CN-122021612-A - Information processing method and system based on map sensing

CN122021612ACN 122021612 ACN122021612 ACN 122021612ACN-122021612-A

Abstract

The application provides an information processing method and system based on spectrum sensing, wherein the method comprises the steps of obtaining an original information text input by a user, creating a corresponding semantic spectrum based on the original information text, carrying out language identification and emotion analysis on the communication information text according to a preset emotion identification classifier, obtaining a corresponding language intensity score and an attitude polarity score, calculating a conflict probability score and a risk grade of the original information text according to the semantic spectrum, the language intensity score and the attitude polarity score, calling a language model to carry out expression optimization on the communication information text based on the conflict probability score and the risk grade, generating at least one optimization suggestion message, and updating the communication information text according to a selection instruction of the user on the optimization suggestion message. Therefore, by applying the technical scheme of the application, the contradiction and conflict generated in the communication of the user can be effectively avoided, and the communication and cooperation efficiency is improved.

Inventors

CHAI MENGZHU
XIONG YAMENG
CHEN YING

Assignees

广州工程技术职业学院

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. An information processing method based on spectrum sensing is characterized by comprising the following steps: Acquiring an original information text input by a user, wherein the original information text at least comprises a communication information text sent by the user, identity information of a corresponding sender and identity information of a corresponding receiver; creating a corresponding semantic graph based on the original information text, wherein the semantic graph corresponds to a communication session corresponding to the original information text, and the semantic graph at least comprises a corresponding sender node, a corresponding receiver node and a relation edge connecting the sender node and the receiver node; According to a preset emotion recognition classifier, performing language and emotion recognition and emotion analysis on the communication information text to obtain a corresponding language intensity score and attitude polarity score; calculating conflict probability scores and risk grades of the original information text according to the semantic graph, the mood intensity scores and the attitude polarity scores; based on the conflict probability score and the risk level, invoking a preset language model to perform expression optimization on the communication information text, and generating at least one optimization suggestion message; And responding to a selection instruction of the user for the optimization suggestion information, and updating the communication information text.
2. The information processing method based on spectrum sensing according to claim 1, wherein the step of acquiring the original information text input by the user comprises: Acquiring a communication information text written by a user in a current communication session through a preset communication interface; Based on a preset enterprise organization architecture interface, acquiring the identity information of a sender and the identity information of a receiver corresponding to the communication information text, wherein the identity information at least comprises a affiliated department and a corresponding job level; and recording the context environment parameters of the communication session, wherein the context environment parameters at least comprise a transmission channel and a timestamp.
3. The information processing method based on the spectrum sensing according to claim 2, wherein the step of creating the corresponding semantic spectrum based on the original information text includes: creating corresponding sender personnel nodes, corresponding receiver personnel nodes, corresponding department nodes of the sender and corresponding department nodes of the receiver according to the identity information of the sender and the identity information of the receiver; Constructing at least one relation edge between the sender personnel node and the receiver personnel node based on a history communication record in a preset time period, wherein the relation edge at least comprises one of a communication frequency edge for quantifying the history communication closeness, a superior-subordinate relation edge for identifying an organization hierarchy relation and a cooperation label edge for identifying an item cooperation history; And generating the semantic graph based on the personnel node, the department node and the relation edge.
4. The information processing method based on spectrum sensing according to claim 1, wherein the emotion recognition classifier is constructed based on a large language model; According to a preset emotion recognition classifier, performing language and emotion recognition and emotion analysis on the communication information text, and obtaining a corresponding language intensity score and attitude polarity score comprises the following steps: Inputting the communication information text to the emotion recognition classifier; Based on the output of the emotion recognition classifier, calculating and obtaining a mood intensity score and an attitude polarity score corresponding to the communication information text; the calculation formula of the mood intensity score is as follows: , Wherein ToneIntensity is a corresponding language intensity score, σ (·) is a preset sigmoid function for mapping a corresponding calculation result to a range from 0 to 1, s cls is a feature vector representing the whole semantics of the communication information text extracted from the emotion recognition classifier, s avg is a whole feature vector obtained by averaging all word vectors in the communication information text, w 1 is a preset first linear weight coefficient, w 2 is a preset second linear weight coefficient, and b is a preset bias term; the attitude polarity score is calculated and obtained by an attitude polarity score calculation formula, wherein the calculation mode of the attitude polarity score comprises the steps of obtaining the probability P positive of positive emotion and the probability P negative of negative emotion of the communication information text output by the emotion recognition classifier: , Wherein, polarity is the corresponding attitude Polarity score, and the value range is [ -1, +1].
