Intelligent scoring method and system for eloquence competition

Abstract

The present invention provides an intelligent scoring method and system for an eloquence competition, the method comprising: obtaining contestant speech data, using blockchain technology to store the contestant speech data in a blockchain, and using distributed storage technology to store the contestant speech data in multiple distributed storage nodes; obtaining multiple eloquence dimension indicators and weights of each eloquence dimension indicator, and automatically generating scoring rules according to the multiple eloquence dimension indicators and the weights of each eloquence dimension indicator; obtaining competition information and preset competition rules input by a user, using a smart contract programming language to convert the preset competition rules and scoring rules into contract codes, executing the contract codes to obtain smart contracts, and storing the smart contracts in a contract library and a blockchain; calling smart contracts in the blockchain to score the contestant speech data, and obtaining the contestant's eloquence competition score; realizing intelligent, standardized and personalized scoring of eloquence competitions, and improving the efficiency of scoring of eloquence competitions.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to an intelligent scoring method and system for an eloquence competition. Background Art

[0002] Eloquence competitions can improve personal speaking skills and on-the-spot reactions, but existing eloquence competition organizations mostly use manual scoring to score the eloquence speeches of contestants, which is largely influenced by subjective thinking. At the same time, existing eloquence competition organizations do not specifically score multiple eloquence dimensions such as content structure, speech emotions, and on-the-spot reactions, making it difficult to provide specific eloquence speech suggestions to contestants.

[0003] At present, the existing contract creation platform can help eloquence competition organizations create smart contracts for eloquence competitions, but this method requires eloquence competition organizations to have certain programming skills and the ability to understand the underlying logic. Summary of the invention

[0004] The present invention provides an intelligent scoring method and system for eloquence competitions, which score eloquence speeches from multiple modes, realize intelligent, standardized and personalized scoring of eloquence competitions, and improve the efficiency of eloquence competition scoring.

[0005] In order to solve the above technical problems, the present invention provides an intelligent scoring method for eloquence competition, comprising:

[0006] Acquire the contestant's speech data, use blockchain technology to store the contestant's speech data on the blockchain, and use distributed storage technology to store the contestant's speech data in multiple distributed storage nodes; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data, and the contestant's speech video data;

[0007] Acquire multiple eloquence dimension indicators and weights of each eloquence dimension indicator, and automatically generate scoring rules according to the multiple eloquence dimension indicators and weights of each eloquence dimension indicator;

[0008] Obtaining the competition information and preset competition rules input by the user, converting the preset competition rules and the scoring rules into contract codes using a smart contract programming language, executing the contract codes to obtain a smart contract, and storing the smart contract in a contract library and a blockchain;

[0009] The smart contract is called in the blockchain to score the contestant's speech data to obtain the contestant's eloquence competition score.

[0010] The present invention can process various types of contestant speech data to meet the scoring requirements of various eloquence competitions and improve the efficiency of eloquence competition scoring; automatically generate scoring rules according to various eloquence dimension indicators, intelligently adjust preset competition rules according to user input information, generate a comprehensive eloquence competition smart contract by combining the scoring rules and competition rules, and score the contestant speech data according to the smart contract. It can score eloquence speeches from multiple modes, thus realizing the intelligence, standardization and personalization of eloquence competitions.

[0011] Furthermore, the acquiring of the contestant's speech data, storing the contestant's speech data on a blockchain using blockchain technology, and storing the contestant's speech data in a plurality of distributed storage nodes using distributed storage technology, includes:

[0012] Acquire the contestant's speech data; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data;

[0013] Using an encryption algorithm to convert the contestant's speech data into encrypted speech data, and generating a digital signature of the contestant's speech data based on the encrypted speech data;

[0014] Using blockchain technology to verify the hash value or digital signature of the contestant's speech data, and after the hash value or digital signature of the contestant's speech data is successfully verified, storing the contestant's speech data on the blockchain;

[0015] The contestant's speech data and the encrypted speech data are sharded according to preset algorithm rules to obtain multiple shards of the contestant's speech data and multiple shards of the encrypted speech data, and the multiple shards of the contestant's speech data and the multiple shards of the encrypted speech data are stored in multiple distributed storage nodes using distributed storage technology.

[0016] The present invention uses distributed storage technology to encrypt and disperse the contestant's speech data, uses a data redundant storage strategy to ensure the availability and persistence of the contestant's speech data, and at the same time makes the data evenly distributed to improve the response speed of data calls; uses blockchain technology to store the contestant's speech data, which can monitor and trace the data flow during the eloquence competition to ensure the confidentiality and integrity of the contestant's speech data.

[0017] Furthermore, the obtaining of multiple eloquence dimension indicators and weights of each eloquence dimension indicator, and automatically generating scoring rules according to the multiple eloquence dimension indicators and weights of each eloquence dimension indicator, includes:

[0018] Obtaining multiple eloquence dimension indicators and weights of each eloquence dimension indicator; wherein the eloquence dimension indicators include grammatical structure indicators, sentiment analysis indicators, vocabulary richness indicators and speech speed indicators;

[0019] Scoring rules are automatically generated according to the multiple eloquence dimension indicators and the weights of the respective eloquence dimension indicators; wherein the scoring rules include: establishing scoring models for the multiple eloquence dimension indicators respectively to obtain multiple eloquence dimension scoring models, the multiple eloquence dimension scoring models include a grammatical structure scoring model, a sentiment analysis scoring model, a vocabulary richness scoring model and a speech speed scoring model; using natural language processing technology to extract features of the multiple eloquence dimension indicators from the contestant's speech data to obtain eloquence dimension features of the multiple contestant's speech data, the eloquence dimension features include structural features, sentiment features, vocabulary features and speech speed features; inputting the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring models corresponding to the eloquence dimension features respectively, and outputting the scores of the contestant's respective eloquence dimension indicators; calculating the total score of the contestant's first speech according to the weights of the respective eloquence dimension indicators and the scores of the respective eloquence dimension indicators of the contestant.

[0020] Furthermore, the eloquence dimension features of the contestant's speech data are respectively input into the eloquence dimension scoring model corresponding to the eloquence dimension features, and the scores of the contestants' eloquence dimension indicators are output, including:

[0021] The structural features of the contestant's speech data are input into the grammatical structure scoring model, and the contestant's grammatical structure score is calculated based on the syntax tree analysis method; wherein the grammatical structure score is specifically as follows:

[0022]

[0023] In the formula, represents the grammatical structure score of the contestant; g(xi) represents the grammatical structure score of the contestant's speech data;

[0024] The emotional features of the contestant's speech data are input into the emotional analysis scoring model, and the contestant's emotional analysis score is output; wherein the emotional analysis score is specifically as follows:

[0025]

[0026] In the formula, represents the sentiment analysis score of the contestant; h(xi) represents the sentiment analysis score of the contestant’s speech data;

[0027] The lexical features of the contestant's speech data are input into the lexical richness scoring model, the lexical importance score of the contestant's speech data is calculated using the TF-IDF method and the topic modeling method, the lexical diversity score of the contestant's speech data is calculated based on the lexical diversity index, and the lexical richness score of the contestant is calculated by combining the lexical importance score and the lexical diversity score; wherein the lexical richness score is specifically as follows:

[0028]

[0029] In the formula, represents the lexical richness score of the contestant; p(xi) represents the lexical richness score of the contestant's speech data;

[0030] The speech speed feature of the contestant's speech data is input into the speech speed scoring model, and the contestant's speech speed score is output; wherein the speech speed score is specifically as follows:

[0031]

[0032] In the formula, represents the contestant's speech speed score; q(xi) represents the speech rhythm score and speech fluency score of the contestant's speech data.

[0033] Furthermore, the calculation of the total score of the first speech of the contestant according to the weights of the various eloquence dimension indicators and the scores of the various eloquence dimension indicators of the contestant includes:

[0034] According to the weights of the eloquence dimension indicators and the scores of the eloquence dimension indicators of the contestant, the total score of the first speech of the contestant is calculated; wherein the total score of the first speech of the contestant is as follows:

[0035]

[0036] In the formula, S x Indicates the total score of the contestant's first speech; W j represents the weight of each eloquence dimension indicator, and j represents the number of eloquence dimension indicators; Indicates the eloquence dimension index y of the i-th paragraph of the contestant's speech data j The score on

[0037] Among them, the weight of each eloquence dimension indicator is W j The details are as follows:

[0038] W j =f(Δy i ,Δx)

[0039] Where Wj Indicates the weight of each eloquence dimension indicator; Δy i Represents the eloquence dimension index y j The degree of variation in the contestant's speech data, Δx, represents the complexity of the contestant's speech data.

[0040] The present invention is based on various contestants' speech characteristics such as structural characteristics, emotional characteristics, vocabulary characteristics and speech speed characteristics, uses multiple eloquence dimension scoring models to score various types of contestants' speech data, and calculates the total score of the contestant's first speech according to the weight of each eloquence dimension indicator, which can achieve scientific, comprehensive and multimodal feedback of eloquence competition scoring.

[0041] Furthermore, the obtaining of the competition information and the preset competition rules input by the user, converting the preset competition rules and the scoring rules into contract codes using a smart contract programming language, executing the contract codes to obtain a smart contract, and storing the smart contract in a contract library and a blockchain, includes:

[0042] Obtaining the competition information input by the user and the preset competition rules; wherein the competition information input by the user includes an audience participation factor, multiple speech dimensions and the weight of each speech dimension; the preset competition rules include: obtaining multiple speech dimensions, the speech dimensions including content richness, language expression ability, emotional appeal, reaction speed and interactivity dimensions, and structural organization and time control dimensions; using the audience participation factor to dynamically adjust the weight of each speech dimension to obtain the adjusted weight of each speech dimension; using natural language processing technology and machine learning technology to process the contestant's speech data to obtain the contestant's score for each speech dimension; combining the weight of each speech dimension and the contestant's score for each speech dimension, calculating the contestant's second speech total score;

[0043] The preset competition rules and the scoring rules are converted into contract codes using a smart contract programming language, the contract codes are executed to obtain a smart contract, and the smart contract is stored in a contract library and a blockchain.

[0044] Furthermore, the method of dynamically adjusting the weight of each speech dimension by using the audience participation factor to obtain the adjusted weight of each speech dimension includes:

[0045] The weights of various speech dimensions are dynamically adjusted using the audience participation factor to obtain the adjusted weights of various speech dimensions; wherein the adjusted weights of various speech dimensions are specifically as follows:

[0046]

[0047] In the formula, represents the weight of each speech dimension after adjustment; w X represents the weight of each speech dimension before adjustment, f() represents the nonlinear mapping function, V X represents the audience participation factor, X represents the various speech dimensions of the contestants, X∈{C,L,E,R,T}, C represents the content richness, L represents the language expression ability, E represents the emotional appeal, R represents the reaction speed and interactivity dimension, and T represents the structural organization and time control dimension.

