

[0075] Specifically, in the process of counting the multiple risk labels carried by the target news event, the keywords matching the risk labels in the news text are extracted by the large model, and the number of occurrences of each level label is counted. Based on the above formula for determining the risk score:

[0076] α = 12 x log(c4 + 1) + 3 x log(c3 + 1) + 1 x log(c2 + 1) + 0.5 x log(c1 + 1);

[0077] Wherein, ci is the number of the i-th level risk label, and the coefficients 12, 3, 1 and 0.5 are the weights of each level of risk, and the weight of the fourth level risk is the highest, which reflects the principle of priority of high-impact labels.

[0078] It can be seen that in the present embodiment, the type and number of risk labels in each target news event are determined; according to the type and number of risk labels, the risk score of each target news event is calculated. In this way, the large model automatically counts the label type and number, effectively improving the timeliness requirement of news event monitoring; when the news event continues to develop, new risk labels are added, ensuring that the risk score is updated in real time with the evolution of the event, solving the problem of lagging behind of traditional static score.

[0079] In other possible embodiments, the weights of each risk level configuration are not fixed values, but can be dynamically adjusted according to the differences in risk sensitivity of different industries to adapt to the current risk monitoring needs. It can also be that the same enterprise subject is combined with historical data to increase additional weights for repeated risk labels, reflecting the cumulative effect of risk, so that the score is more in line with the evolution trend of actual risk.

[0080] In a possible embodiment, before the push priority of each target news event in the plurality of target news events is determined according to the risk score, the timeliness indicator, and the topic indicator, the method further includes: determining a publishing time of each target news event; determining a publishing time difference of each target news event according to the publishing time and a current time; determining a level of the timeliness indicator according to the publishing time difference, wherein the level of the timeliness indicator includes, from high to low, a first timeliness, a second timeliness, a third timeliness, and a fourth timeliness according to the timeliness of the occurrence of the target news event; wherein the publishing time difference is less than or equal to a first time length, the timeliness indicator is the first timeliness; the publishing time difference is greater than the first time length and less than or equal to a second time length, the timeliness indicator is the second timeliness; the publishing time difference is greater than the second time length and less than or equal to a third time length, the timeliness indicator is the third timeliness; the publishing time difference is greater than the third time length, the timeliness indicator is the fourth timeliness; determining a number of news sources reporting each target news event within a preset time window; and determining a total number of the plurality of news sources; and calculating the topic indicator according to the number of news sources reporting each target news event and the total number of the plurality of news sources.

[0081] The timeliness indicator is used to measure the time and timeliness of the occurrence of a news event, and the timeliness indicator is greater, the closer the news occurs. The calculation method is to use a step attenuation strategy through a timestamp difference, and weight distribution is achieved through a segmented function. The specific formula is as follows:

[0082] B = current_time (current time) publish_time (publishing time);

[0083] ;

[0084] In the formula, B is the publishing time difference, Time-score is the timeliness indicator, the first timeliness in the above formula refers to the timeliness of the publishing time difference being less than or equal to the first time length, that is, one hour, at this time, the first timeliness is used to reflect the most timely and rapid news event reporting; the second timeliness refers to the timeliness of the publishing time difference being greater than the first time length and less than or equal to the second time length, that is, 24 hours; the third timeliness refers to the timeliness of the publishing time difference being greater than the second time length and less than or equal to the third time length, that is, 7 days; the fourth timeliness refers to the timeliness of the publishing time difference being greater than 7 days, which corresponds to 0 on the timeliness indicator, that is, losing timeliness.

[0085] The topic indicator refers to the heat of a news being reported by multiple news media. The higher the topic indicator, the wider the news dissemination. The specific formula is as follows:

[0086] Topic indicator = ;

[0087] For example, the server retrieves full-volume news sources, 200 by default, including XX Information, XX Fortune Network, etc., and finds that there are 30 news sources reporting news Z; the topic index is equal to 30 / 200 = 0.15.