5. The information processing method based on the atlas awareness according to claim 1, wherein the step of calculating the collision probability score and the risk level of the original information text based on the semantic atlas, the mood intensity score, and the attitude polarity score includes: Calculating and acquiring a map perception risk factor based on the relation side information in the semantic map; fusing the mood intensity score, the absolute value of the attitude polarity score and the map perception risk factor, and calculating and obtaining a conflict probability score of the original information text; and determining the corresponding risk level according to the preset threshold interval where the conflict probability score is located.
6. The information processing method based on spectrum sensing according to claim 5, wherein the collision probability score is calculated according to the following formula: , Wherein sigma (-) is a preset sigmoid function for mapping the corresponding calculation result to the interval of 0 to 1, P conflict is the corresponding collision probability score, For the score of the intensity of the mood, For the attitude polarity score, R ctx is the atlas-aware risk factor, and w 3 、w 4 and w 5 are respectively a preset first weight adjustment coefficient, a preset second weight adjustment coefficient, and a preset third weight adjustment coefficient.
7. The information processing method based on graph perception according to claim 1, wherein the step of calling a preset language model to perform expression optimization on the communication information text based on the collision probability score and the risk level, and generating at least one optimization suggestion information comprises: the communication information text, the conflict concept score, the risk level and the relation information extracted from the semantic graph are used as prompt information and are input into a preset language model, wherein the relation information is information represented by relation edges of the semantic graph; generating a plurality of candidate optimized texts based on the prompt information through the language model, wherein the language model is configured to perform at least one of replacing vocabulary in the communication information text, adjusting the arrangement sequence of vocabulary in the communication information text or adding language buffer term in the process of generating the candidate optimized texts; scoring and sorting the plurality of candidate optimized texts according to a preset optimization strategy, wherein the preset optimization strategy at least comprises semantic similarity with the communication information text; And according to the preset output quantity, outputting the candidate optimization texts with the highest ranking corresponding output quantity as the optimization suggestion information.
8. The information processing method based on spectrum sensing according to claim 7, wherein the step of scoring and ranking the plurality of candidate optimized texts according to a preset optimization strategy comprises: calculating a text optimization score for each candidate optimized text according to a text scoring formula, and sorting according to the text optimization scores: , Wherein S opt is a text optimization score of the candidate optimized text, S is the candidate optimized text, T user is the communication information text, sim (S, T user ) is semantic similarity between the candidate optimized text S and the communication information text T user , polarity (S) represents an attitude Polarity score of the candidate optimized text S, and Formality (S) represents a language score of the candidate optimized text S; the first harmonic parameter, the second harmonic parameter and the third harmonic parameter are preset respectively.
9. A method of processing information based on a spectrum sensing according to any one of claims 1 to 3, further comprising a step for updating a semantic spectrum, the step comprising: Acquiring behavior information of a user in a current communication session; and dynamically updating the attribute information of the relation edges in the semantic graph based on the behavior information.
10. An information processing system based on spectrum sensing, comprising: The system comprises an original information text acquisition unit, a storage unit and a storage unit, wherein the original information text acquisition unit is used for acquiring an original information text input by a user, and the original information text at least comprises a communication information text sent by the user, identity information of a corresponding sender and identity information of a corresponding receiver; A semantic graph creation unit, configured to create a corresponding semantic graph based on the original information text, where the semantic graph corresponds to a communication session corresponding to the original information text, and the semantic graph includes at least a corresponding sender node, a corresponding receiver node, and a relationship edge connecting the two; the mood analysis unit is used for carrying out mood recognition and mood analysis on the communication information text according to a preset mood recognition classifier to obtain a corresponding mood intensity score and an attitude polarity score; The conflict risk analysis unit is used for calculating a conflict probability score and a risk grade of the original information text according to the semantic graph, the mood intensity score and the attitude polarity score; The optimization suggestion information generation unit is used for calling a preset language model to perform expression optimization on the communication information text based on the conflict probability score and the risk level, and generating at least one optimization suggestion information; And the information text updating unit is used for responding to the selection instruction of the user for the optimization suggestion information and updating the communication information text.