[0048] Furthermore, the natural language processing technology and machine learning technology are used to process the contestant's speech data to obtain the contestant's scores in various speech dimensions, including:

[0049] Use text mining technology and NLP algorithms to analyze the contestants' speech data and obtain the contestants' content richness scores;

[0050] Use natural language processing technology to analyze contestants’ speech data and obtain contestants’ language expression ability scores;

[0051] Use sentiment analysis algorithms to analyze contestants’ speech data and obtain contestants’ emotional appeal scores;

[0052] Use machine learning models to analyze contestants’ speech data and obtain contestants’ scores on reaction speed and interactivity;

[0053] Natural language processing technology is used to analyze the contestants' speech data to obtain the contestants' scores on structural organization and time control dimensions.

[0054] Furthermore, the combining of the weights of the various speech dimensions and the scores of the contestants in the various speech dimensions to calculate the overall score of the contestant's second speech includes:

[0055] The total score of the contestant's second speech is calculated by combining the weights of each speech dimension and the contestant's score of each speech dimension; wherein the total score of the contestant's second speech is as follows:

[0056] S=ω C ·C+ω L ·L+ω E ·E+ω R ·R+ω T ·T

[0057] Where S represents the total score of the contestant’s second speech; ω C represents the weight coefficient of content richness, C represents the content richness score; ω L represents the weight coefficient of language expression ability, L represents the language expression ability score; ω Erepresents the weight coefficient of emotional appeal, E represents the emotional appeal score; ω R Represents the weight coefficient of the reaction speed and interactivity dimensions, R represents the reaction speed and interactivity dimension score; ω T It represents the weight coefficient of the structural organization and time control dimension, and T represents the score of the structural organization and time control dimension.

[0058] The present invention can adaptively adjust the weights of multiple speech dimensions and intelligently generate competition rules according to the competition information input by the user, and then generate a smart contract by combining the scoring rules and the competition rules that reflect the user's needs, and conduct multiple comprehensive evaluations of the contestants based on the total score of the first speech and the total score of the second speech, thereby improving the accuracy and scientificity of the eloquence competition scoring.

[0059] Furthermore, calling the smart contract in the blockchain to score the contestant's speech data and obtain the contestant's eloquence competition score includes:

[0060] Get user request;

[0061] Downloading the contestant's speech data from the blockchain and multiple distributed storage nodes according to the user's request;

[0062] Calling the smart contract from the blockchain according to the user request;

[0063] According to the smart contract, the contestant's speech data is scored in multiple stages of the eloquence competition to obtain a total eloquence competition score.

[0064] Furthermore, the contestant's speech data is scored in multiple stages of eloquence competition according to the smart contract to obtain a total eloquence competition score, including:

[0065] According to the smart contract, the contestant's speech data is scored in multiple stages of the eloquence competition to obtain a total eloquence competition score; wherein the total eloquence competition score is as follows:

[0066]

[0067] In the formula, S total Indicates the total score of the eloquence competition; S i represents the score of the eloquence competition at the i-th stage, n represents that there are n scoring stages in the eloquence competition, a i Represents the relative importance coefficient of the i-th stage.

[0068] The present invention integrates the eloquence competition scores of each stage in a weighted recursive manner based on the differences in scores at different stages, obtains the final eloquence competition total score, and realizes intelligent, standardized and personalized eloquence competition scores.

[0069] Furthermore, it also includes: using cloud storage technology to store training resources for eloquence competitions;

[0070] Get training resources for eloquence competitions;

[0071] By combining document storage technology, video media storage technology and cloud storage technology, the training resources of the eloquence competition are stored in a training resource database;

[0072] Use relational or non-relational databases to manage user information and perform user authentication;

[0073] After the user verification is successful, the training resources for the eloquence competition are obtained through the training resource database.

[0074] Based on the above method embodiment, the present invention provides a corresponding system embodiment, which provides an intelligent scoring system for eloquence competition, including: a speech data module, an eloquence dimension module, a smart contract module and a competition scoring module;

[0075] The speech data module is used to obtain the contestant's speech data, use blockchain technology to store the contestant's speech data on the blockchain, and use distributed storage technology to store the contestant's speech data in multiple distributed storage nodes; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data;

[0076] The eloquence dimension module is used to obtain multiple eloquence dimension indicators and the weights of each eloquence dimension indicator, and automatically generate scoring rules according to the multiple eloquence dimension indicators and the weights of each eloquence dimension indicator;

[0077] The smart contract module is used to obtain the competition information and preset competition rules input by the user, convert the preset competition rules and the scoring rules into contract codes using a smart contract programming language, execute the contract code to obtain a smart contract, and store the smart contract in a contract library and a blockchain;

[0078] The competition scoring module is used to call the smart contract in the blockchain to score the contestant's speech data and obtain the contestant's eloquence competition score.

[0079] Furthermore, the speech data module includes: a data acquisition submodule, a data encryption submodule, a data verification submodule and a shard storage submodule;

[0080] The data acquisition submodule is used to acquire the contestant's speech data; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data;

[0081] The data encryption submodule is used to convert the contestant's speech data into encrypted speech data using an encryption algorithm, and generate a digital signature of the contestant's speech data based on the encrypted speech data;

[0082] The data verification submodule is used to verify the hash value or digital signature of the contestant's speech data using blockchain technology, and after the hash value or digital signature of the contestant's speech data is successfully verified, the contestant's speech data is stored on the blockchain;

[0083] The shard storage submodule is used to shard the contestant's speech data and the encrypted speech data according to preset algorithm rules, obtain multiple shards of the contestant's speech data and multiple shards of the encrypted speech data, and use distributed storage technology to store the multiple shards of the contestant's speech data and the multiple shards of the encrypted speech data in multiple distributed storage nodes.

[0084] Furthermore, the eloquence dimension module includes: an eloquence dimension indicator submodule and an eloquence dimension score submodule;

[0085] The eloquence dimension index submodule is used to obtain multiple eloquence dimension indicators and the weight of each eloquence dimension indicator; wherein the eloquence dimension indicators include grammatical structure indicators, sentiment analysis indicators, vocabulary richness indicators and speech speed indicators;

[0086] The eloquence dimension scoring submodule is used to automatically generate scoring rules according to the multiple eloquence dimension indicators and the weights of the various eloquence dimension indicators; wherein the scoring rules include: establishing scoring models for the multiple eloquence dimension indicators respectively to obtain multiple eloquence dimension scoring models, and the multiple eloquence dimension scoring models include a grammatical structure scoring model, a sentiment analysis scoring model, a vocabulary richness scoring model and a speech speed scoring model; using natural language processing technology to extract features of the contestant's speech data regarding the multiple eloquence dimension indicators to obtain eloquence dimension features of the multiple contestant's speech data, and the eloquence dimension features include structural features, sentiment features, vocabulary features and speech speed features; inputting the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring models corresponding to the eloquence dimension features respectively, and outputting the scores of the contestant's various eloquence dimension indicators; and calculating the total score of the contestant's first speech according to the weights of the various eloquence dimension indicators and the scores of the various eloquence dimension indicators of the contestant.

[0087] Furthermore, the submodule for scoring the eloquence dimension includes: a scoring model unit;

[0088] The scoring model unit is used to input the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring model corresponding to the eloquence dimension features, and output the scores of each eloquence dimension indicator of the contestant; wherein the scoring model unit includes: a grammatical structure scoring model sub-unit, a sentiment analysis scoring model sub-unit, a vocabulary richness scoring model sub-unit and a speech speed scoring model sub-unit;

[0089] The grammatical structure scoring model subunit is used to input the structural features of the contestant's speech data into the grammatical structure scoring model, and calculate the contestant's grammatical structure score based on the syntax tree analysis method; wherein the grammatical structure score is specifically as follows:

[0090]

[0091] In the formula, represents the grammatical structure score of the contestant; g(xi) represents the grammatical structure score of the contestant's speech data;

[0092] The sentiment analysis scoring model subunit is used to input the sentiment features of the contestant's speech data into the sentiment analysis scoring model and output the contestant's sentiment analysis score; wherein the sentiment analysis score is specifically as follows:

[0093]

[0094] In the formula, represents the sentiment analysis score of the contestant; h(xi) represents the sentiment analysis score of the contestant’s speech data;

[0095] The lexical richness scoring model subunit is used to input the lexical features of the contestant's speech data into the lexical richness scoring model, calculate the lexical importance score of the contestant's speech data using the TF-IDF method and the topic modeling method, calculate the lexical diversity score of the contestant's speech data based on the lexical diversity index, and calculate the contestant's lexical richness score by combining the lexical importance score and the lexical diversity score; wherein the lexical richness score is specifically as follows:

[0096]

[0097] In the formula, represents the lexical richness score of the contestant; p(xi) represents the lexical richness score of the contestant's speech data;

[0098] The speech speed scoring model subunit is used to input the speech speed feature of the contestant's speech data into the speech speed scoring model, and output the contestant's speech speed score; wherein the speech speed score is specifically as follows:

[0099]

[0100] In the formula, represents the contestant's speech speed score; q(xi) represents the speech rhythm score and speech fluency score of the contestant's speech data.

[0101] Furthermore, the eloquence dimension scoring submodule includes: a first speech total scoring unit;

[0102] The first speech total scoring unit is used to calculate the contestant's first speech total score according to the weights of the eloquence dimension indicators and the contestant's eloquence dimension indicator scores; wherein the contestant's first speech total score is as follows:

[0103]

[0104] In the formula, S x Indicates the total score of the contestant's first speech; W j represents the weight of each eloquence dimension indicator, and j represents the number of eloquence dimension indicators; Indicates the eloquence dimension index y of the i-th paragraph of the contestant's speech data j The score on

[0105] Among them, the weight of each eloquence dimension indicator is W j The details are as follows:

[0106] W j =f(Δy i ,Δx)

[0107] Where W j Indicates the weight of each eloquence dimension indicator; Δy i Represents the eloquence dimension index y j The degree of variation in the contestant's speech data, Δx, represents the complexity of the contestant's speech data.

[0108] Furthermore, the smart contract module includes: a competition information submodule and a contract code submodule;

[0109] The competition information submodule is used to obtain the competition information input by the user and the preset competition rules; wherein the competition information input by the user includes audience participation factors, multiple speech dimensions and weights of each speech dimension; the preset competition rules include: obtaining multiple speech dimensions, the speech dimensions include content richness, language expression ability, emotional appeal, reaction speed and interactivity dimensions, and structural organization and time control dimensions; using the audience participation factor to dynamically adjust the weight of each speech dimension to obtain the adjusted weight of each speech dimension; using natural language processing technology and machine learning technology to process the contestant's speech data to obtain the contestant's scores for each speech dimension; combining the weights of each speech dimension and the contestant's scores for each speech dimension, calculating the contestant's total score for the second speech;

[0110] The contract code submodule is used to convert the preset competition rules and the scoring rules into contract codes using a smart contract programming language, execute the contract code to obtain a smart contract, and store the smart contract in a contract library and a blockchain.