[0088] Specifically, the news sources can also be classified by source authority, into core sources and ordinary sources, and the topic index calculation formula is adjusted as follows: topic index = (core source reporting number x 1.5 + ordinary source reporting number x 1.0) / total source number.

[0089] As can be seen, in this embodiment, the timeliness and topic heat of news are converted into objective numerical values through time difference segmentation and reporting source proportion, so that the determination of the push priority is more objective, and the risk tracking efficiency of news event risk monitoring is improved.

[0090] In one possible embodiment, the push priority of each target news event in the plurality of target news events is calculated and determined, including: normalizing the risk score, the timeliness index, and the topic index; configuring a weight parameter for the normalized risk score, the timeliness index, and the topic index, wherein the weight parameter corresponding to the risk score is greater than the weight parameter corresponding to the timeliness index, and the weight parameter corresponding to the timeliness index is greater than the weight parameter corresponding to the topic index; and calculating and determining the push priority according to the normalized risk score, the timeliness index, and the topic index, and the weight parameter.

[0091] Among them, the server normalizes the risk score, the timeliness index, and the topic index, maps the indexes of different magnitudes to the range of 0-1, and ensures that the weighted calculation can be performed; the specific normalization processing includes: risk score normalization, the original risk score of a certain target news event is 10.3965, the maximum value of the risk score in the full-volume news is 15.0, and the minimum value is 0, so the normalized risk score = 10.3965 / 15.0 ≈ 0.693; timeliness index normalization, the news publishing time difference is 45 minutes (first-level timeliness), and the first-level timeliness corresponds to 1.0 by default, so the normalized timeliness index = 1.0; topic index normalization: the news is reported by 30 sources (total sources 200), and the topic index = 30 / 200 = 0.15.

[0092] Among them, the weighted calculation formula is as follows:

[0093] Push priority = 4 x risk score + 3 x timeliness index + 2 x topic index.

[0094] Among them, the risk score weight is the highest, because the risk is the core concern element; the timeliness index weight is the second, to ensure real-time events first; the topic index weight is used for assistance, reflecting the spread of the breadth.

[0095] It can be seen that in the embodiment, the push priority is calculated and determined according to the normalized risk score, timeliness index and topic index, and the weight parameter. In this way, compared with the traditional sorting method, the push priority of the present solution reduces subjective interference in priority determination, and the differences between different news can be directly reflected by the score; so as to push the news with the highest importance to the customer in the first time from the risk point of view, and improve the efficiency of news event risk monitoring.

[0096] In other possible embodiments, for different news types, the weight parameter can be reconfigured after a large amount of data statistical test; for example, when detecting an emergency event such as cross-market fluctuation, the system triggers temporary adjustment of the weight: the risk score weight is increased to 6, and the timeliness index weight is increased to 4, to ensure that the emergency event is displayed at the top within a few minutes.

[0097] Specifically, please refer to Figure 5 , Figure 5 is a flowchart of another information processing method based on news events provided by the embodiment of the present application. In one possible embodiment, after comparing the cosine similarity of any two vectors in the semantic vector, as shown in Figure 5 , the following steps are included:

[0098] S510, if the cosine similarity of the two vectors is less than the first similarity threshold and greater than or equal to the second similarity threshold, it is determined that the news events corresponding to the two vectors are associated news events of the same event.

[0099] Specifically, for example, the first similarity threshold W1=91% is preset, that is, the repeated news determination threshold, and the second similarity threshold W2=70%; when it is determined according to the above scheme that the H report of A news event is the news to be published first (target news event), and the cosine similarity of the L report is 85%, since 70%<85%<91%, it is determined that the H report and the L report are associated news events of the same A news event, and the L report is the related news of the news to be published first, which can be the subsequent development of the A news event, or the related background of the A news event, etc.

[0100] S520, statistics of a plurality of associated news events, and establishing a news event list respectively.

