Event development prediction method and device, program product and electronic equipment

Abstract

The invention provides an event development prediction method and device, a program product and electronic equipment, and relates to the technical field of computers. The method comprises the following steps: acquiring first event information of a to-be-predicted event; processing the first event information by using a plurality of professional prediction models corresponding to a plurality of fields to obtain a plurality of first development prediction results of the to-be-predicted event; if the consensus degree of the plurality of first development prediction results does not reach a consensus threshold value, generating a reasoning chain of the to-be-predicted event according to prior information related to the first event information; and determining a second development prediction result of the to-be-predicted event according to the reasoning chain. According to the invention, the accuracy and interpretability of the event development prediction result are improved.

Description

Technical Field

[0001] The embodiments of this disclosure relate to the field of computer technology, and more specifically, the embodiments of this disclosure relate to event development prediction methods, event development prediction devices, computer program products, and electronic devices. Background Technology

[0002] This section is intended to provide background or context for the technical solutions stated in the claims. The description of the relevant content in this section does not imply an admission that it is prior art.

[0003] Event prediction refers to forecasting the future trend or outcome of an event based on existing information. With the development of the internet and communications, the speed and impact of event dissemination are increasing daily. For individuals or groups, accurately predicting event development is beneficial for optimizing decision-making and strategic planning, and reducing risks.

[0004] In related technologies, events are typically predicted manually by domain experts, or by using traditional machine learning models to predict the development of events. Summary of the Invention

[0005] However, human-made predictions of event development rely on the expertise, experience, and subjective perception of domain experts, making their accuracy and objectivity unreliable. Furthermore, traditional machine learning models often possess knowledge only in a single domain, resulting in biased predictions, low accuracy, and poor interpretability.

[0006] In view of the above problems, embodiments of this disclosure provide an event development prediction method, an event development prediction device, a computer program product, and an electronic device.

[0007] According to a first aspect of this disclosure, an event development prediction method is provided, the method comprising: acquiring first event information of an event to be predicted; processing the first event information using multiple professional prediction models corresponding to multiple fields to obtain multiple first development prediction results of the event to be predicted; if the consensus of the multiple first development prediction results does not reach a consensus threshold, generating an inference chain of the event to be predicted based on prior information related to the first event information; and determining a second development prediction result of the event to be predicted based on the inference chain.

[0008] Optionally, the step of processing the first event information using multiple professional prediction models corresponding to multiple fields to obtain multiple first development prediction results for the event to be predicted includes: obtaining first domain knowledge associated with the first event information from a knowledge base; the first domain knowledge includes knowledge of the target domain to which the event to be predicted belongs and knowledge of non-target domains; fusing the first domain knowledge to obtain second domain knowledge; inputting the first event information and the second domain knowledge into the multiple professional prediction models, and outputting the corresponding first development prediction result through each professional prediction model.

[0009] Optionally, fusing the first domain knowledge to obtain the second domain knowledge includes: determining the influence of the domain to which the first domain knowledge belongs on the event to be predicted; and weighting and fusing the first domain knowledge according to the influence to obtain the second domain knowledge.

[0010] Optionally, the prior information includes a first historical event similar to the event to be predicted; the inference chain includes a first inference chain; generating the inference chain of the event to be predicted based on the prior information related to the first event information includes: searching for the first historical event in the event memory bank based on the first event information; and generating the first inference chain based on the first event development information corresponding to the first historical event.

[0011] Optionally, the prior information includes third-domain knowledge related to the event to be predicted; the inference chain includes a second inference chain; generating the inference chain for the event to be predicted based on the prior information related to the first event information includes: matching the first event information with rule information in a knowledge base, determining the third-domain knowledge based on the matching result; and generating the second inference chain based on the third-domain knowledge.

[0012] Optionally, the prior information includes a second historical event of the same category as the event to be predicted; the inference chain includes a third inference chain; generating the inference chain of the event to be predicted based on the prior information related to the first event information includes: searching for the second historical event in the event memory bank according to the category of the event to be predicted; performing statistics on the development information of the second event corresponding to the second historical event, and generating the third inference chain based on the statistical results.

[0013] Optionally, the event memory bank includes a short-term memory sub-bank, a medium-term memory sub-bank, and a long-term memory sub-bank; the method further includes: determining short-term historical events whose time distance from the present does not exceed a first time threshold, and storing their event information in the short-term memory sub-bank; determining medium-term historical events whose time distance from the present exceeds the first time threshold but does not exceed a second time threshold, selecting medium-term historical events whose importance index is greater than a first importance threshold, and storing their event information in the medium-term memory sub-bank; determining long-term historical events whose time distance from the present exceeds a second time threshold, selecting long-term historical events whose importance index is greater than a second importance threshold, and storing their event information in the long-term memory sub-bank; wherein, the first time threshold is less than the second time threshold, and the first importance threshold is less than the second importance threshold.

[0014] Optionally, the method further includes: determining the influence of the historical event based on its dissemination scope, attention, and response; determining the scarcity of the historical event based on its novelty and uniqueness; determining the relevance of the historical event based on the number of other events associated with it and the depth of its association with the field; and determining the importance index of the historical event based on its influence, scarcity, and relevance.

[0015] Optionally, the short-term memory sub-database uses a cache database, the medium-term memory sub-database uses a document database, and the long-term memory sub-database uses a graph database.

[0016] Optionally, the method further includes: finding a third historical event based on the reasoning result corresponding to the reasoning chain, wherein the development information of the third event corresponding to the third historical event is opposite to the reasoning result; generating reverse questioning information of the reasoning chain based on the third historical event; and determining the second development prediction result of the event to be predicted based on the reasoning chain, which includes: determining the second development prediction result based on the reasoning chain and the reverse questioning information.

[0017] Optionally, determining the second development prediction result based on the reasoning chain and the counter-questioning information includes: assessing a first credibility of the reasoning chain; determining a contradiction index of the reasoning result based on the counter-questioning information; determining a second credibility of the reasoning result based on the first credibility and the contradiction index; and determining the second development prediction result based on the second credibility.