[0111] Furthermore, the competition information submodule includes: a weight adjustment unit;

[0112] The weight adjustment unit is used to dynamically adjust the weight of each speech dimension using the audience participation factor to obtain the adjusted weight of each speech dimension; wherein the adjusted weight of each speech dimension is specifically as follows:

[0113]

[0114] In the formula, represents the weight of each speech dimension after adjustment; w X represents the weight of each speech dimension before adjustment, f() represents the nonlinear mapping function, V X represents the audience participation factor, X represents the various speech dimensions of the contestants, X∈{C,L,E,R,T}, C represents the content richness, L represents the language expression ability, E represents the emotional appeal, R represents the reaction speed and interactivity dimension, and T represents the structural organization and time control dimension.

[0115] Furthermore, the competition information submodule includes: a speech dimension scoring unit;

[0116] The speech dimension scoring unit is used to process the contestant's speech data using natural language processing technology and machine learning technology to obtain the contestant's scores for each speech dimension; wherein the speech dimension scoring unit includes: a first dimension sub-unit, a second dimension sub-unit, a third dimension sub-unit, a fourth dimension sub-unit and a fifth dimension sub-unit;

[0117] The first dimension sub-unit is used to analyze the contestant's speech data using text mining technology and NLP algorithm to obtain the contestant's content richness score;

[0118] The second dimension sub-unit is used to analyze the contestant's speech data using natural language processing technology to obtain the contestant's language expression ability score;

[0119] The third dimension sub-unit is used to analyze the contestant's speech data using a sentiment analysis algorithm to obtain the contestant's sentiment appeal score;

[0120] The fourth dimension sub-unit is used to analyze the contestant's speech data using a machine learning model to obtain the contestant's reaction speed and interactivity dimension scores;

[0121] The fifth dimension sub-unit is used to analyze the contestant's speech data using natural language processing technology to obtain the contestant's structural organization and time control dimension scores.

[0122] Further, the competition information submodule includes: a second speech total score unit;

[0123] The second speech total scoring unit is used to calculate the contestant's second speech total score by combining the weights of each speech dimension and the contestant's score of each speech dimension; wherein the contestant's second speech total score is specifically as follows:

[0124] S=ω C ·C+ω L ·L+ω E ·E+ω R ·R+ω T ·T

[0125] Where S represents the total score of the contestant’s second speech; ω C represents the weight coefficient of content richness, C represents the content richness score; ω L represents the weight coefficient of language expression ability, L represents the language expression ability score; ω E represents the weight coefficient of emotional appeal, E represents the emotional appeal score; ω R Represents the weight coefficient of the reaction speed and interactivity dimensions, R represents the reaction speed and interactivity dimension score; ω T It represents the weight coefficient of the structural organization and time control dimension, and T represents the score of the structural organization and time control dimension.

[0126] Furthermore, the competition scoring module includes: a user request acquisition submodule, a speech data calling submodule, a smart contract calling submodule and an eloquence competition scoring submodule;

[0127] The user request acquisition submodule is used to acquire user requests;

[0128] The speech data calling submodule is used to download the contestant's speech data from the blockchain and multiple distributed storage nodes according to the user request;

[0129] The smart contract calling submodule is used to call the smart contract from the blockchain according to the user request;

[0130] The eloquence competition scoring submodule is used to perform multiple stages of eloquence competition scoring on the contestant's speech data according to the smart contract to obtain a total eloquence competition score.

[0131] Furthermore, the eloquence competition scoring submodule is used to score the contestant's speech data at multiple stages of eloquence competition according to the smart contract to obtain a total eloquence competition score, including:

[0132] The eloquence competition scoring submodule is used to score the contestant's speech data in multiple stages of eloquence competition according to the smart contract to obtain a total score of the eloquence competition; wherein the total score of the eloquence competition is as follows:

[0133]

[0134] In the formula, S total Indicates the total score of the eloquence competition; S i represents the score of the eloquence competition at the i-th stage, n represents that there are n scoring stages in the eloquence competition, a i Represents the relative importance coefficient of the i-th stage.

[0135] Further, it also includes: training resource module;

[0136] The training resource module is used to store the training resources for the eloquence competition using cloud storage technology; wherein the training resource module includes: a resource acquisition submodule, a resource storage submodule, a user verification submodule and a resource download submodule;

[0137] The resource acquisition submodule is used to acquire training resources for eloquence competitions;

[0138] The resource storage submodule is used to store the training resources of the eloquence competition in a training resource database by combining document storage technology, video media storage technology and cloud storage technology;

[0139] The user verification submodule is used to manage user information and perform user verification using a relational database or a non-relational database;

[0140] The resource downloading submodule is used to obtain the training resources for the eloquence competition through the training resource database after the user verification is successful. BRIEF DESCRIPTION OF THE DRAWINGS

[0141] Figure 1 : A method flow chart of an embodiment of the intelligent scoring method for eloquence competition provided by the present invention;

[0142] Figure 2 : A system structure diagram of an embodiment of the intelligent scoring system for eloquence competition provided by the present invention. DETAILED DESCRIPTION

[0143] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0144] In the description of the present application, it should be understood that the terms "first", "second" and "third" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first", "second" and "third" may explicitly or implicitly include one or more of the features. In the description of the present application, unless otherwise specified, "several" means two or more.

[0145] The present invention provides an intelligent scoring method and system for eloquence competitions, which scores eloquence speeches from multiple modes, realizes intelligent, standardized and personalized scoring of eloquence competitions, and improves the efficiency of eloquence competition scoring.

[0146] Example 1

[0147] Based on the above requirements, an embodiment of the present invention provides an intelligent scoring method for eloquence competition, the method flow is as follows: Figure 1 As shown, the method includes steps S1 to S4, and each step is specifically as follows:

[0148] S1. Obtain the contestant's speech data, use blockchain technology to store the contestant's speech data on the blockchain, and use distributed storage technology to store the contestant's speech data in multiple distributed storage nodes; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data, and the contestant's speech video data. This step includes steps S1.1 to S1.4, and each step is as follows:

[0149] S1.1. Acquire the contestant's speech data; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data.

[0150] This embodiment needs to obtain the contestant's speech data input by the contestant. The contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data.

[0151] The data entered by the entrants include the following:

[0152] (1) Text input: Participants can submit their speeches or transcripts of their speeches in plain text. This method is particularly useful for scoring eloquence competitions because natural language processing technology can be easily applied to analyze and score them.

[0153] (2) Voice input: Participants can submit voice recordings or audio files of their speeches. This method is more suitable for evaluating voice expression ability in eloquence competitions, including pronunciation, speaking speed, intonation, etc.

[0154] (3) Video input: In some public speaking competitions, contestants may be required to submit a complete speech video so that the judges can fully evaluate the speech. This method can include the evaluation of both audio and visual elements.

[0155] (4) Other methods: Based on the specific requirements of the eloquence competition, other forms of speech content input may also be allowed, such as slide presentations, real-time speeches, etc.

[0156] S1.2. Use an encryption algorithm to convert the contestant's speech data into encrypted speech data, and generate a digital signature of the contestant's speech data based on the encrypted speech data.

[0157] This embodiment uses an encryption algorithm to perform end-to-end encryption on the contestant's speech data, converting the contestant's speech data into encrypted speech data. Only authorized users can decrypt the encrypted speech data and access the original contestant's speech data, thereby ensuring the security of the contestant's speech data during transmission and storage.

[0158] This embodiment uses digital signature technology to sign the encrypted speech data and generate a digital signature of the contestant's speech data. The digital signature of the contestant's speech data is generated based on the hash value of the contestant's speech data and the signer's private key, and the digital signature is used to verify the integrity of the contestant's speech data and authenticate the identity of the signer.

[0159] S1.3. Use blockchain technology to verify the hash value or digital signature of the contestant's speech data. After the hash value or digital signature of the contestant's speech data is successfully verified, the contestant's speech data is stored on the blockchain.

[0160] This embodiment records the hash value and digital signature of the contestant's speech data on the blockchain, and uses the immutability and transparency of the blockchain to verify the integrity of the stored data. This embodiment uses blockchain technology to verify the hash value or digital signature of the contestant's speech data, and after the hash value or digital signature of the contestant's speech data is successfully verified, the contestant's speech data is stored on the blockchain.

[0161] S1.4. Slice the contestant's speech data and the encrypted speech data according to preset algorithm rules to obtain multiple slicings of the contestant's speech data and multiple slicings of the encrypted speech data, and use distributed storage technology to store the multiple slicings of the contestant's speech data and the multiple slicings of the encrypted speech data in multiple distributed storage nodes.

[0162] Distributed storage nodes are distributed in different physical locations or cloud servers. Multiple distributed storage nodes build a multi-node storage architecture. In this embodiment, the number of distributed storage nodes can be set and expanded according to the performance requirements of the load equipment in actual conditions.

[0163] First, the contestant's speech data and the encrypted speech data are sharded according to a preset algorithm rule to obtain multiple shards of the contestant's speech data and multiple shards of the encrypted speech data. Then, using distributed storage technology (such as IPFS or Filecoin), the multiple shards of the contestant's speech data and the multiple shards of the encrypted speech data are stored in multiple distributed storage nodes, and the decentralized storage and sharing of data is achieved by allocating different data shards to different distributed storage nodes.

[0164] This embodiment adopts a data redundancy strategy, with redundancy between distributed storage nodes. The same data fragment may be stored on multiple distributed storage nodes to cope with the failure or damage of a single distributed storage node, so as to ensure the integrity, availability and persistence of the stored data. If a distributed storage node fails or data is damaged, this embodiment can automatically repair the data and restore the data from the backup to ensure the preservation of important speech data.

[0165] This embodiment can also monitor the load conditions of each distributed storage node to avoid overloading of a certain node, thereby ensuring uniform distribution of stored data, which is beneficial to improving the performance and response speed of the distributed storage system.

[0166] This embodiment also performs performance tests on the distributed storage and blockchain storage of data, as follows:

[0167] Data Integrity Verification:

[0168] Allow users to verify the integrity of speech content and ensure that data has not been tampered with during storage and transmission.

[0169] Use records on the blockchain to verify the authenticity and immutability of data.

[0170] Data access control testing:

[0171] Test access control and permission management mechanisms to ensure that only authorized users can access presentation content.