[0101] Among them, the news event list includes multiple associated news events of the same event theme and the target news event. For example, the H and L reports in the above example, the H event is the target news event for priority push, and the L report is the associated news event, which is counted into the list of A news events.

[0102] S530, calculate the push priority of the multiple associated news events.

[0103] Among them, for multiple associated news events, the push priority needs to be calculated to determine the display order in the news event list, so that the user can quickly locate the more urgent, more important and higher heat news events when pulling down the news event list.

[0104] S540, if the push priority of at least one event in the multiple associated news events is higher than that of the target news event, replace the target news event with the event in the news event list.

[0105] Among them, there may be a subsequent news event whose push priority is higher than that of the target news event, and the new subsequent news event replaces the target news event as the priority published news for push and display on the terminal device. It should be understood that at this time, the previous target news event is folded into the news event list and displayed by the user taking the click-down operation or other viewing operation.

[0106] S550, send the news event list and prompt information to the terminal device.

[0107] Among them, the prompt information is used to prompt the user to execute a pull-down instruction at the terminal device to browse the multiple associated news events included in the news event list. Specifically, the prompt information can be text, pattern, etc., and the guide key of the news event list is displayed around the target news event, which can be presented as an ellipsis, referring to the pull-down key 412 in Figure 4 , the user executes a pull-down instruction at the terminal device for the guide key, including: finger clicking the guide key, mouse clicking the guide key, etc., which is not limited here; at this time, in response to the user's pull-down instruction, the news event list is expanded to display the multiple related news events about the target news event.

[0108] It can be seen that in this embodiment, through specific threshold judgment, list construction and terminal interaction, the development context of the news event is intuitively and efficiently displayed, and the efficiency and completeness of news event analysis are effectively improved.

[0109] In a possible embodiment, the method further comprises: determining the time of the plurality of target news events; for the plurality of target news events occurring on the current day, re-computing the push priority of the plurality of target news events according to a first preset period; and for the plurality of target news events occurring before the current day, re-computing the push priority of the plurality of target news events according to a second preset period, wherein the length of the first preset period is less than the length of the second preset period.

[0110] The first preset period can be an hour, and the second preset period can be a day; the news score re-computation and the timing update are performed every hour for the news of the current day; and the news score re-computation is performed every day for the news of the previous day.

[0111] Specifically, the server monitors the newly pushed news of the news source in real time or according to a monitoring period, determines the push priority of the newly pushed news according to the method, and dynamically sorts the target news events in combination with the previously determined target news events.

[0112] It can be seen that, in the embodiment, the push priority is re-computed in a short period for the news of the current day through the score re-computation mechanism, the priority fluctuation caused by the timeliness index, the topic degree change or the newly added risk label can be captured in real time, and the news with high risk and high timeliness can be continuously pushed to the top; the historical news before the current day is updated in a long period, the waste of server computing power caused by indiscriminate high-frequency calculation is avoided, and the efficiency of news pushing is improved.

[0113] The embodiments of the application can divide the functional units of the electronic device according to the above method examples, for example, each functional unit can be divided according to each function, or two or more functions can be integrated in one processing unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit. It should be noted that the division of the units in the embodiments of the application is illustrative, and is only a logical functional division. When actually implemented, another division mode can be used.