[0018] Optionally, the step of searching for a third historical event based on the reasoning result corresponding to the reasoning chain includes: searching for third event development information and corresponding third historical events that are opposite to the reasoning result from the event memory bank; and / or, adding the event to be predicted to an event association network, and searching for nodes with a reverse association relationship with the event to be predicted from the event association network to determine the third historical event.

[0019] Optionally, the method further includes: acquiring multiple historical events, treating each historical event as a node, generating edges between different nodes based on the association between different historical events, and constructing the event association network; the association includes positive association, negative association, causal relationship, temporal association, and domain association.

[0020] Optionally, after adding the event to be predicted to the event association network, the method further includes: monitoring the traffic of the node corresponding to the event to be predicted in the event association network and the traffic of the associated edge; and determining whether the event to be predicted has experienced abnormal fluctuations based on the traffic of the node or the traffic of the edge.

[0021] According to a second aspect of this disclosure, an event development prediction apparatus is provided, the apparatus comprising: a first event information acquisition module configured to acquire first event information of an event to be predicted; a first prediction module configured to process the first event information using multiple professional prediction models corresponding to multiple fields respectively, to obtain multiple first development prediction results of the event to be predicted; an inference chain generation module configured to generate an inference chain of the event to be predicted based on prior information related to the first event information if the consensus of the multiple first development prediction results does not reach a consensus threshold; and a second prediction module configured to determine a second development prediction result of the event to be predicted based on the inference chain.

[0022] According to a third aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method of the first aspect described above and possible implementations thereof.

[0023] According to a fourth aspect of this disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the method of the first aspect and possible implementations thereof by executing the executable instructions.

[0024] The embodiments disclosed herein have the following technical effects:

[0025] This method utilizes multiple specialized prediction models from various fields to process the first event information of the event to be dealt with, resulting in multiple initial development predictions. If the consensus among these initial predictions does not reach a consensus threshold, an inference chain is generated based on prior information related to the first event information. This inference chain then determines a second development prediction, which serves as the final event development prediction. On one hand, employing specialized prediction models from different fields to generate multiple initial development predictions ensures the comprehensiveness of the prediction result. When the consensus among the initial predictions is insufficient, generating an inference chain and determining a second development prediction improves the accuracy and reliability of the final result. On the other hand, the inference chain reveals the information and logic upon which the prediction result is based, enhancing the interpretability of the prediction result. Attached Figure Description

[0026] Figure 1 A flowchart of an event development prediction method according to an embodiment of this disclosure is shown.

[0027] Figure 2 A flowchart illustrating one embodiment of this disclosure for determining a first development prediction result is shown.

[0028] Figure 3 A flowchart illustrating one embodiment of constructing an event memory bank is shown.

[0029] Figure 4 A flowchart illustrating one method for determining an importance index according to an embodiment of this disclosure is shown.

[0030] Figure 5 A schematic diagram of an event development prediction device according to an embodiment of this disclosure is shown.

[0031] Figure 6 A schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure is shown.

[0032] In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed Implementation

[0033] The principles and spirit of this disclosure are described in detail below with reference to several representative embodiments. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement this disclosure, and are not intended to limit the scope of this disclosure in any way. Rather, these embodiments are provided to make this disclosure more thorough and complete, and to fully convey the scope of this disclosure to those skilled in the art.

[0034] The embodiments of this disclosure can be implemented as systems, apparatuses, devices, methods, and computer program products. Therefore, this disclosure can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. Invention Overview

[0036] In related technologies, event predictions are typically made manually by domain experts or using traditional machine learning models. However, human predictions rely on the domain expert's expertise, experience, and subjective perception, making their accuracy and objectivity uncertain. Traditional machine learning models, on the other hand, often possess knowledge only in a single domain, neglecting cross-domain connections and interactions between events. This inability to fully grasp the complexity of events leads to biased and inaccurate predictions. Furthermore, traditional machine learning models suffer from poor interpretability; users cannot understand the information and logic upon which the model arrives at its predictions, making verification difficult and reducing the credibility of the predictions.

[0037] In view of the foregoing, this disclosure provides an event development prediction method, an event development prediction device, a computer program product, and an electronic device. Multiple professional prediction models corresponding to various fields are used to process the first event information of the event to be processed, resulting in multiple first development prediction results. If the consensus level of these results does not reach a consensus threshold, an inference chain is generated based on prior information related to the first event information. A second development prediction result is then determined based on the inference chain, serving as the final event development prediction result. On the one hand, using professional prediction models from different fields to perform predictions and obtain multiple first development prediction results ensures the comprehensiveness of the prediction result. When the consensus level of the first development prediction result does not meet the requirements, an inference chain is generated and a second development prediction result is determined, improving the accuracy and credibility of the final result. On the other hand, the inference chain can reveal the information and logic upon which the prediction result is based, enhancing the interpretability of the prediction result.

[0038] After introducing the basic principles of this disclosure, various non-limiting embodiments of this disclosure will be described in detail below.

[0039] Application Scenarios Overview

[0040] It should be noted that the following application scenarios are shown only to facilitate understanding of the spirit and principles of this disclosure. The implementation of this disclosure is not limited in this respect and can be applied to any applicable scenario.

[0041] The embodiments disclosed herein can be applied to platforms with dissemination capabilities such as news, social media, and video, to predict the subsequent development of news events, facilitating advance planning and resource allocation. For example, based on the predicted development of events within the current unit of time (a unit of time could be a day), a resource allocation plan for the next unit of time can be determined. Alternatively, if it is predicted that an event will subsequently develop into an abnormal situation, an early warning mechanism can be triggered in advance, and corresponding measures can be formulated.

[0042] Exemplary methods

[0043] The event development prediction method in the embodiments of this disclosure will be described by way of example below. Figure 1 The flowchart of an event development prediction method is shown, which may include the following steps S110 to S140:

[0044] Step S110: Obtain the first event information of the event to be predicted.

[0045] The first event information refers to the existing information about the event to be predicted, that is, the information that can be obtained from the event up to the present. In one implementation, the information of the event to be predicted is organized into structured information, such as RDF triples (Resource Description Framework, whose triples are usually composed of subject, predicate, and object), "entity + attribute" and other forms of information, as the first event information, which facilitates subsequent processing.