[0172] Simulate the access of different user roles to verify the effectiveness of permission management.

[0173] Distributed storage redundancy test:

[0174] Simulate storage node failures to verify the redundancy and data availability of distributed storage.

[0175] Ensure that the presentation content can still be accessed even if there is a node failure.

[0176] S2. Obtain multiple eloquence dimension indicators and the weights of each eloquence dimension indicator, and automatically generate scoring rules based on the multiple eloquence dimension indicators and the weights of each eloquence dimension indicator. This step includes steps S2.1 to S2.2, and each step is as follows:

[0177] S2.1. Obtain multiple eloquence dimension indicators and the weight of each eloquence dimension indicator; wherein the eloquence dimension indicators include a grammatical structure indicator, a sentiment analysis indicator, a vocabulary richness indicator and a speech speed indicator.

[0178] This embodiment needs to obtain multiple eloquence dimension indicators and the weights of each eloquence dimension indicator; wherein the eloquence dimension indicators include grammatical structure indicators, sentiment analysis indicators, vocabulary richness indicators and speech speed indicators.

[0179] In this embodiment, the eloquence dimension index may be a scoring standard and score correspondence table of the eloquence dimension index pre-defined by the jury or competition organizer in the eloquence competition. A specific example of an eloquence dimension index is as follows:

[0180] (1) Pronunciation and voice skills: 5 points: clear pronunciation, moderate speaking speed, fluent voice expression, no verbal errors. 4 points: basically clear pronunciation, moderate speaking speed, relatively fluent voice expression, with a few verbal errors. 3 points: average pronunciation, slightly fast or slow speaking speed, basically understandable voice expression, with verbal errors or hesitations. 2 points: unclear pronunciation, too fast or too slow speaking speed, voice expression needs to be improved, with obvious verbal errors or hesitations. 1 point: extremely unclear pronunciation, inappropriate speaking speed, incomprehensible voice expression, serious verbal errors.

[0181] (2) Content structure and logic: 5 points: The content of the speech has a clear structure and strict logic. It has an introduction, main body and conclusion. The viewpoints are clear and the supporting materials are sufficient. 4 points: The content of the speech has a relatively clear structure and good logic. It has an introduction, main body and conclusion. The viewpoints are clear and the supporting materials are sufficient. 3 points: The content of the speech has an average structure and logic. There may be some unclear structure or weak logic. The viewpoints are not clear enough and the supporting materials are limited. 2 points: The content of the speech has a relatively chaotic structure and weak logic. It lacks an introduction, main body and conclusion. The viewpoints are vague and the supporting materials are insufficient. 1 point: The content of the speech has an extremely chaotic structure and poor logic. It lacks a clear introduction, main body and conclusion. The viewpoints are incomprehensible and the supporting materials are insufficient.

[0182] Through eloquence dimension indicators and scoring criteria similar to the above examples, the jury or competition organizer can assign specific scores to each eloquence dimension and calculate the overall score of the contestant's speech based on these scores. This helps to ensure the objectivity and consistency of the scoring, thereby improving the fairness and credibility of the competition.

[0183] In this embodiment, the weights of the various eloquence dimension indicators are pre-assigned by the jury or competition organizer in the eloquence competition. These weights represent the relative importance of each indicator in the total score. For example, if fluency is more important for a particular competition, the organizer can assign a higher weight to it.

[0184] This embodiment provides an example of a comprehensive score calculation formula based on multiple eloquence dimensions, which is as follows:

[0185] Assume that there are N selected eloquence dimension indicators (D1, D2, ..., D N ) and the corresponding weights (W1, W2, ..., W N ); where the score of each indicator (S1, S2, ..., S N ) is assigned to each contestant's speech by the competition organizers and the judging panel. The score for each indicator should be between 0 and 1, with 1 indicating perfect performance and 0 indicating the worst performance. The weight of each indicator indicates its relative importance, and the sum of the weights should equal 1.

[0186] The calculation formula for the overall score is as follows:

[0187] Overall Score=(W1 S1+W2 S2+...+W N S N ) / (W1+W2+...+W N )

[0188] This formula multiplies the score of each eloquence dimension indicator by its corresponding weight, then adds the weighted scores of all indicators, and finally divides by the sum of all weights to obtain a composite score. The final composite score can be used to measure the overall quality of the speech.

[0189] In this embodiment, the above-mentioned comprehensive scoring calculation formula can be customized according to the needs of the eloquence competition, and different eloquence dimension indicators and weights can be selected to meet different types of competition rules and scoring standards. In this way, the fairness and accuracy of the eloquence competition can be improved, while ensuring the originality and personalization of the competition.

[0190] S2.2. Automatically generate scoring rules based on the multiple eloquence dimension indicators and the weights of the various eloquence dimension indicators; wherein the scoring rules include: establishing scoring models for the multiple eloquence dimension indicators respectively to obtain multiple eloquence dimension scoring models, the multiple eloquence dimension scoring models include a grammatical structure scoring model, a sentiment analysis scoring model, a vocabulary richness scoring model and a speech speed scoring model; using natural language processing technology to extract features of the multiple eloquence dimension indicators from the contestant's speech data to obtain eloquence dimension features of the multiple contestants' speech data, the eloquence dimension features include structural features, sentiment features, vocabulary features and speech speed features; inputting the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring models corresponding to the eloquence dimension features respectively, and outputting the scores of the contestants' various eloquence dimension indicators; calculating the total score of the contestant's first speech based on the weights of the various eloquence dimension indicators and the scores of the various eloquence dimension indicators of the contestants.

[0191] The specific implementation steps of the scoring rule include steps S2.2.1 to S2.2.4, each of which is as follows:

[0192] S2.2.1. Establish scoring models for the multiple eloquence dimension indicators respectively to obtain multiple eloquence dimension scoring models, wherein the multiple eloquence dimension scoring models include a grammatical structure scoring model, a sentiment analysis scoring model, a vocabulary richness scoring model and a speech speed scoring model.

[0193] This embodiment establishes a corresponding scoring model for each eloquence dimension indicator, and uses large-scale training data for training to obtain multiple eloquence dimension scoring models, which include a grammatical structure scoring model, a sentiment analysis scoring model, a vocabulary richness scoring model, and a speech speed scoring model.

[0194] S2.2.2. Use natural language processing technology to extract features of the multiple eloquence dimension indicators from the contestants' speech data to obtain eloquence dimension features of the multiple contestants' speech data, wherein the eloquence dimension features include structural features, emotional features, vocabulary features and speech speed features.

[0195] First, the contestants’ speech data were subjected to text preprocessing, including operations such as word segmentation, sentiment analysis, and syntactic analysis, to prepare the data for subsequent analysis.

[0196] Then, natural language processing technology is used to extract features of the multiple eloquence dimension indicators from the contestants' speech data to obtain eloquence dimension features of the multiple contestants' speech data, wherein the eloquence dimension features include structural features, emotional features, vocabulary features and speech speed features.

[0197] S2.2.3, respectively input the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring model corresponding to the eloquence dimension features, and output the scores of each eloquence dimension indicator of the contestant. This step includes steps (1) to (4), and each step is as follows:

[0198] (1) Inputting the structural features of the contestant's speech data into the grammatical structure scoring model, and calculating the contestant's grammatical structure score based on the syntax tree analysis method; wherein the grammatical structure score is specifically as follows:

[0199]

[0200] In the formula, represents the grammatical structure score of the contestant; g(xi) represents the grammatical structure score of the contestant's speech data.

[0201] (2) Inputting the emotional features of the contestant's speech data into the emotional analysis scoring model, and outputting the contestant's emotional analysis score; wherein the emotional analysis score is specifically as follows:

[0202]

[0203] In the formula, represents the sentiment analysis score of the contestant; h(xi) represents the sentiment analysis score of the contestant’s speech data.

[0204] (3) Inputting the lexical features of the contestant's speech data into the lexical richness scoring model, using the TF-IDF method and the topic modeling method to calculate the lexical importance score of the contestant's speech data, calculating the lexical diversity score of the contestant's speech data based on the lexical diversity index, and calculating the contestant's lexical richness score by combining the lexical importance score and the lexical diversity score; wherein the lexical richness score is specifically as follows:

[0205]

[0206] In the formula, represents the lexical richness score of the contestant; p(xi) represents the lexical richness score of the contestant’s speech data.

[0207] The importance of a word is usually related to the role and meaning of the word in the speech content. In this embodiment, the following method can be used to calculate the importance score of the word:

[0208] TF-IDF (Term Frequency-Inverse Document Frequency): The importance of each word is calculated using the TF-IDF method. TF represents the frequency of the word in the speech, while IDF represents the importance of the word in the entire corpus. The importance score of a word is equal to TF multiplied by IDF.

[0209] Topic Modeling: Use topic modeling methods to identify the topics and important words in the speech, such as Latent Dirichlet Allocation (LDA) or Non-Negative Matrix Factorization (NMF). Importance scores can be calculated based on the relevance of the words to the topic.

[0210] Sentiment Analysis: Use sentiment analysis to identify emotional words in a speech, such as positive, negative, or neutral, and assign word importance scores based on their sentiment strength.

[0211] Lexical diversity indicates how many different types of words the contestant used in his speech. In this embodiment, the lexical diversity score can be calculated using the following method:

[0212] Lexical diversity index: Use a lexical diversity index such as text-based richness index or information entropy to measure the diversity of vocabulary in the speech. Higher diversity scores indicate that the contestant used more diverse vocabulary.

[0213] Vocabulary frequency statistics: Count the frequency of each word in the speech and calculate the number of unique words. Contestants who use more unique words in their speeches can get higher vocabulary richness scores.

[0214] (4) Inputting the speech speed feature of the contestant's speech data into the speech speed scoring model, and outputting the contestant's speech speed score; wherein the speech speed score is specifically as follows:

[0215]

[0216] In the formula, represents the contestant's speech speed score; q(xi) represents the speech rhythm score and speech fluency score of the contestant's speech data.

[0217] The speech speed is usually related to the speech speed and rhythm. In this embodiment, the speech speed can be evaluated in the following ways:

[0218] Syllable Count: Count the syllables in the speech content to determine the speed of the speech. A faster syllable playback speed may cause the speech to sound rushed, while a slower speed may cause the speech to sound dragged out. You can use syllable count to measure speech speed.

[0219] Speed ​​standardization: Compare the speed of a speech to a certain standardized speed. The standardized speed can be adjusted according to the requirements of the oratory competition, usually at a medium speed, aiming to maintain the rhythm and fluency of the speech.

[0220] Natural Language Processing Tools: Use natural language processing tools to analyze the reading time of speech text to determine speech speed. This can include the average reading time of each text paragraph, as well as the total length of the entire speech.