[0114] Consistent with the embodiments of Figure 3 , see Figure 6 , Figure 6 is a functional unit composition block diagram of a news event-based information processing apparatus provided by the embodiments of the application, as shown in Figure 6As shown, the news event-based information processing apparatus 600 comprises a collection unit 610, a screening unit 620, a selection unit 630, a determination unit 640, a calculation unit 650, and a pushing unit 660. The collection unit 610 is specifically configured to collect news events pushed by a plurality of news sources, and extract feature information in the news events, the feature information comprising at least one of the following: a person, a place, a time, an event subject, and an event summary. The screening unit 620 is specifically configured to exclude duplicate news events from different news sources according to the feature information, and obtain a plurality of target news events. The selection unit 630 is specifically configured to select risk labels for the plurality of target news events within a risk label range, the risk label range being a label range pre-determined by an operation and maintenance personnel of a news pushing system, the risk label being a set of keywords related to financial risks, and the risk label representing an influence degree of the target news event on a financial business. The determination unit 640 is specifically configured to determine a risk score of each target news event in the plurality of target news events according to the risk label. The calculation unit 650 is specifically configured to calculate a pushing priority of each target news event in the plurality of target news events according to the risk score, a timeliness index, and a topic index, the pushing priority being used to reflect an importance degree of the target news event, the timeliness index being used to reflect a timeliness of occurrence of the target news event, and the topic index being used to reflect a propagation degree of the target news event. The pushing unit 660 is specifically configured to determine a pushing order of the plurality of target news events according to the pushing priority, and send the plurality of target news events carrying the pushing order to a terminal device, for instructing the terminal device to display the plurality of target news events according to the pushing order.

[0115] In one possible embodiment, in terms of excluding duplicate news events from different news sources according to the feature information, the screening unit 620 is specifically configured to: convert text information of the plurality of feature information into semantic vectors by a pre-trained deep learning model based on a natural language processing algorithm, the semantic vector being a vector with fixed dimensions and used to represent deep semantic meanings of the feature information; compare cosine similarities of any two vectors in the semantic vectors; if the cosine similarity of the two vectors is greater than or equal to a first similarity threshold, it is determined that the feature information corresponding to the two vectors is consistent, and the consistent feature information represents that the associated news events are duplicate news events; and determine times of the plurality of duplicate news events according to the feature information, and retain a news event with the earliest time.

[0116] In a possible implementation, in the aspect of determining the risk score of each target news event in the plurality of target news events according to the risk label, the determining unit 640 is specifically configured to: determine the type and quantity of the risk label in each target news event, each target news event including a plurality of risk labels, the type of the risk label being classified into first-level risk, second-level risk, third-level risk, and fourth-level risk from low to high according to the degree of influence on the financial business; and calculate the risk score of each target news event according to the type and quantity of the risk label, the risk score being positively correlated with the degree of influence of the risk label on the financial business and positively correlated with the quantity of the risk label.

[0117] In a possible implementation, before the aspect of calculating the push priority of each target news event in the plurality of target news events according to the risk score, the timeliness index, and the topic index, the calculating unit 650 is specifically configured to: determine the publishing time of each target news event; calculate the publishing time difference of each target news event according to the publishing time and the current time; and determine the level of the timeliness index according to the publishing time difference, wherein the level of the timeliness index includes first-level timeliness, second-level timeliness, third-level timeliness, and fourth-level timeliness from high to low according to the timeliness of the occurrence of the target news event, wherein the publishing time difference is less than or equal to a first time length, the timeliness index is the first-level timeliness, the publishing time difference is greater than the first time length and less than or equal to a second time length, the timeliness index is the second-level timeliness, the publishing time difference is greater than the second time length and less than or equal to a third time length, the timeliness index is the third-level timeliness, the publishing time difference is greater than the third time length, and the timeliness index is the fourth-level timeliness; determine the number of news sources reporting each target news event within a preset time window; and determine the total number of the plurality of news sources; and calculate the topic index according to the number of news sources reporting each target news event and the total number of the plurality of news sources.

[0118] In a possible implementation, in the aspect of calculating the push priority of each target news event in the plurality of target news events, the calculating unit 650 is specifically configured to: perform normalization processing on the risk score, the timeliness index, and the topic index; configure a weight parameter for the risk score, the timeliness index, and the topic index after the normalization processing, wherein the weight parameter corresponding to the risk score is greater than the weight parameter corresponding to the timeliness index, and the weight parameter corresponding to the timeliness index is greater than the weight parameter corresponding to the topic index; and calculate the push priority according to the risk score, the timeliness index, and the topic index after the normalization processing and the weight parameter.