[0046] Step S120: The first event information is processed using multiple professional prediction models corresponding to multiple fields to obtain multiple first development prediction results of the event to be predicted.

[0047] These can include multiple fields such as economics, science and technology, and politics. A corresponding specialized forecasting model is built for each field to predict events within that field. The specialized forecasting model can employ algorithms and architectures suitable for that field. For example, a specialized forecasting model for the economic field could use a combined ARIMA (Autoregressive Integrated Moving Average) and GARCH (Generalized Autoregressive Conditional Heteroskedasticity) model. Its input is historical or current economic data related to economic events, and its output is a predicted sequence of economic indicators for a future period, such as the predicted GDP growth rate for the next three months. A specialized forecasting model for the science and technology field could use a deep learning model with a Transformer architecture. Its input is patent texts, technical reports, etc., related to science and technology events, and its output is a probability distribution of technological development trends, such as predicting the probability of a technological breakthrough within the next two years. Professional prediction models in the political field can adopt LSTM (Long Short-Term Memory) and Attention architectures. The input is policy texts, social media comments, etc. in the political field, and the output is a public opinion tendency classification vector (such as [0.8,0.1,0.1] representing 80% support, 10% neutral, and 10% opposition) and confidence level.

[0048] The information about the first event can be input into each specialized prediction model, and after processing, the first development prediction result corresponding to each specialized prediction model can be output.

[0049] In one implementation, reference Figure 2 As shown, the above-mentioned process utilizes multiple professional prediction models corresponding to multiple fields to process the information of the first event, thereby obtaining multiple first development prediction results for the event to be predicted, including the following steps S210 to S230:

[0050] Step S210: Obtain first domain knowledge associated with the first event information from the knowledge base; the first domain knowledge includes knowledge of the target domain to which the event to be predicted belongs and knowledge of non-target domains.

[0051] The knowledge base is used to store domain-specific knowledge. This knowledge can originate from well-known principles, rules, and policies across various domains, or from historical events, such as extracting commonalities from a large number of historical events. The domain-specific knowledge in the knowledge base is categorized and stored according to each domain. For example, a knowledge set for each domain can be established within the knowledge base to record knowledge within that domain, facilitating searches by domain. In one implementation, a domain-specific knowledge set can be configured for each domain-specific prediction model, enabling the model to utilize this domain-specific knowledge during training or application.

[0052] First-domain knowledge refers to knowledge related to the information of the first event. The target domain to which the event to be predicted belongs is the target domain, and other domains are non-target domains. First-domain knowledge includes not only knowledge from the target domain but also knowledge from non-target domains. For example, domain knowledge is presented in the form of RDF triples, such as [a policy, an impact, an economic indicator]. The information of the first event is matched with the RDF triples, such as by calculating semantic similarity. If the information of the first event matches any meta-information in the RDF triple, then the RDF triple is added to the first-domain knowledge.

[0053] Step S220: Integrate the knowledge from the first domain to obtain the knowledge from the second domain.

[0054] In this context, second-domain knowledge can be viewed as a collection of first-domain knowledge. For example, a matrix containing all first-domain knowledge can be generated based on first-domain knowledge in the form of triples, and this matrix can serve as second-domain knowledge. Furthermore, during fusion, the first-domain knowledge can be deduplicated and conflict-removed (for example, when first-domain knowledge from different domains conflicts or contradicts, a voting mechanism can be used to retain knowledge supported by the majority of domains) to improve the quality of second-domain knowledge.

[0055] In one implementation, the above-described fusion of knowledge in the first domain to obtain knowledge in the second domain includes the following steps:

[0056] Determine the influence of the domain to which the knowledge in the first domain belongs on the predicted event;

[0057] The knowledge in the first domain is weighted and fused according to its influence to obtain the knowledge in the second domain.

[0058] Generally, the higher the relevance of the event to be predicted to a certain domain, the greater the influence of that domain on the event. For example, if the event to be predicted is in the economic domain, the economic domain, as the target domain, has a greater influence on the event than non-target domains, such as 1 or 0.8. The political domain has a moderate influence on the event, such as 0.5. The sports domain has a low influence on the event, such as 0.2. The relevance between different domains can be pre-calculated or set, and the influence of the domain to which the first domain knowledge belongs can be determined based on the relevance. Alternatively, the similarity between the first event information and keywords in different domains can be calculated, and the influence can be determined based on the similarity.

[0059] Given a defined level of influence, the first-domain knowledge from each domain is weighted and fused. For example, the first-domain knowledge can be encoded into a unified feature space to obtain the features corresponding to the first-domain knowledge. These features can then be weighted and fused according to their influence to obtain the second-domain knowledge, or the second-domain knowledge can be obtained through further decoding.

[0060] In one implementation, the first-domain knowledge and its corresponding influence in each domain can be formed into a vector or matrix, such as an n*4 matrix, where n represents the number of first-domain knowledge items, and 4 represents that each first-domain knowledge item has 4 dimensions, with triples being 3 dimensions and influence being the 4th dimension. This vector or matrix is ​​then used as the second-domain knowledge.

[0061] In one implementation, domain knowledge can be transferred between different specialized prediction models. This allows a model specializing in one domain to acquire and utilize knowledge from other domains, such as through a RESTful API (Representational State Transfer Application Programming Interface). This knowledge transfer enables different specialized prediction models to share knowledge from either a first domain or a second domain.

[0062] Step S230: Input the first event information and the second domain knowledge into multiple professional prediction models, and output the corresponding first development prediction result through each professional prediction model.

[0063] For example, information about the first event can be combined with knowledge from the second domain, and then input into each specialized prediction model. The specialized prediction model learns the essential characteristics of the event from the information about the first event and relevant knowledge from the second domain, which helps improve the accuracy of the initial development prediction results output by the specialized prediction model.

[0064] Step S130: If the consensus of multiple first development prediction results does not reach the consensus threshold, then generate the inference chain of the event to be predicted based on the prior information related to the first event information.