[0221] In this embodiment, speech rhythm and speech fluency can be evaluated in the following ways:

[0222] Human scoring: A panel of judges or experts can watch a speech video or listen to a speech recording and evaluate the rhythm and fluency of the speech based on their subjective impressions and experience. They can consider whether the speech has appropriate pauses, changes in tone, and coherence of expression.

[0223] Natural language processing tools: Use natural language processing tools to analyze speech text for grammar and coherence. This can include checking for grammatical errors, reasonable sentence and paragraph structure, and logical and contextual coherence.

[0224] Audio analysis: Analyzes the audio of speech to identify signs of incoherent speech, interruptions in speech, or poor fluency. Audio analysis can detect changes in speech speed and audio breaks.

[0225] S2.2.4. Calculate the total score of the first speech of the contestant based on the weights of the various eloquence dimension indicators and the scores of the various eloquence dimension indicators of the contestant.

[0226] This embodiment dynamically adjusts the weight of each eloquence dimension indicator according to the change degree of the contestant's speech data and eloquence dimension indicator. Then, according to the weight of each eloquence dimension indicator and the score of each eloquence dimension indicator of the contestant, the total score of the contestant's first speech is calculated; wherein, the total score of the contestant's first speech is as follows:

[0227]

[0228] In the formula, S x Indicates the total score of the contestant's first speech; W j represents the weight of each eloquence dimension indicator, and j represents the number of eloquence dimension indicators; Indicates the eloquence dimension index y of the i-th paragraph of the contestant's speech data j The score on

[0229] Among them, the weight of each eloquence dimension indicator is W j The details are as follows:

[0230] W j =f(Δy i ,Δx)

[0231] Where W j Indicates the weight of each eloquence dimension indicator; Δy i Represents the eloquence dimension index y j The degree of variation in the contestant's speech data, Δx, represents the complexity of the contestant's speech data.

[0232] S3, obtaining the competition information and preset competition rules input by the user, converting the preset competition rules and the scoring rules into contract code using the smart contract programming language, executing the contract code to obtain a smart contract, and storing the smart contract in the contract library and blockchain. This step includes steps S3.1 to S3.2, and each step is as follows:

[0233] S3.1. Obtain the competition information input by the user and the preset competition rules; wherein the competition information input by the user includes an audience participation factor, multiple speech dimensions and the weight of each speech dimension; the preset competition rules include: obtaining multiple speech dimensions, the speech dimensions including content richness, language expression ability, emotional appeal, reaction speed and interactivity dimensions and structural organization and time control dimensions; using the audience participation factor to dynamically adjust the weight of each speech dimension to obtain the adjusted weight of each speech dimension; using natural language processing technology and machine learning technology to process the contestant's speech data to obtain the contestant's scores for each speech dimension; combining the weights of each speech dimension and the contestant's scores for each speech dimension to calculate the contestant's total score for the second speech.

[0234] The specific implementation steps of the preset competition rules include steps S3.1.1 to S3.1.4, each of which is as follows:

[0235] S3.1.1. Obtain multiple speech dimensions, including content richness, language expression ability, emotional appeal, response speed and interactivity, and structural organization and time control.

[0236] The preset competition rules in this embodiment are based on scoring on multiple speech dimensions, where the multiple speech dimensions include content richness C, language expression ability L, emotional appeal E, reaction speed and interactivity dimension R, and structural organization and time control dimension T.

[0237] S3.1.2. Use the audience participation factor to dynamically adjust the weight of each speech dimension to obtain the adjusted weight of each speech dimension.

[0238] In this embodiment, the audience participation factor represents the audience's recognition of a certain speech dimension, and may be a function of the number of votes or satisfaction scores.

[0239] The weights of various speech dimensions are dynamically adjusted using the audience participation factor to obtain the adjusted weights of various speech dimensions; wherein the adjusted weights of various speech dimensions are specifically as follows:

[0240]

[0241] In the formula, represents the weight of each speech dimension after adjustment; w X represents the weight of each speech dimension before adjustment, f() represents the nonlinear mapping function, V Xrepresents the audience participation factor, X represents the various speech dimensions of the contestants, X∈{C,L,E,R,T}, C represents the content richness, L represents the language expression ability, E represents the emotional appeal, R represents the reaction speed and interactivity dimension, and T represents the structural organization and time control dimension.

[0242] S3.1.3. Use natural language processing technology and machine learning technology to process the contestant's speech data and obtain the contestant's scores for each speech dimension. This step includes steps (1) to (5), and each step is as follows:

[0243] (1) Use text mining technology and NLP algorithms to analyze the contestants’ speech data and obtain the contestants’ content richness scores.

[0244] Content richness (C) is used to evaluate factors such as the number of key information in the speech content, the depth and breadth of information, factuality and opinion, etc. This embodiment uses text mining technology and NLP algorithms to analyze the contestant's speech text data to determine the richness of its content.

[0245] (2) Use natural language processing technology to analyze the contestants’ speech data and obtain the contestants’ language expression ability scores.

[0246] Language expression ability (L) is used to evaluate factors such as grammar, vocabulary, sentence structure, rhetoric, etc. of a speech to determine the quality and clarity of language expression. This embodiment uses a natural language processing tool to analyze the contestant's speech data to evaluate the contestant's language expression ability.

[0247] (3) Use sentiment analysis algorithms to analyze the contestants’ speech data and obtain the contestants’ emotional appeal scores.

[0248] Emotional appeal (E) is used to evaluate the emotional appeal of speech. This embodiment uses a sentiment analysis algorithm to identify the emotional color in the contestant's speech data, including the use of emotional vocabulary, emotional polarity and emotional intensity. The sentiment analysis algorithm can determine the emotional appeal of the speech.

[0249] (4) Use machine learning models to analyze the contestants’ speech data and obtain the contestants’ reaction speed and interactivity scores.

[0250] The reaction speed and interactivity dimension (R) is used to evaluate the dialogue and interactive elements in the speech, including the speed of responding to audience questions, the frequency and quality of interaction, etc. This embodiment uses a machine learning model to analyze the contestant's speech data to obtain the contestant's reaction speed and interactivity dimension score.

[0251] (5) Use natural language processing technology to analyze the contestants’ speech data and obtain the contestants’ scores on structural organization and time control dimensions.

[0252] The structural organization and time control dimension (T) is used to evaluate and analyze the structure, paragraph organization, logical coherence and time allocation of the speech. This embodiment uses NLP tools to help determine the control of structure and time in the contestant's speech data.

[0253] S3.1.4. Calculate the total score of the contestant's second speech by combining the weights of each speech dimension and the contestant's score of each speech dimension.

[0254] This embodiment combines the weights of each speech dimension and the standardized scores of each speech dimension of the contestant to calculate the total score of the contestant's second speech; wherein the total score of the contestant's second speech is as follows:

[0255] S=ω C ·C+ω L ·L+ω E ·E+ω R ·R+ω T ·T

[0256] Where S represents the total score of the contestant’s second speech; ω C represents the weight coefficient of content richness, C represents the content richness score; ω L represents the weight coefficient of language expression ability, L represents the language expression ability score; ω E represents the weight coefficient of emotional appeal, E represents the emotional appeal score; ω R Represents the weight coefficient of the reaction speed and interactivity dimensions, R represents the reaction speed and interactivity dimension score; ω T It represents the weight coefficient of the structural organization and time control dimension, and T represents the score of the structural organization and time control dimension.

[0257] S3.2. Use a smart contract programming language to convert the preset competition rules and the scoring rules into contract codes, execute the contract codes to obtain smart contracts, and store the smart contracts in a contract library and a blockchain.

[0258] This embodiment first uses advanced natural language processing libraries (including NLTK, spaCy, BERT, etc.) to perform text analysis and grammatical parsing on the scoring rules and competition rules, and converts the scoring rules and competition rules described by the user in natural language into the code logic of the smart contract. Then a customized smart contract generation engine is developed, and the preset competition rules and the scoring rules are converted into executable smart contract code using a smart contract programming language (such as Solidity, etc.) to ensure the correctness and consistency of the contract code. Finally, the smart contract is stored in the contract library and blockchain.

[0259] In this embodiment, the contract library also provides a series of pre-designed smart contract templates specifically formulated for different types of eloquence competitions, including templates for debates, speeches, eloquence competitions, etc. Each template has specific competition rules and logic. The specific use process of the above smart contract templates is as follows:

[0260] Template creation and selection:

[0261] Competition organizers or system administrators can access the Eloquence Competition Smart Contract Library and select existing Eloquence Competition Smart Contract templates in the contract library. These templates already contain common competition rules and eloquence dimension indicators. At the same time, competition organizers can also create custom Eloquence Competition Smart Contract templates according to specific needs.

[0262] Template customization:

[0263] If the competition organizer chooses to create a customized eloquence competition smart contract template, the competition rules, scoring criteria, participant qualifications, and eloquence dimension indicators can be customized in the contract library. These customizations can be personalized according to the specific requirements of the eloquence competition through interface operations or natural language input.

[0264] Eloquence dimension indicator settings:

[0265] In the eloquence competition smart contract template, competition organizers can freely select the included eloquence dimension indicators, and then set the weight for each eloquence dimension indicator to reflect its importance in the overall score. At the same time, competition organizers can also customize the specific calculation method and scoring range of the eloquence dimension indicators.

[0266] Comprehensive template:

[0267] After the template is created and customized, the eloquence competition smart contract library will store the created templates, which can cover different types of eloquence competitions, such as debates, speeches, eloquence competitions, etc.

[0268] Template selection and instantiation:

[0269] When the competition organizer is ready to create a new eloquence competition, he can select the most suitable template from the contract library according to the actual situation, and then instantiate the template to create a new eloquence competition smart contract instance, which includes the defined rules, scoring criteria and eloquence dimension indicators.

[0270] By using the above contract library, competition organizers can easily create smart contracts without writing complex codes. The contract library of this embodiment is developed with a user-friendly web interface, which allows competition organizers to customize smart contracts by filling out forms. Users can fill in the basic information and rules of the competition on the interface, and then the interface automatically translates the verbal description provided by the user into smart contract code with the help of natural language processing tools and a customized syntax analyzer, thereby automatically generating a smart contract that meets the needs of the eloquence competition.

[0271] By using the above contract library, users can use the programmability of smart contracts to develop special interactive contract templates. These special interactive contract templates can include timer contracts, question and answer contracts, interactive voting contracts, etc., and use the automatic execution capabilities of smart contracts to manage the interactive process in the competition. The specific implementation process of the above interactive contract template is as follows:

[0272] Select an interactive template:

[0273] In the contract library, contestants can choose to use interactive smart contract templates to increase the interactivity of the competition. These templates are usually created and deployed in advance by competition organizers or platform administrators. The selected templates may include different functions such as countdown timers, question and answer sessions, interactive voting, etc. Contestants can freely choose according to the type of competition and needs.