[0119] In one possible embodiment, after comparing the cosine similarity of any two vectors in the semantic vectors, the filtering unit 620 is further configured to: if the cosine similarity of the two vectors is greater than a first similarity threshold and less than or equal to a second similarity threshold, then determine that the news events corresponding to the two vectors are related news events of the same event; count multiple related news events and establish news event lists respectively, the news event lists including multiple related news events with the same event theme and a target news event; calculate and determine the push priority of multiple related news events; if at least one of the multiple related news events has a push priority higher than the push priority of the target news event, then replace the target news event with that event in the news event list; send the news event list and a prompt message to the terminal device, the prompt message being used to prompt the user to execute a pull-down command on the terminal device to browse the multiple related news events included in the news event list.

[0120] In one possible embodiment, the information processing device 600 based on news events is further configured to: determine the time of multiple target news events; for multiple target news events occurring on the same day, recalculate and determine the push priority of multiple target news events according to a first preset period; for multiple target news events occurring before the same day, recalculate and determine the push priority of multiple target news events according to a second preset period, wherein the duration of the first preset period is less than the duration of the second preset period.

[0121] It is understood that since the method embodiments and the device embodiments are different presentations of the same technical concept, the content of the method embodiment section in this application should be adapted to the device embodiment section in a synchronous manner, and will not be repeated here.

[0122] Figure 7 This is a structural block diagram of an electronic device provided in an embodiment of this application. For example... Figure 7 As shown, electronic device 700 may include one or more components: a processor 701 and a memory 702 coupled to the processor 701, wherein the memory 702 may store one or more computer programs, which may be configured to implement the methods described in the examples above when executed by one or more processors 701. Electronic device 700 may be as follows: Figure 1 The server shown is 110.

[0123] The processor 701 can include one or more processing cores. The processor 701 connects various parts within the entire electronic device 700 by various interfaces and lines, performs various functions of the electronic device 700 and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 702, and calling data stored in the memory 702. Optionally, the processor 701 can be implemented in at least one of a hardware form of a digital signal processing (DSP), a field-programmable gate array (FPGA), a programmable logic array (PLA). The processor 701 can integrate a combination of one or several of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. It can be understood that the above-mentioned modem can also not be integrated into the processor 701, but can be implemented by a separate communication chip.

[0124] The memory 702 can include a random access memory (RAM) and can also include a read-only memory (ROM). The memory 702 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 702 can include a program storage area and a data storage area, wherein the program storage area can store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-mentioned various method examples, etc. The data storage area can also store data created by the electronic device 700 in use, etc.

[0125] It can be understood that the electronic device 700 can include more or fewer structural elements than those in the above structural block diagram, for example, a power module, a physical key, a WiFi (Wireless Fidelity) module, a speaker, a Bluetooth module, a sensor, etc., which are not limited herein.

[0126] The embodiments of the present application also provide a computer storage medium, wherein a computer program / instructions are stored on the computer storage medium, and the computer program / instructions are executed by a processor to implement some or all steps of any method described in the above method embodiments.

[0127] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments.

[0128] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0129] In the several embodiments provided in this application, it should be understood that the disclosed methods, apparatuses, and systems can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for example, the division of units is merely a logical functional division, and there may be other division methods in actual implementation; for example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0130] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0131] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can be physically comprised separately, or two or more units can be integrated into one unit. The integrated unit described above can be implemented in hardware or in the form of hardware plus software functional units.

[0132] The integrated unit in the form of software function unit can be stored in a computer readable storage medium. The software function unit is stored in a storage medium, including a plurality of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute part of the steps of the method of various embodiments of the present application. The storage medium includes a U disk, a mobile hard disk, a magnetic disk, an optical disk, a volatile memory or a non-volatile memory. The non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM) or a flash memory. The volatile memory can be a random access memory (RAM) used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) can be used, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM) and direct rambus random access memory (DR RAM), etc. Various media that can store program codes.