[0065] Each specialized prediction model outputs a primary development forecast, and there may be multiple conflicting primary development forecasts. The most prevalent forecast (i.e., the one that appears most frequently) can be identified, and its support level can be calculated as the consensus among the multiple primary development forecasts. The method for calculating the consensus level is shown below:

[0066]

[0067]

[0068] A higher degree of consensus indicates greater agreement among different professional prediction models, making the most mainstream prediction result more credible. Conversely, a lower degree of consensus indicates a less credible most mainstream prediction result, meaning the credibility of all primary development prediction results is low. The consensus threshold is a threshold set for the degree of consensus, used to measure whether the consensus of multiple primary development prediction results reaches the required level. This threshold can be set based on experience or specific business needs. In one implementation, if the consensus of multiple primary development prediction results reaches the consensus threshold, the most mainstream prediction result among them is taken as the final prediction result. If the consensus of multiple primary development prediction results does not reach the consensus threshold, it indicates that all primary development prediction results are unreliable, and a reasoning chain for the event to be predicted can be further generated.

[0069] Prior information can include pre-obtained information related to the first event, such as relevant knowledge and information about similar events. The inference chain includes information about the analysis and prediction process, making it easier for users to understand the basis of the computer's predictions.

[0070] In one implementation, the prior information includes a first historical event similar to the event to be predicted; the inference chain includes a first inference chain. Generating the inference chain for the event to be predicted based on the prior information related to the first event includes the following steps:

[0071] The first historical event is retrieved from the event memory based on the first event information.

[0072] The first reasoning chain is generated based on the development information of the first event corresponding to the first historical event.

[0073] The event memory is used to store information about events that have occurred (i.e., historical events), which may include basic event information, event development information (such as event development trend information, event development outcome information), etc. The following explains how to construct an event memory.

[0074] In one implementation, the event memory bank includes a short-term memory sub-bank, a medium-term memory sub-bank, and a long-term memory sub-bank. (See reference) Figure 3 As shown, the event development prediction method also includes the following steps S310 to S330:

[0075] Step S310: Determine short-term historical events whose time distance from the current event does not exceed the first time threshold, and store their event information in the short-term memory sub-database;

[0076] Step S320: Determine intermediate historical events whose time distance from the current event exceeds the first time threshold but does not exceed the second time threshold, filter out intermediate historical events whose importance index is greater than the first importance threshold, and store their event information in the intermediate memory sub-database;

[0077] Step S330: Determine long-term historical events whose time distance from the current event exceeds the second time threshold, filter out long-term historical events whose importance index is greater than the second importance threshold, and store their event information in the long-term memory sub-database.

[0078] The first time threshold is less than the second time threshold, and the first importance threshold is less than the second importance threshold. The first time threshold, second time threshold, first importance threshold, and second importance threshold can all be set based on experience or specific business needs. For example, the first time threshold could be 7 days, the second time threshold could be 3 months, the first importance threshold could be 0.3, and the second importance threshold could be 0.5.

[0079] Event time can be either the time the event occurred or the time the event ended. The event importance index is used to quantify the importance of the event. Historical events whose time is within the first time threshold (e.g., events within 7 days) are considered short-term historical events, and their information can be stored in a short-term memory sub-database. This sub-database can use a caching database, such as Redis, to ensure fast access to the latest data and support real-time analysis. Historical events whose time is within the first but not the second time threshold (e.g., events within 7 days to 3 months) are considered medium-term historical events. From these, events with an importance index greater than the first importance threshold are selected and their information stored in a medium-term memory sub-database. This sub-database can use a document database, such as MongoDB, offering flexible storage and query options at a moderate cost, suitable for large datasets with moderate access frequency. Historical events whose time is more than the second time threshold (e.g., events more than 3 months) are considered long-term historical events. From these, events with an importance index greater than the second importance threshold are selected and their information stored in a long-term memory sub-database. The long-term memory sub-database utilizes a graph database, such as Neo4j, which facilitates the discovery of relationships between long-term historical events, forming a systematic knowledge base and enabling in-depth analysis. In one implementation, the long-term memory sub-database can also be used to store a knowledge base.

[0080] Based on a three-tiered memory architecture (short-term, medium-term, and long-term), hierarchical management and dynamic updates of historical events are achieved. Cache databases, document databases, and graph databases are employed to address the characteristics and access requirements of different event levels, achieving a good balance between performance, cost, functionality, and scalability.

[0081] The intermediate memory and long-term memory sub-banks do not store information on all historical events. Instead, they select a subset of more important historical events based on their importance index and store their information there. The following section explains in detail how the importance index is determined.

[0082] In one implementation, reference Figure 4 As shown, the importance index of historical events can be determined through the following steps S410 to S440:

[0083] Step S410: Determine the impact of the historical event based on its dissemination scope, level of attention, and degree of reaction.

[0084] For example, the number of media reports and reposts of historical events can be quantified and standardized to a value between 0 and 1, representing the dissemination range. The search index of keywords related to historical events can be quantified and standardized to a value between 0 and 1, representing attention. The volume of social media discussions related to historical events can be quantified and standardized to a value between 0 and 1, representing attention and response. Weights can be assigned to each indicator based on experience or specific business needs; for example, the weight for dissemination range could be 0.3, attention 0.4, and response 0.3. The influence can then be calculated by weighting the dissemination range, attention, and response.

[0085] Step S420: Determine the scarcity of historical events based on their novelty and uniqueness.

[0086] For example, determining the similarity of a historical event to other existing events can be achieved using models like BERT to calculate semantic similarity. The novelty of the historical event is then determined based on this similarity; lower similarity indicates higher novelty. Alternatively, the highest k similarity scores between the historical event and other existing events can be obtained, and a negative correlation function can be used to calculate the novelty value based on the average and maximum of these k similarities. Uniqueness refers to the rarity of a historical event within its domain. Keywords from the historical event can be extracted, and uniqueness can be calculated based on their frequency within the domain. Weights for each indicator can be set based on experience or specific business needs; for example, novelty might have a weight of 0.6, and uniqueness a weight of 0.4. Finally, a weighted average of novelty and uniqueness can be used to calculate scarcity.

[0087] Step S430: Determine the degree of association of the historical event based on the number of other events associated with the historical event and the depth of association with the domain.