[0274] Configure template parameters:

[0275] If you choose a special interactive template, you need to configure the specific parameters of the template to meet the requirements of the competition. For example, if you choose the countdown timer template, you need to set the countdown duration and start time; if you choose the question-and-answer session, you need to define the content of the question and the answer format, etc.

[0276] Interactive execution:

[0277] During the eloquence contest, the interactive smart contract template will automatically execute according to the configured parameters and rules. For example, if the countdown timer template is used, the countdown will be displayed on the screen and the contestants need to complete the speech within the specified time; if the question-and-answer session is used, the contestants and the audience can submit questions and answers at specific times, and the smart contract will automatically record and display these interactive contents.

[0278] Interaction results:

[0279] The interactive smart contract template will record and store all interactive results, including countdown time records, questions and answers, interactive voting results, etc. These results can be used for competition scoring and feedback

[0280] In this embodiment, the smart contract stored in the blockchain has the following functions:

[0281] Contract deployment:

[0282] Competition organizers or system administrators can deploy relevant rules of eloquence competitions described in natural language on the blockchain network, then convert the relevant rules into smart contract codes, and finally upload the smart contract codes to the blockchain, and allow the configuration of relevant parameters of the smart contract codes, such as permission management rules and basic information of the competition.

[0283] Speech content record:

[0284] After the contestant enters the contestant's speech data, the smart contract in the blockchain will calculate the hash value or digital signature of the contestant's speech data and is responsible for recording these hash values ​​or digital signatures in the corresponding data fields on the blockchain.

[0285] Data Validation:

[0286] The smart contract in the blockchain will verify the integrity and consistency of the hash value or digital signature of the contestant's speech data after it is entered. If the verification finds that the hash value or digital signature does not match the input contestant's speech data, the smart contract will reject the input data and may trigger corresponding alarms or processing measures.

[0287] Permission management:

[0288] The smart contract in the blockchain contains permission management rules that define the user permissions that can access the contestant speech data stored on the blockchain. The rules can be configured based on factors such as role, authentication, and competition stage, and the smart contract will restrict or authorize user access according to the rules.

[0289] Data Access:

[0290] Contestants, judges and other authorized users can access and read contestants’ speech data by connecting and interacting with smart contracts in the blockchain. Among them, smart contracts in the blockchain act as access interfaces for contestants’ speech data to ensure data security and transparency.

[0291] Event Logging and Auditing:

[0292] Smart contracts in the blockchain can record various events, such as data submission, access requests, permission changes, etc. The records of these events can be used to audit, trace and monitor the integrity and compliance of the competition process.

[0293] S4. Call the smart contract in the blockchain to score the contestant's speech data and obtain the contestant's eloquence competition score. This step includes steps S4.1 to S4.3, and each step is as follows:

[0294] S4.1. Obtain user request.

[0295] When a user needs to access and call contestant speech data and smart contracts, this embodiment will obtain the corresponding user request.

[0296] S4.2. Download the contestant's speech data from the blockchain and multiple distributed storage nodes according to the user's request.

[0297] In this embodiment, according to the user request, the integrity of the requested data and the authority of the requesting user are first verified on the blockchain, and after the verification is successful, the contestant's speech data is downloaded from the blockchain and multiple distributed storage nodes.

[0298] S4.3. Call the smart contract from the blockchain according to the user request.

[0299] S4.4. According to the smart contract, the contestant's speech data is scored in multiple stages of the eloquence competition to obtain a total score of the eloquence competition.

[0300] Based on the consideration of the difference in scores at different stages, this embodiment uses a recursive or integral formula to integrate the scores at each stage, and the final total score of the eloquence competition is the weighted average of the scores at each stage.

[0301] In this embodiment, the contestant's speech data is scored in multiple stages of eloquence competition according to the smart contract to obtain a total eloquence competition score; wherein the total eloquence competition score is as follows:

[0302]

[0303] In the formula, S total Indicates the total score of the eloquence competition; S i represents the score of the eloquence competition at the i-th stage, n represents that there are n scoring stages in the eloquence competition, a i Represents the relative importance coefficient of the i-th stage.

[0304] This embodiment also provides a specific example of user registration and identity authentication:

[0305] User Registration:

[0306] New users enter the registration page of the web interface and fill in the necessary registration information, usually including user name, email address, password, etc. After the user submits the registration information, the legitimacy and integrity of the information will be verified to ensure that the user has provided valid information.

[0307] Email Verification:

[0308] The user needs to verify the email address they provided, and the web interface will send an email containing a verification link to the email address provided by the user. Once the user receives the verification email, they click the link to complete the email verification to confirm that they own the email address they provided.

[0309] Password security:

[0310] The web interface requires users to create strong passwords to ensure the security of their accounts. Strong passwords usually include a combination of uppercase letters, lowercase letters, numbers, and special characters. This helps prevent unauthorized access and protect the privacy of user accounts.

[0311] Authentication:

[0312] The web interface can use different methods to authenticate the user, such as requiring the user to provide a scanned copy of an identity document or other relevant proof. The authentication method can be performed according to the specific requirements of the platform to ensure the legitimacy of the user's identity.

[0313] create Account:

[0314] When the user's registration information and identity verification are passed, their account will be created and can use the smart contract of the contract library.

[0315] This embodiment also provides a specific example of user personal information management:

[0316] User login:

[0317] Users first need to log in to the platform using their registered username and password, such as social media login using the OAuth protocol.

[0318] Visit the Personal Information page:

[0319] After a user successfully logs in, the user can access the personal information management page. Usually, this page is located in the settings or profile section of the user account.

[0320] View personal information:

[0321] Users can view currently saved personal information, including name, contact information, skill tags, personal profile, etc.

[0322] To edit your personal information:

[0323] If the user wishes to update or modify their personal information, they can select the Edit option. In the Edit option, you can change various aspects of your personal information such as name, email address, contact number, skill tags, etc.

[0324] Save the changes:

[0325] After editing personal information, the user needs to select the save changes option to ensure that the new information is submitted to the database.

[0326] Information update:

[0327] After the changes are saved, the user's personal information will be updated to reflect their latest information. This means that other users or related functions on the platform will be able to access and display the user's latest information.

[0328] The intelligent scoring method for an eloquence contest provided in this embodiment further includes: using cloud storage technology to store training resources for the eloquence contest, specifically:

[0329] Get training resources for public speaking competitions.

[0330] This embodiment establishes a library of rich training resources including tutorials, demonstration speeches, professional opinions, etc. to improve the user's eloquence and expression skills.

[0331] By combining document storage technology, video media storage technology and cloud storage technology (such as Amazon S3 or Google Cloud Storage), the training resources of the eloquence competition are stored in a training resource database.

[0332] Use relational or non-relational databases to manage user information and perform user authentication.

[0333] After the user is successfully authenticated, the training resources for the eloquence competition are obtained through the training resource database. The training resource database includes the following user functions:

[0334] Resource browsing: After logging in, users can browse the training resource library, which is usually located in the training or education section of the user account. The resource library includes various training resources such as tutorials, demonstration speeches, professional advice, etc.

[0335] Search and filter: Users can use the search function or filter options to find resources for a specific topic, skill level, or content type, which enables users to find the most relevant and useful training materials.

[0336] Resource selection: When users find a resource of interest, they can click the resource link to view detailed information.

[0337] Resource Viewing: Users can view or download training resources online, such as tutorial videos, demonstration speech texts, or professional advice documents, which enables users to learn and improve their public speaking skills.

[0338] Learn and practice: Users can learn and practice public speaking skills based on training resources, which may include watching demonstration speeches and imitating them, reading expert opinions and applying them to their own speeches.

[0339] Feedback and Assessment: During the learning and practice process, users can self-assess their progress against the training resources or seek feedback from other users or educators.

[0340] Training records: Training resource databases usually track users' training progress and learning history so that users can view their learning records and progress at any time.

[0341] This embodiment also provides a specific example of a user support and question-answering platform:

[0342] User login:

[0343] Users first need to log in and enter the user support and Q&A platform interface.

[0344] Questions and issue submissions:

[0345] On the User Support and Q&A platform, users can ask questions or submit queries. These questions can cover any topic related to eloquence competition, presentation skills, or platform usage. Users can write detailed descriptions of their questions to ensure that other users or administrators can understand the context and needs of the question.

[0346] Question categories and tags:

[0347] Users are usually required to select a category for their question or add relevant tags so that the question can be more easily found by other users or experts, which helps to organize and categorize the questions and answers on the platform.

[0348] Question posted:

[0349] After the question is submitted and completed, the user can choose to publish the question, making it visible to other users and administrators.

[0350] answer the questions:

[0351] Other users or experts can view the list of questions and choose to answer questions that they are interested in or have experience with. Users can also answer questions by providing text answers, linking to relevant resources, uploading files, etc.

[0352] Discussion and Interaction:

[0353] On user support and Q&A platforms, users can discuss and interact with each other under questions to further clarify questions or provide more insights. This interaction is usually public, so other users can also benefit from it.

[0354] Admin Support:

[0355] Platform administrators or experts are often also involved in answering questions and providing high-quality advice and support. This helps ensure that users receive trustworthy answers and help.

[0356] Issue Solving and Tagging:

[0357] When a question is answered or solved satisfactorily, users can mark it as "solved" so that other users know that the question has been solved. This helps organize questions and answers on the user support and Q&A platform.

[0358] Search and find:

[0359] On the user support and Q&A platform, users can use the search function to find previously asked questions and related answers to find information about eloquence competitions and speech skills.

[0360] This embodiment also provides a specific example of the effect verification of the intelligent scoring method for eloquence competition:

[0361] User activity and satisfaction: User activity is assessed by the number of user registrations, login frequency, and participation in community interactions. User feedback is collected regularly to understand their satisfaction with the platform and their suggestions for improvement.

[0362] Training resource utilization: Analyze users’ browsing, downloading, and collection of different training resources. Improve and update training content based on resource utilization data.

[0363] Community Interaction: Track community interactions, including the number of posts, replies, and likes. Verify community activity through increased social media sharing and online discussions.

[0364] The implementation of the above embodiments of the present invention has the following beneficial effects:

[0365] The intelligent scoring method for an eloquence competition provided by the present invention can process various types of contestant speech data to meet the scoring requirements of various eloquence competitions and improve the efficiency of eloquence competition scoring; automatically generate scoring rules according to various eloquence dimension indicators, intelligently adjust preset competition rules according to user input information, generate a comprehensive eloquence competition smart contract in combination with the scoring rules and competition rules, and score the contestant's speech data according to the smart contract. It can score eloquence speeches from multiple modes, thereby realizing the intelligence, standardization and personalization of eloquence competitions.