[0133] Although the present application is disclosed as above, the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions without departing from the spirit and scope of the present application, and various modifications can be made, including combinations of different functions and implementation steps, including software and hardware implementations, which are all within the scope of the present application.

Claims

1. A news event-based information processing method characterized by comprising: The application discloses a server applied to a news pushing system, and a method for pushing news events. The method comprises the following steps: Collecting news events pushed by multiple news sources, and extracting feature information in the news events, wherein the feature information comprises at least one of the following: a person, a place, a time and an event summary; According to the feature information, excluding repeated news events from different news sources to obtain multiple target news events; Selecting a risk label in a risk label range to mark the multiple target news events, wherein the risk label range is a label range pre-determined by an operation and maintenance personnel of the news pushing system, the risk label is a keyword set related to financial risk, and the risk label represents an influence degree of the target news event on financial business; According to the risk label, determining a risk score of each target news event in the multiple target news events; Determining a publishing time of each target news event, calculating a publishing time difference of each target news event according to the publishing time and a current time, determining a timeliness index level according to the publishing time difference, wherein the timeliness index level comprises, from high to low, a first-level timeliness, a second-level timeliness, a third-level timeliness and a fourth-level timeliness; wherein the publishing time difference is less than or equal to a first time length, and the timeliness index is the first-level timeliness; the publishing time difference is greater than the first time length and less than or equal to a second time length, and the timeliness index is the second-level timeliness; the publishing time difference is greater than the second time length and less than or equal to a third time length, and the timeliness index is the third-level timeliness; the publishing time difference is greater than the third time length, and the timeliness index is the fourth-level timeliness; determining a number of news sources reporting each target news event in a preset time window; determining a total number of the multiple news sources; and calculating a topic index according to the number of news sources reporting each target news event and the total number of the multiple news sources; According to the risk score, the timeliness index and the topic index, calculating a pushing priority of each target news event, wherein the pushing priority is used to reflect an importance degree of the target news event, the timeliness index is used to reflect a timeliness degree of the target news event, and the topic index is used to reflect a spread degree of the target news event; 2. The method of claim 1, wherein, According to the pushing priority, determining a pushing order of the multiple target news events, and sending the multiple target news events carrying information of the pushing order to the terminal device, so as to instruct the terminal device to display the multiple target news events according to the pushing order. According to the feature information, excluding repeated news events from different news sources comprises the following steps: Based on a natural language processing algorithm, converting text information of multiple feature information into semantic vectors by using a pre-trained deep learning model, wherein the semantic vectors are vectors with fixed dimensions and used to represent deep semantic meanings of the feature information. comparing cosine similarities of any two vectors in the semantic vectors; if the cosine similarity of the two vectors is greater than or equal to a first similarity threshold, determining that the feature information corresponding to the two vectors is consistent, and the consistent feature information represents that the associated news event is a repeated news event; determining a time of a plurality of repeated news events according to the feature information, and retaining a news event with the earliest time.

3. The method of claim 2, wherein, The method further comprises: determining a type and a number of the risk labels in each target news event, wherein the target news event comprises a plurality of risk labels, and the type of the risk label is classified into a first-level risk, a second-level risk, a third-level risk, and a fourth-level risk according to an influence degree on the financial business from low to high; calculating the risk score of each target news event according to the type and the number of the risk labels, wherein the risk score is positively correlated with the influence degree of the risk label on the financial business, and the risk score is positively correlated with the number of the risk labels.

4. The method of claim 3, wherein, The method further comprises: normalizing the risk score, the timeliness index, and the topic index; configuring weight parameters for the normalized risk score, the timeliness index, and the topic index, wherein the weight parameter corresponding to the risk score is greater than the weight parameter corresponding to the timeliness index, and the weight parameter corresponding to the timeliness index is greater than the weight parameter corresponding to the topic index; calculating and determining the push priority according to the normalized risk score, the timeliness index, and the topic index, and the weight parameters.