[0088] For example, in an event association network constructed from multiple events, the number of edges between a specific historical event and other events can be determined and standardized to a value in the range of 0 to 1, representing the number of other events associated with that historical event. The path length between a historical event and domain knowledge or important events in the domain can also be determined and standardized to a value in the range of 0 to 1, representing the association depth between that historical event and the domain. Weights can be set for each indicator based on experience or specific business needs; for example, the weight for the number of associated other events could be 0.5, and the weight for the association depth with the domain could also be 0.5. The association degree can then be calculated by weighting the number of associated other events and the association depth with the domain.

[0089] Step S440: Determine the importance index of historical events based on their influence, scarcity, and relevance.

[0090] For example, the weights for influence, scarcity, and relevance can be set based on experience or specific business needs. For instance, the weight for influence could be 0.5, scarcity 0.3, and relevance 0.2. The importance index is then calculated by weighting the influence, scarcity, and relevance of historical events.

[0091] based on Figure 4 The method considers three factors: the impact, scarcity, and relevance of historical events, and specifically considers the sub-indices under each factor. By combining these indicators, the importance index of historical events is calculated, ensuring the comprehensiveness and accuracy of the importance index.

[0092] The above details how to determine the importance index of historical events and how to construct an event memory bank. Based on the information of the first event, the first historical event is retrieved from the event memory bank. For example, the similarity between the information of the first event and the event information of each historical event in the event memory bank can be calculated, and one or more historical events with the highest similarity can be selected as the first historical event. A first inference chain is generated based on the development information of the first historical event. For example, the event information of the first historical event and the development information of the first event can be combined to form a binary tuple with a causal relationship, which serves as the first inference chain. The first inference chain can be an analogical inference chain based on similar historical events.

[0093] In one implementation, the prior information includes third-domain knowledge related to the event to be predicted; the inference chain includes a second inference chain. Generating the inference chain for the event to be predicted based on the prior information related to the first event information includes the following steps:

[0094] The first event information is matched with the rule information in the knowledge base, and the third domain knowledge is determined based on the matching results;

[0095] A second reasoning chain is generated based on knowledge from the third domain.

[0096] The third domain knowledge can be the same as or different from the first domain knowledge. The knowledge base can record rule information corresponding to the domain knowledge, representing the event conditions within the domain knowledge. When the rule information is satisfied, the development trend or result indicated in the domain knowledge can be generated. The first event information is matched with the rule information in the knowledge base. For example, the rule information can be a regular expression. The system checks whether the first event information satisfies the regular expression; if it does, the first event information matches the rule information. This determines the rule information satisfied by the first event information, and its corresponding domain knowledge is taken as the third domain knowledge. A second reasoning chain is generated based on the third domain knowledge. For example, the rule information, development trend, or development result in the third domain knowledge can be combined to form a binary tuple with a causal relationship, which serves as the second reasoning chain. The second reasoning chain can be a logical reasoning chain based on domain knowledge.

[0097] In one implementation, the prior information includes a second historical event of the same category as the event to be predicted; the inference chain includes a third inference chain. Generating the inference chain for the event to be predicted based on prior information related to the first event information includes the following steps:

[0098] Based on the category of the event to be predicted, search for a second historical event in the event memory bank;

[0099] The development information of the second event corresponding to the second historical event is statistically analyzed, and a third inference chain is generated based on the statistical results.

[0100] This process involves classifying the event to be predicted using an event classification model to determine its category, or determining its category based on entities, attribute values, etc., in the first event information. An event memory can store the categories of historical events. Historical events with the same category as the event to be predicted are identified as second historical events; typically, multiple second historical events can be identified. The development information of these second historical events is statistically analyzed, and a third inference chain is generated based on the statistical results. This third inference chain can be an inductive inference chain based on data statistics.

[0101] The following are examples of algorithms for generating the first, second, and third inference chains:

[0102]

[0103]

[0104] To further illustrate the three inference chains, consider the release of an economic policy as an example. To predict its subsequent impact on the economy, the following three inference chains can be generated: The first inference chain is based on analogical analysis of the historical impact of similar economic policies in other regions. The second inference chain is based on logical analysis using economic theory. The third inference chain is based on statistical analysis using current economic data.

[0105] Continue to refer to Figure 1 In step S140, the second development prediction result of the event to be predicted is determined according to the inference chain.

[0106] In this approach, a second development prediction result can be determined based on the reasoning result corresponding to the reasoning chain, and this result serves as the final prediction result. In one implementation, the second development prediction result can be determined based on the first development prediction result and the reasoning chain. For example, by comparing the first development prediction result with the reasoning result corresponding to the reasoning chain, determining the first development prediction result and the reasoning result with the same prediction direction, and then combining the two, the second development prediction result is obtained.

[0107] In one implementation, the event development prediction method further includes the following steps:

[0108] The third historical event is found based on the reasoning result corresponding to the reasoning chain. The development information of the third historical event is the opposite of the reasoning result.

[0109] Reverse challenge information is generated based on the third historical event to form a chain of reasoning.

[0110] Accordingly, the second development prediction result for determining the event to be predicted based on the inference chain includes the following steps:

[0111] The second development prediction result is determined based on the reasoning chain and the reverse questioning information.

[0112] For example, information on a third event development that contradicts the inference result and the corresponding third historical event can be retrieved from the event memory. Alternatively, the event to be predicted can be added to an event association network, and nodes with a reverse correlation to the event to be predicted can be found in the event association network to determine the third historical event. Based on the third historical event, reverse challenge information is generated for the inference chain. This reverse challenge information is information that contradicts the conclusion of the inference chain and can challenge the inference chain. Based on the inference chain and the reverse challenge information, the second development prediction result is determined by combining the positive and negative information. This ensures that the second development prediction result has undergone reverse challenge, further increasing its reliability.

[0113] In one implementation, determining the second development prediction result based on the reasoning chain and reverse questioning information includes the following steps:

[0114] Assess the first degree of credibility of the reasoning chain;

[0115] Determine the contradiction index of the reasoning result based on the information of counter-questioning;

[0116] The second credibility of the reasoning result is determined based on the first credibility and the contradiction index, and the second development prediction result is determined based on the second credibility.