[0366] Example 2

[0367] Based on the content of the above embodiments, an embodiment of the present invention provides an intelligent scoring system for eloquence competition, including: speech data module 101, eloquence dimension module 102, intelligent contract module 103 and competition scoring module 104. The system structure is as follows Figure 2 shown.

[0368] The speech data module 101 is used to obtain the contestant's speech data, use the blockchain technology to store the contestant's speech data on the blockchain, and use the distributed storage technology to store the contestant's speech data in multiple distributed storage nodes; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data;

[0369] The eloquence dimension module 102 is used to obtain multiple eloquence dimension indicators and the weights of each eloquence dimension indicator, and automatically generate a scoring rule according to the multiple eloquence dimension indicators and the weights of each eloquence dimension indicator;

[0370] The smart contract module 103 is used to obtain the competition information and preset competition rules input by the user, convert the preset competition rules and the scoring rules into contract codes using a smart contract programming language, execute the contract codes to obtain a smart contract, and store the smart contract in a contract library and a blockchain;

[0371] The competition scoring module 104 is used to call the smart contract in the blockchain to score the contestant's speech data and obtain the contestant's eloquence competition score.

[0372] In a possible implementation, the speech data module 101 includes: a data acquisition submodule 201, a data encryption submodule 202, a data verification submodule 203 and a shard storage submodule 204;

[0373] The data acquisition submodule 201 is used to acquire the contestant's speech data; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data;

[0374] The data encryption submodule 202 is used to convert the contestant's speech data into encrypted speech data using an encryption algorithm, and generate a digital signature of the contestant's speech data based on the encrypted speech data;

[0375] The data verification submodule 203 is used to verify the hash value or digital signature of the contestant's speech data using blockchain technology, and after the hash value or digital signature of the contestant's speech data is successfully verified, the contestant's speech data is stored on the blockchain;

[0376] The shard storage submodule 204 is used to shard the contestant's speech data and the encrypted speech data according to preset algorithm rules, obtain multiple shards of the contestant's speech data and multiple shards of the encrypted speech data, and use distributed storage technology to store the multiple shards of the contestant's speech data and the multiple shards of the encrypted speech data in multiple distributed storage nodes.

[0377] In a possible implementation, the eloquence dimension module 102 includes: an eloquence dimension indicator submodule 301 and an eloquence dimension score submodule 302;

[0378] The eloquence dimension index submodule 301 is used to obtain multiple eloquence dimension indicators and the weight of each eloquence dimension indicator; wherein the eloquence dimension indicators include grammatical structure indicators, sentiment analysis indicators, vocabulary richness indicators and speech speed indicators;

[0379] The eloquence dimension scoring submodule 302 is used to automatically generate scoring rules according to the multiple eloquence dimension indicators and the weights of the various eloquence dimension indicators; wherein the scoring rules include: establishing scoring models for the multiple eloquence dimension indicators respectively to obtain multiple eloquence dimension scoring models, the multiple eloquence dimension scoring models include a grammatical structure scoring model, a sentiment analysis scoring model, a vocabulary richness scoring model and a speech speed scoring model; using natural language processing technology to extract features of the multiple eloquence dimension indicators from the contestant's speech data to obtain eloquence dimension features of the multiple contestant's speech data, the eloquence dimension features include structural features, sentiment features, vocabulary features and speech speed features; inputting the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring models corresponding to the eloquence dimension features respectively, and outputting the scores of the contestant's various eloquence dimension indicators; calculating the total score of the contestant's first speech according to the weights of the various eloquence dimension indicators and the scores of the various eloquence dimension indicators of the contestant.

[0380] Furthermore, the eloquence dimension scoring submodule 302 includes: a scoring model unit 401;

[0381] The scoring model unit 401 is used to input the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring model corresponding to the eloquence dimension features, and output the scores of each eloquence dimension index of the contestant; wherein the scoring model unit 401 includes: a grammatical structure scoring model subunit 501, a sentiment analysis scoring model subunit 502, a vocabulary richness scoring model subunit 503 and a speech speed scoring model subunit 504;

[0382] The grammatical structure scoring model subunit 501 is used to input the structural features of the contestant's speech data into the grammatical structure scoring model, and calculate the contestant's grammatical structure score based on the syntax tree analysis method; wherein the grammatical structure score is specifically as follows:

[0383]

[0384] In the formula, represents the grammatical structure score of the contestant; g(xi) represents the grammatical structure score of the contestant's speech data;

[0385] The sentiment analysis scoring model subunit 502 is used to input the sentiment features of the contestant's speech data into the sentiment analysis scoring model and output the contestant's sentiment analysis score; wherein the sentiment analysis score is specifically as follows:

[0386]

[0387] In the formula, represents the sentiment analysis score of the contestant; h(xi) represents the sentiment analysis score of the contestant’s speech data;

[0388] The lexical richness scoring model subunit 503 is used to input the lexical features of the contestant's speech data into the lexical richness scoring model, calculate the lexical importance score of the contestant's speech data using the TF-IDF method and the topic modeling method, calculate the lexical diversity score of the contestant's speech data based on the lexical diversity index, and calculate the contestant's lexical richness score by combining the lexical importance score and the lexical diversity score; wherein the lexical richness score is specifically as follows:

[0389]

[0390] In the formula, represents the lexical richness score of the contestant; p(xi) represents the lexical richness score of the contestant's speech data;

[0391] The speech speed scoring model subunit 504 is used to input the speech speed feature of the contestant's speech data into the speech speed scoring model, and output the contestant's speech speed score; wherein the speech speed score is specifically as follows:

[0392]

[0393] In the formula, represents the contestant's speech speed score; q(xi) represents the speech rhythm score and speech fluency score of the contestant's speech data.

[0394] Furthermore, the eloquence dimension scoring submodule 302 includes: a first speech overall scoring unit 402;

[0395] The first speech total score unit 402 is used to calculate the contestant's first speech total score according to the weights of the eloquence dimension indicators and the contestant's eloquence dimension indicator scores; wherein the contestant's first speech total score is as follows:

[0396]

[0397] In the formula, S x Indicates the total score of the contestant's first speech; W j represents the weight of each eloquence dimension indicator, and j represents the number of eloquence dimension indicators; Indicates the eloquence dimension index y of the i-th paragraph of the contestant's speech data j The score on

[0398] Among them, the weight of each eloquence dimension indicator is W j The details are as follows:

[0399] W j =f(Δy i ,Δx)

[0400] Where W j Indicates the weight of each eloquence dimension indicator; Δy i Represents the eloquence dimension index y j The degree of variation in the contestant's speech data, Δx, represents the complexity of the contestant's speech data.

[0401] In a possible implementation, the smart contract module 103 includes: a competition information submodule 601 and a contract code submodule 602;

[0402] The competition information submodule 601 is used to obtain the competition information input by the user and the preset competition rules; wherein the competition information input by the user includes audience participation factors, multiple speech dimensions and weights of each speech dimension; the preset competition rules include: obtaining multiple speech dimensions, the speech dimensions including content richness, language expression ability, emotional appeal, reaction speed and interactivity dimensions, and structural organization and time control dimensions; using the audience participation factor to dynamically adjust the weights of each speech dimension to obtain the adjusted weights of each speech dimension; using natural language processing technology and machine learning technology to process the contestant's speech data to obtain the contestant's scores for each speech dimension; combining the weights of each speech dimension and the contestant's scores for each speech dimension to calculate the contestant's total score for the second speech;

[0403] The contract code submodule 602 is used to convert the preset competition rules and the scoring rules into contract codes using a smart contract programming language, execute the contract codes to obtain a smart contract, and store the smart contract in a contract library and a blockchain.

[0404] Furthermore, the competition information submodule 601 includes: a weight adjustment unit 701;

[0405] The weight adjustment unit 701 is used to dynamically adjust the weight of each speech dimension using the audience participation factor to obtain the adjusted weight of each speech dimension; wherein the adjusted weight of each speech dimension is as follows:

[0406]

[0407] In the formula, represents the weight of each speech dimension after adjustment; w X represents the weight of each speech dimension before adjustment, f() represents the nonlinear mapping function, V X represents the audience participation factor, X represents the various speech dimensions of the contestants, X∈{C,L,E,R,T}, C represents the content richness, L represents the language expression ability, E represents the emotional appeal, R represents the reaction speed and interactivity dimension, and T represents the structural organization and time control dimension.

[0408] Furthermore, the competition information submodule 601 includes: a speech dimension scoring unit 702;

[0409] The speech dimension scoring unit 702 is used to process the contestant's speech data using natural language processing technology and machine learning technology to obtain the contestant's scores in various speech dimensions; wherein the speech dimension scoring unit 702 includes: a first dimension sub-unit 801, a second dimension sub-unit 802, a third dimension sub-unit 803, a fourth dimension sub-unit 804 and a fifth dimension sub-unit 805;

[0410] The first dimension subunit 801 is used to analyze the contestant's speech data using text mining technology and NLP algorithm to obtain the contestant's content richness score;

[0411] The second dimension subunit 802 is used to analyze the contestant's speech data using natural language processing technology to obtain the contestant's language expression ability score;

[0412] The third dimension subunit 803 is used to analyze the contestant's speech data using a sentiment analysis algorithm to obtain the contestant's sentiment appeal score;

[0413] The fourth dimension subunit 804 is used to analyze the contestant's speech data using a machine learning model to obtain the contestant's reaction speed and interactivity dimension scores;

[0414] The fifth dimension subunit 805 is used to analyze the contestant's speech data using natural language processing technology to obtain the contestant's structural organization and time control dimension scores.

[0415] Furthermore, the competition information submodule 601 includes: a second speech overall scoring unit 703;

[0416] The second speech total score unit 703 is used to calculate the contestant's second speech total score by combining the weights of each speech dimension and the contestant's score of each speech dimension; wherein the contestant's second speech total score is specifically as follows:

[0417] S=ω C ·C+ω L ·L+ω E ·E+ω R ·R+ω T ·T

[0418] Where S represents the total score of the contestant’s second speech; ω C represents the weight coefficient of content richness, C represents the content richness score; ω L represents the weight coefficient of language expression ability, L represents the language expression ability score; ω E represents the weight coefficient of emotional appeal, E represents the emotional appeal score; ω RRepresents the weight coefficient of the reaction speed and interactivity dimensions, R represents the reaction speed and interactivity dimension score; ω T It represents the weight coefficient of the structural organization and time control dimension, and T represents the score of the structural organization and time control dimension.