5. The method of claim 2, wherein, After the comparison of the cosine similarities of any two vectors in the semantic vectors, the method further comprises: if the cosine similarity of the two vectors is less than the first similarity threshold and greater than or equal to a second similarity threshold, determining that the two vectors correspond to associated news events of the same event; counting a plurality of associated news events, and establishing a news event list respectively, wherein the news event list comprises the plurality of associated news events of the same event theme and the target news event; calculating and determining the push priority of the plurality of associated news events; if the push priority of at least one event in the plurality of associated news events is higher than the push priority of the target news event, replacing the target news event with the event in the news event list; sending the news event list and prompt information to the terminal device, wherein the prompt information is used to prompt the user to execute a pull-down instruction at the terminal device to browse the plurality of associated news events included in the news event list.

6. The method of claim 1, wherein, The method further comprises: determining a time of the plurality of target news events; for the plurality of target news events occurring on the same day, recalculating and determining the push priority of the plurality of target news events according to a first preset period. For the plurality of target news events occurring before the day, the push priority of the plurality of target news events is recalculated according to a second preset period, wherein the first preset period is shorter than the second preset period.

7. A news event-based information processing apparatus characterized by comprising: The server is applied to a news push system, and the news push system comprises the server and a terminal device of a user. The collection unit is specifically configured to collect news events pushed by a plurality of news sources, and extract feature information in the news events, the feature information comprising at least one of the following: a person, a place, a time, an event subject, and an event summary. The screening unit is specifically configured to exclude duplicate news events from different news sources according to the feature information, and obtain a plurality of target news events. The selection unit is specifically configured to select a risk label within a risk label range to mark the plurality of target news events, the risk label range being a label range pre-determined by an operation and maintenance personnel of the news push system, the risk label being a set of keywords related to financial risks, and the risk label representing an influence degree of the target news event on financial business. The determination unit is specifically configured to determine a risk score of each target news event in the plurality of target news events according to the risk label, determine a publication time of the each target news event, calculate a publication time difference of the each target news event according to the publication time and a current time, determine a grade of a timeliness index according to the publication time difference, wherein the grade of the timeliness index comprises, from high to low, a first timeliness, a second timeliness, a third timeliness, and a fourth timeliness, wherein the publication time difference is less than or equal to a first time length, the timeliness index is the first timeliness, the publication time difference is greater than the first time length and less than or equal to a second time length, the timeliness index is the second timeliness, the publication time difference is greater than the second time length and less than or equal to a third time length, the timeliness index is the third timeliness, the publication time difference is greater than the third time length, the timeliness index is the fourth timeliness, determine a number of news sources reporting the each target news event within a preset time window, determine a total number of the plurality of news sources, and calculate a topic index according to the number of news sources reporting the each target news event and the total number of the plurality of news sources. The calculation unit is specifically configured to calculate a push priority of the each target news event according to the risk score, the timeliness index, and the topic index, the push priority being used to reflect an importance degree of the target news event, the timeliness index being used to reflect a timeliness degree of the target news event, and the topic index being used to reflect a propagation degree of the target news event. The pushing unit is specifically configured to determine a pushing sequence of the plurality of target news events according to the pushing priorities, and send the plurality of target news events carrying the pushing sequence to the terminal device, so as to instruct the terminal device to display the plurality of target news events according to the pushing sequence.

8. An electronic device, comprising: A computer program product comprising a processor, a memory storing one or more programs configured for execution by the processor, the programs comprising instructions for performing the steps of any of the methods of claims 1-6.

9. A computer-readable storage medium, characterized in that, A computer program for electronic data interchange, wherein the computer program causes a computer to perform the method of any of claims 1-6.

Images