[0117] The first level of credibility can be assessed based on indicators such as the completeness of the reasoning chain, data support, and universality. The contradiction index of the reasoning result can be determined based on indicators such as the completeness of the counter-questioning information, data support, and universality. The higher the evaluation of the counter-questioning information, the higher the contradiction index of the reasoning result, indicating a greater degree of doubt caused by the counter-questioning information. Combining the first level of credibility and the contradiction index, the second level of credibility of the reasoning result is calculated. This can be achieved by subtracting the contradiction index from the first level of credibility, or by fusing the first level of credibility and the contradiction index, and this second level of credibility is used as the final credibility. The reasoning result with the highest second level of credibility is selected as the second development prediction result, or multiple reasoning results can be fused based on the second level of credibility to obtain the second development prediction result.

[0118] In addition, multiple inference results can be weighted and fused based on the first credibility and the contradiction index, and reverse questioning information can also be fused to obtain the second development prediction result.

[0119] In one implementation, the event development prediction method further includes the following steps:

[0120] Multiple historical events are acquired, each historical event is used as a node, and edges are generated between different nodes based on the relationships between different historical events to construct an event association network; the relationships include positive association, negative association, causal relationship, temporal association, and domain association.

[0121] Event association networks can intuitively display the complex relationships between events, making them ideal for correlation and in-depth analysis, as well as for finding similar and related events. For example, by tracing the evolution of nodes, the entire process of event development can be obtained, helping the system discover patterns and trends in event development. For instance, the Neo4j graph database can be used to store and manage event association networks, supporting real-time updates and facilitating quick querying and analysis.

[0122] In one implementation, after adding the event to be predicted to the event association network, the event development prediction method further includes the following steps:

[0123] Monitor the traffic of the nodes corresponding to the event to be predicted in the event association network, as well as the traffic of the associated edges;

[0124] Determine whether the event to be predicted has experienced abnormal fluctuations based on the flow of nodes or edges.

[0125] The flow of a node or edge can be the sum of data such as the number of news reports or social media discussions related to the node or edge within a unit of time. The mean and variance of the flow within the time window are calculated using a sliding window algorithm. When the flow exceeds "mean + 3 times variance", it is determined that the event to be predicted has experienced abnormal fluctuations, indicating that there may be an important event or potential risk, which can trigger an early warning mechanism and prompt corresponding measures to be taken.

[0126] Exemplary device

[0127] The event development prediction device in the embodiments of this disclosure will now be described. (See reference...) Figure 5 As shown, the event development prediction device 500 may include the following modules:

[0128] The first event information acquisition module 510 is configured to acquire the first event information of the event to be predicted.

[0129] The first prediction module 520 is configured to process the first event information using multiple professional prediction models corresponding to multiple fields, and obtain multiple first development prediction results of the event to be predicted.

[0130] The inference chain generation module 530 is configured to generate an inference chain for the event to be predicted based on prior information related to the first event information if the consensus of the plurality of first development prediction results does not reach a consensus threshold.

[0131] The second prediction module 540 is configured to determine a second development prediction result of the event to be predicted based on the inference chain.

[0132] In one implementation, the step of processing the first event information using multiple professional prediction models corresponding to multiple fields to obtain multiple first development prediction results for the event to be predicted includes:

[0133] Obtain first domain knowledge associated with the first event information from the knowledge base; the first domain knowledge includes knowledge of the target domain to which the event to be predicted belongs and knowledge of non-target domains;

[0134] By fusing knowledge from the first domain, knowledge from the second domain is obtained;

[0135] The first event information and the second domain knowledge are input into the multiple professional prediction models, and the corresponding first development prediction result is output through each professional prediction model.

[0136] In one implementation, the fusion of the first domain knowledge to obtain the second domain knowledge includes:

[0137] Determine the degree of influence of the domain to which the first domain knowledge belongs on the event to be predicted;

[0138] The knowledge in the first domain is weighted and fused according to the influence level to obtain the knowledge in the second domain.

[0139] In one implementation, the prior information includes a first historical event similar to the event to be predicted; the inference chain includes a first inference chain; generating the inference chain for the event to be predicted based on prior information related to the first event information includes:

[0140] The first historical event is retrieved from the event memory based on the first event information;

[0141] The first inference chain is generated based on the development information of the first event corresponding to the first historical event.

[0142] In one implementation, the prior information includes third-domain knowledge related to the event to be predicted; the inference chain includes a second inference chain; generating the inference chain for the event to be predicted based on the prior information related to the first event information includes:

[0143] The first event information is matched with the rule information in the knowledge base, and the third domain knowledge is determined based on the matching result;

[0144] The second reasoning chain is generated based on the knowledge in the third domain.

[0145] In one implementation, the prior information includes a second historical event of the same category as the event to be predicted; the inference chain includes a third inference chain; generating the inference chain for the event to be predicted based on prior information related to the first event information includes:

[0146] Based on the category of the event to be predicted, the second historical event is retrieved from the event memory bank;

[0147] The development information of the second event corresponding to the second historical event is statistically analyzed, and the third inference chain is generated based on the statistical results.

[0148] In one embodiment, the event memory includes a short-term memory sub-repository, a medium-term memory sub-repository, and a long-term memory sub-repository; the device is further configured to:

[0149] Identify short-term historical events whose time distance from the current time does not exceed a first time threshold, and store their event information in the short-term memory sub-database;

[0150] Identify intermediate historical events whose time distance from the current event exceeds a first time threshold but does not exceed a second time threshold, filter out intermediate historical events whose importance index is greater than the first importance threshold, and store their event information in the intermediate memory sub-database;

[0151] Determine long-term historical events whose time distance from the present exceeds a second time threshold, filter out long-term historical events whose importance index is greater than the second importance threshold, and store their event information in the long-term memory sub-database;

[0152] Wherein, the first time threshold is less than the second time threshold, and the first importance threshold is less than the second importance threshold.

[0153] In one embodiment, the device is further configured to:

[0154] The influence of a historical event is determined based on its scope of dissemination, level of attention, and degree of impact.

[0155] The scarcity of a historical event is determined based on its novelty and uniqueness.

[0156] The degree of relevance of the historical event is determined based on the number of other events associated with it and the depth of its association with the domain.

[0157] The importance index of the historical event is determined based on its influence, scarcity, and relevance.