Description

Information processing method and system based on map sensing Technical Field The application relates to the technical field of artificial intelligence, in particular to an information processing method and system based on spectrum sensing. Background With the digital popularization of job site communication, electronic mail and instant messaging tools have become core collaboration carriers. Under the background, users commonly have the inherent willingness to optimize self expression and improve the communication portability and effect, especially when facing complex scenes such as upper-level and lower-level communication, cross-department collaboration and the like. However, there is a significant gap between the support capability provided by the prior art and the fine optimization requirement of the user, so the prior art becomes a bottleneck for restricting the improvement of the communication efficiency. Existing intelligent authoring aids focus mostly on normative processing of language representations, such as grammar correction, spell checking, or text rendering based on fixed templates. However, the technology is essentially static correction to language form, but not dynamic optimization to communication effect, and the fundamental defect is that the dynamic context when communication occurs, including the key elements such as the organization relationship of the two communication parties, the historical interaction background or the current task state, cannot be understood. Thus, even if the tool provides modification suggestions, generalized templates that deviate from a particular context are often not able to achieve the precise optimization objective of "what is appropriate for a particular object under a particular relationship" for the user. Therefore, the technical support layer is lost, the communication efficiency is low, when the diagrammed expression of the system advice is inconsistent with the communication effect of the fine and agreeable situation expected to be achieved by the user, the user has to expend extra effort to judge, modify and even completely abandon the system advice, the willingness of the active optimization expression is consumed by invalid tool interaction, and worse, the incorrect advice can cause misunderstanding or negative emotion of an information receiver, so that conflict needing to be compensated for by subsequent communication is generated, and the cooperation efficiency is fundamentally reduced. Disclosure of Invention Based on the above, the application aims to provide an information processing method and system based on map sensing, which can effectively avoid contradiction and conflict generated in daily communication of users and improve communication efficiency and cooperation efficiency. The aim of the application can be achieved by the following technical scheme: An information processing method based on spectrum sensing comprises the following steps of obtaining an original information text input by a user, wherein the original information text at least comprises a communication information text sent by the user, identity information of a corresponding sender and identity information of a corresponding receiver, creating a corresponding semantic spectrum based on the original information text, wherein the semantic spectrum corresponds to a communication session corresponding to the original information text, the semantic spectrum at least comprises a corresponding sender node, a corresponding receiver node and a relation edge connecting the sender node and the receiver node, performing language identification and emotion analysis on the communication information text according to a preset emotion recognition classifier to obtain a corresponding language intensity score and an attitude polarity score, calculating a conflict probability score and a risk grade of the original information text according to the semantic spectrum, the language intensity score and the attitude polarity score, calling a preset language model to perform expression optimization on the communication information text based on the conflict probability score and the risk grade to generate at least one optimization suggestion information, and updating the communication suggestion information according to a selected communication instruction of the user. An information processing system based on spectrum sensing comprises an original information text acquisition unit, a semantic graph creation unit, an optimization suggestion information generation unit and an optimization suggestion information generation unit, wherein the original information text acquisition unit is used for acquiring an original information text input by a user, the original information text at least comprises a communication information text sent by the user, identity information of a corresponding sender and identity information of a corresponding receiver, the semantic graph creation unit is used for c