[0158] In one implementation, the short-term memory sub-database uses a cache database, the medium-term memory sub-database uses a document database, and the long-term memory sub-database uses a graph database.

[0159] In one embodiment, the device is further configured to:

[0160] The third historical event is found based on the reasoning result corresponding to the reasoning chain, and the development information of the third historical event is the opposite of the reasoning result.

[0161] Based on the third historical event, reverse challenge information is generated for the reasoning chain;

[0162] The step of determining the second development prediction result of the event to be predicted based on the inference chain includes:

[0163] The second development prediction result is determined based on the inference chain and the reverse questioning information.

[0164] In one implementation, determining the second development prediction result based on the reasoning chain and the reverse questioning information includes:

[0165] Evaluate the first credibility of the inference chain;

[0166] The contradiction index of the reasoning result is determined based on the reverse questioning information;

[0167] The second credibility of the reasoning result is determined based on the first credibility and the contradiction index, and the second development prediction result is determined based on the second credibility.

[0168] In one implementation, the step of searching for a third historical event based on the reasoning result corresponding to the reasoning chain includes:

[0169] Search the event memory for third event development information and corresponding third historical events that are contrary to the reasoning result; and / or, add the event to be predicted to the event association network, and search the event association network for nodes that have a reverse association relationship with the event to be predicted, so as to determine the third historical event.

[0170] In one embodiment, the device is further configured to:

[0171] Multiple historical events are acquired, each historical event is used as a node, and edges are generated between different nodes based on the relationships between different historical events to construct the event association network; the relationships include positive association, negative association, causal relationship, temporal association, and domain association.

[0172] In one implementation, after adding the event to be predicted to the event association network, the device is further configured to:

[0173] Monitor the traffic of the nodes corresponding to the event to be predicted in the event association network and the traffic of the associated edges;

[0174] Whether the event to be predicted has experienced abnormal fluctuations is determined based on the flow of the node or the flow of the edge.

[0175] Furthermore, other specific details of the embodiments of this disclosure have been described in detail in the embodiments of the above methods, and will not be repeated here.

[0176] Exemplary program product

[0177] The computer program product in the embodiments of this disclosure will now be described. The computer program product includes a computer program that, when executed by a processor, implements the methods described above in this disclosure.

[0178] In one implementation, the computer program product can be a tangible product, such as a computer-readable storage medium storing a computer program. The readable storage medium can be based on electrical, magnetic, optical, electromagnetic, infrared, or other signals, and includes, but is not limited to: random access memory (RAM), read-only memory (ROM), magnetic tape, floppy disk, flash memory, hard disk drive (HDD), solid-state drive (SSD), etc. For example, the computer program product can be a non-volatile storage medium storing a computer program, such as read-only memory, NAND flash memory, etc.

[0179] In one implementation, the computer program product can be an intangible product. For example, the computer program product can be a virtual digital product, such as an executable file or installation package containing a computer program.

[0180] Computer program code can be written in one or more programming languages. Programming languages ​​include, but are not limited to, C, Java, and C++. Program code can execute entirely on the user's computing device, or partially on the user's computing device, or as a standalone software package, or partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, such as a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via an internet connection provided by a mobile network operator).

[0181] Computer programs can be carried or transmitted via signals such as electricity, magnetism, light, electromagnetic fields, and infrared radiation. Electronic devices can convert signals carrying computer programs into digital signals, thereby running the computer programs. When a computer program runs on an electronic device, its code is used to cause the electronic device to execute (more specifically, the processor of the electronic device to execute) the method steps of various embodiments of this disclosure, such as: Step S110, obtaining first event information of the event to be predicted. Step S120, processing the first event information using multiple professional prediction models corresponding to multiple fields to obtain multiple first development prediction results of the event to be predicted. Step S130, if the consensus of the multiple first development prediction results does not reach a consensus threshold, generating an inference chain for the event to be predicted based on prior information related to the first event information. Step S140, determining a second development prediction result for the event to be predicted based on the inference chain.

[0182] The above steps are executed by a computer program, utilizing multiple professional prediction models corresponding to various fields to process the first event information of the event to be dealt with, resulting in multiple first development prediction results. If the consensus level of these first development prediction results does not reach the consensus threshold, an inference chain is generated based on prior information related to the first event information. A second development prediction result is then determined based on the inference chain, serving as the final event development prediction result. On one hand, using professional prediction models from different fields to perform predictions and obtain multiple first development prediction results ensures the comprehensiveness of the prediction result. When the consensus level of the first development prediction result does not meet the requirements, an inference chain is generated and a second development prediction result is determined, improving the accuracy and credibility of the final result. On the other hand, the inference chain can reveal the information and logic upon which the prediction result is based, enhancing the interpretability of the prediction result.

[0183] Exemplary electronic devices

[0184] The electronic device described below is an embodiment of the present disclosure. The electronic device includes a processor and a memory, the memory storing executable instructions of the processor. The processor is configured to perform the methods described above by executing the executable instructions.

[0185] refer to Figure 6 An electronic device according to embodiments of the present disclosure will be described by way of example. Figure 6 The electronic device 600 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0186] like Figure 6 As shown, the electronic device 600 is presented in the form of a general-purpose computing device. The components of the electronic device 600 may include, but are not limited to: a processor 610, a memory 620, a bus 630 connecting different system components (including the memory 620 and the processor 610), an I / O (input / output) interface 640, and a network adapter 650.

[0187] The memory 620 stores program code, which can be executed by the processor 610 to perform the method steps of this embodiment, such as: Step S110, obtaining first event information of the event to be predicted; Step S120, processing the first event information using multiple professional prediction models corresponding to multiple fields to obtain multiple first development prediction results of the event to be predicted; Step S130, if the consensus of the multiple first development prediction results does not reach a consensus threshold, generating an inference chain for the event to be predicted based on prior information related to the first event information; Step S140, determining a second development prediction result for the event to be predicted based on the inference chain.

[0188] The processor 610 executes the above steps, using multiple professional prediction models corresponding to various fields to process the first event information of the event to be processed, obtaining multiple first development prediction results. If the consensus level of these first development prediction results does not reach the consensus threshold, an inference chain is generated based on prior information related to the first event information. A second development prediction result is then determined based on the inference chain, serving as the final event development prediction result. On one hand, using professional prediction models from different fields to perform predictions and obtain multiple first development prediction results ensures the comprehensiveness of the prediction result. When the consensus level of the first development prediction result does not meet the requirements, generating an inference chain and determining a second development prediction result improves the accuracy and credibility of the final result. On the other hand, the inference chain can reveal the information and logic upon which the prediction result is based, enhancing the interpretability of the prediction result.

[0189] Memory 620 may include volatile memory, such as random access memory (RAM) 621 and / or cache unit 622, and may also include non-volatile memory, such as read-only memory (ROM) 623. Memory 620 may also include one or more program modules 624, including but not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. For example, program module 624 may include the modules in the above-described device.

[0190] The processor 610 may include processing units such as an AP (Application Processor), a modem processor, a GPU, an ISP (Image Signal Processor), a controller, an encoder, a decoder, a DSP, a baseband processor, and / or an NPU (Neural-Network Processing Unit).

[0191] Bus 630 is used to connect different parts of electronic device 600, and may include data bus, address bus and control bus, etc.

[0192] Electronic device 600 can also communicate with one or more external devices 700 (such as keyboards, pointing devices, Bluetooth devices, etc.), and this communication can be carried out through I / O interface 640.

[0193] Electronic device 600 can also communicate with one or more networks via network adapter 650. For example, network adapter 650 can provide mobile communication solutions such as 3G / 4G / 5G, or wireless communication solutions such as wireless LAN, Bluetooth, and near-field communication. Network adapter 650 can communicate with other modules of electronic device 600 via bus 630.

[0194] Although not shown in the figure, other hardware and / or software modules may be installed on the electronic device 600, including but not limited to: display, audio unit, microcode, device driver, redundancy processing unit, external disk drive array, RAID (Redundant Arrays of Independent Disks) system, tape drive, and data backup storage system.

[0195] It should be noted that although several modules or sub-modules of the apparatus have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0196] Furthermore, although the operations of the methods disclosed herein are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all of the operations shown must be performed to achieve the desired result. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0197] While the spirit and principles of this disclosure have been described with reference to several specific embodiments, it should be understood that this disclosure is not limited to the disclosed specific embodiments, and the division of aspects does not imply that features in these aspects cannot be combined for benefit; such division is merely for convenience of expression. This disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A method for predicting the development of an event, characterized in that, The method includes: Obtain the first event information of the event to be predicted; The first event information is processed by multiple professional prediction models corresponding to multiple fields to obtain multiple first development prediction results of the event to be predicted. If the consensus of the multiple first development prediction results does not reach the consensus threshold, then the inference chain of the event to be predicted is generated based on the prior information related to the first event information. The second development prediction result of the event to be predicted is determined based on the inference chain.

2. The method according to claim 1, characterized in that, The process of using multiple professional prediction models corresponding to multiple fields to process the first event information to obtain multiple first development prediction results for the event to be predicted includes: Obtain first domain knowledge associated with the first event information from the knowledge base; the first domain knowledge includes knowledge of the target domain to which the event to be predicted belongs and knowledge of non-target domains; By fusing knowledge from the first domain, knowledge from the second domain is obtained; The first event information and the second domain knowledge are input into the multiple professional prediction models, and the corresponding first development prediction result is output through each professional prediction model.

3. The method according to claim 2, characterized in that, The fusion of knowledge from the first domain to obtain knowledge from the second domain includes: Determine the degree of influence of the domain to which the first domain knowledge belongs on the event to be predicted; The knowledge in the first domain is weighted and fused according to the influence level to obtain the knowledge in the second domain.

4. The method according to claim 1, characterized in that, The prior information includes a first historical event similar to the event to be predicted; the inference chain includes a first inference chain; generating the inference chain for the event to be predicted based on prior information related to the first event includes: The first historical event is retrieved from the event memory based on the first event information; The first inference chain is generated based on the development information of the first event corresponding to the first historical event.

5. The method according to claim 1, characterized in that, The prior information includes third-domain knowledge related to the event to be predicted; the inference chain includes a second inference chain; generating the inference chain for the event to be predicted based on the prior information related to the first event information includes: The first event information is matched with the rule information in the knowledge base, and the third domain knowledge is determined based on the matching result; The second reasoning chain is generated based on the knowledge in the third domain.

6. The method according to claim 1, characterized in that, The prior information includes a second historical event of the same category as the event to be predicted; the inference chain includes a third inference chain; generating the inference chain for the event to be predicted based on prior information related to the first event information includes: Based on the category of the event to be predicted, the second historical event is retrieved from the event memory bank; The development information of the second event corresponding to the second historical event is statistically analyzed, and the third inference chain is generated based on the statistical results.

7. The method according to claim 4 or 6, characterized in that, The event memory bank includes a short-term memory sub-bank, a medium-term memory sub-bank, and a long-term memory sub-bank; the method further includes: Identify short-term historical events whose time distance from the current event does not exceed a first time threshold, and store their event information in the short-term memory sub-database; Identify intermediate historical events whose time distance from the current event exceeds a first time threshold but does not exceed a second time threshold, filter out intermediate historical events whose importance index is greater than the first importance threshold, and store their event information in the intermediate memory sub-database; Determine long-term historical events whose time distance from the present exceeds a second time threshold, filter out long-term historical events whose importance index is greater than the second importance threshold, and store their event information in the long-term memory sub-database; Wherein, the first time threshold is less than the second time threshold, and the first importance threshold is less than the second importance threshold.

8. An event development prediction device, characterized in that, The device includes: The first event information acquisition module is configured to acquire the first event information of the event to be predicted. The first prediction module is configured to process the first event information using multiple professional prediction models corresponding to multiple fields, and obtain multiple first development prediction results of the event to be predicted. The inference chain generation module is configured to generate an inference chain for the event to be predicted based on prior information related to the first event information if the consensus of the plurality of first development prediction results does not reach a consensus threshold. The second prediction module is configured to determine a second development prediction result of the event to be predicted based on the inference chain.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 7.

10. An electronic device, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to execute the method of any one of claims 1 to 7 by executing the executable instructions.

Images