[0419] In a possible implementation, the competition scoring module 104 includes: a user request acquisition submodule 901, a speech data calling submodule 902, a smart contract calling submodule 903 and an eloquence competition scoring submodule 904;

[0420] The user request acquisition submodule 901 is used to acquire a user request;

[0421] The speech data calling submodule 902 is used to download the contestant's speech data from the blockchain and multiple distributed storage nodes according to the user request;

[0422] The smart contract calling submodule 903 is used to call the smart contract from the blockchain according to the user request;

[0423] The eloquence competition scoring submodule 904 is used to perform multiple stages of eloquence competition scoring on the contestant's speech data according to the smart contract to obtain a total eloquence competition score.

[0424] Furthermore, the eloquence competition scoring submodule 904 is used to score the eloquence competition of the contestant's speech data at multiple stages according to the smart contract to obtain the total score of the eloquence competition, including:

[0425] The eloquence competition scoring submodule 904 is used to score the contestant's speech data in multiple stages of eloquence competition according to the smart contract to obtain a total eloquence competition score; wherein the total eloquence competition score is as follows:

[0426]

[0427] In the formula, S total Indicates the total score of the eloquence competition; S i represents the score of the eloquence competition at the i-th stage, n represents that there are n scoring stages in the eloquence competition, a i Represents the relative importance coefficient of the i-th stage.

[0428] In a possible implementation, the intelligent scoring system for eloquence competition further includes: a training resource module 105;

[0429] The training resource module 105 is used to store the training resources of the eloquence competition using cloud storage technology; wherein the training resource module 105 includes: a resource acquisition submodule 1001, a resource storage submodule 1002, a user verification submodule 1003 and a resource download submodule 1004;

[0430] The resource acquisition submodule 1001 is used to acquire training resources for eloquence competition;

[0431] The resource storage submodule 1002 is used to store the training resources of the eloquence competition in a training resource database by combining document storage technology, video media storage technology, and cloud storage technology;

[0432] The user verification submodule 1003 is used to manage user information and perform user verification using a relational database or a non-relational database;

[0433] The resource downloading submodule 1004 is used to obtain the training resources of the eloquence competition through the training resource database after the user verification is successful.

[0434] The more detailed working principle and step flow of this embodiment can refer to, but are not limited to, the relevant records of Embodiment 1.

[0435] The implementation of the above embodiments of the present invention has the following beneficial effects:

[0436] The present invention provides an intelligent scoring system for eloquence competitions, comprising: a speech data module, an eloquence dimension module, an intelligent contract module and a competition scoring module; the system can score eloquence speeches from multiple modes, realize intelligent, standardized and personalized eloquence competition scoring, and improve the efficiency of eloquence competition scoring.

[0437] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. An intelligent scoring method for eloquence competition, characterized in that: include: Acquire the contestant's speech data, use blockchain technology to store the contestant's speech data on the blockchain, and use distributed storage technology to store the contestant's speech data in multiple distributed storage nodes; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data, and the contestant's speech video data; Acquire multiple eloquence dimension indicators and the weights of each eloquence dimension indicator, and automatically generate scoring rules according to the multiple eloquence dimension indicators and the weights of each eloquence dimension indicator, including: acquire multiple eloquence dimension indicators and the weights of each eloquence dimension indicator; wherein the eloquence dimension indicators include grammatical structure indicators, sentiment analysis indicators, vocabulary richness indicators and speech speed indicators; automatically generate scoring rules according to the multiple eloquence dimension indicators and the weights of each eloquence dimension indicator; wherein the scoring rules include: establish scoring models for the multiple eloquence dimension indicators respectively, obtain multiple eloquence dimension scoring models, and the multiple eloquence dimension scoring models include grammatical structure scoring A model, a sentiment analysis scoring model, a vocabulary richness scoring model and a speech speed scoring model are provided; the natural language processing technology is used to extract the features of the multiple eloquence dimension indicators from the contestant's speech data, and the eloquence dimension features of the multiple contestants' speech data are obtained, and the eloquence dimension features include structural features, sentiment features, vocabulary features and speech speed features; the eloquence dimension features of the contestant's speech data are respectively input into the eloquence dimension scoring model corresponding to the eloquence dimension features, and the scores of the contestants' various eloquence dimension indicators are output; according to the weights of the various eloquence dimension indicators and the scores of the various eloquence dimension indicators of the contestants, the total score of the first speech of the contestant is calculated; The method comprises: obtaining the competition information and preset competition rules input by the user, converting the preset competition rules and the scoring rules into contract codes using a smart contract programming language, executing the contract codes to obtain a smart contract, and storing the smart contract in a contract library and a blockchain, including: obtaining the competition information and the preset competition rules input by the user; wherein the competition information input by the user includes an audience participation factor, a plurality of speech dimensions, and a weight of each speech dimension; the preset competition rules include: obtaining a plurality of speech dimensions, wherein the speech dimensions include content richness, language expression ability, emotional appeal, reaction speed and interactivity dimensions, and structural organization and time control dimensions; dynamically adjusting the weight of each speech dimension using the audience participation factor to obtain the adjusted weight of each speech dimension; processing the speech data of the contestant using natural language processing technology and machine learning technology to obtain the score of each speech dimension of the contestant; combining the weight of each speech dimension and the score of each speech dimension of the contestant to calculate the total score of the second speech of the contestant; converting the preset competition rules and the scoring rules into contract codes using a smart contract programming language, executing the contract codes to obtain a smart contract, and storing the smart contract in a contract library and a blockchain; The smart contract is called in the blockchain to score the contestant's speech data to obtain the contestant's eloquence competition score.

2. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: The obtaining of the contestant's speech data, using blockchain technology to store the contestant's speech data on a blockchain, and using distributed storage technology to store the contestant's speech data in multiple distributed storage nodes, includes: Acquire the contestant's speech data; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data; Using an encryption algorithm to convert the contestant's speech data into encrypted speech data, and generating a digital signature of the contestant's speech data based on the encrypted speech data; Using blockchain technology to verify the hash value or digital signature of the contestant's speech data, and after the hash value or digital signature of the contestant's speech data is successfully verified, storing the contestant's speech data on the blockchain; The contestant's speech data and the encrypted speech data are sharded according to preset algorithm rules to obtain multiple shards of the contestant's speech data and multiple shards of the encrypted speech data, and the multiple shards of the contestant's speech data and the multiple shards of the encrypted speech data are stored in multiple distributed storage nodes using distributed storage technology.

3. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: The eloquence dimension features of the contestant's speech data are respectively input into the eloquence dimension scoring model corresponding to the eloquence dimension features, and the scores of the contestants' eloquence dimension indicators are output, including: The structural features of the contestant's speech data are input into the grammatical structure scoring model, and the contestant's grammatical structure score is calculated based on the syntax tree analysis method; wherein the grammatical structure score is specifically as follows: In the formula, represents the grammatical structure score of the contestant; g(xi) represents the grammatical structure score of the contestant's speech data; The emotional features of the contestant's speech data are input into the emotional analysis scoring model, and the contestant's emotional analysis score is output; wherein the emotional analysis score is specifically as follows: In the formula, represents the sentiment analysis score of the contestant; h(xi) represents the sentiment analysis score of the contestant’s speech data; The lexical features of the contestant's speech data are input into the lexical richness scoring model, the lexical importance score of the contestant's speech data is calculated using the TF-IDF method and the topic modeling method, the lexical diversity score of the contestant's speech data is calculated based on the lexical diversity index, and the lexical richness score of the contestant is calculated by combining the lexical importance score and the lexical diversity score; wherein the lexical richness score is specifically as follows: In the formula, represents the lexical richness score of the contestant; p(xi) represents the lexical richness score of the contestant's speech data; The speech speed feature of the contestant's speech data is input into the speech speed scoring model, and the contestant's speech speed score is output; wherein the speech speed score is specifically as follows: In the formula, represents the contestant's speech speed score; q(xi) represents the speech rhythm score and speech fluency score of the contestant's speech data.

4. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: Calculating the total score of the first speech of the contestant according to the weights of the eloquence dimension indicators and the scores of the eloquence dimension indicators of the contestant includes: According to the weights of the eloquence dimension indicators and the scores of the eloquence dimension indicators of the contestant, the total score of the first speech of the contestant is calculated; wherein the total score of the first speech of the contestant is as follows: In the formula, S x Indicates the total score of the contestant's first speech; W j represents the weight of each eloquence dimension indicator, and j represents the number of eloquence dimension indicators; Indicates the eloquence dimension index y of the i-th paragraph of the contestant's speech data j The score on Among them, the weight of each eloquence dimension indicator W j The details are as follows: W j =f(△y i ,△x) Where W j Indicates the weight of each eloquence dimension indicator; △y i Represents the eloquence dimension index y j The degree of variation in the contestant's speech data, △x represents the complexity of the contestant's speech data.

5. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: The method of dynamically adjusting the weight of each speech dimension using the audience participation factor to obtain the adjusted weight of each speech dimension includes: The weights of various speech dimensions are dynamically adjusted using the audience participation factor to obtain the adjusted weights of various speech dimensions; wherein the adjusted weights of various speech dimensions are specifically as follows: In the formula, represents the weight of each speech dimension after adjustment; w X represents the weight of each speech dimension before adjustment, f() represents the nonlinear mapping function, V X represents the audience participation factor, X represents the various speech dimensions of the contestants, X∈{C,L,E,R,T}, C represents the content richness, L represents the language expression ability, E represents the emotional appeal, R represents the reaction speed and interactivity dimension, and T represents the structural organization and time control dimension.

6. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: The natural language processing technology and machine learning technology are used to process the contestant's speech data to obtain the contestant's scores in various speech dimensions, including: Use text mining technology and NLP algorithms to analyze the contestants' speech data and obtain the contestants' content richness scores; Use natural language processing technology to analyze contestants’ speech data and obtain contestants’ language expression ability scores; Use sentiment analysis algorithms to analyze contestants’ speech data and obtain contestants’ emotional appeal scores; Use machine learning models to analyze contestants’ speech data and obtain contestants’ scores on reaction speed and interactivity. Natural language processing technology is used to analyze the contestants' speech data to obtain the contestants' scores on structural organization and time control dimensions.

7. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: The step of combining the weights of each speech dimension and the score of each speech dimension of the contestant to calculate the total score of the contestant's second speech includes: The total score of the contestant's second speech is calculated by combining the weights of each speech dimension and the contestant's score of each speech dimension; wherein the total score of the contestant's second speech is as follows: S=ω C ·C+ω L ·L+ω E ·E+ω R ·R+ω T ·T Where S represents the total score of the contestant’s second speech; ω C represents the weight coefficient of content richness, C represents the content richness score; ω L represents the weight coefficient of language expression ability, L represents the language expression ability score; ω E represents the weight coefficient of emotional appeal, E represents the emotional appeal score; ω R Represents the weight coefficient of the reaction speed and interactivity dimensions, R represents the reaction speed and interactivity dimension score; ω T It represents the weight coefficient of the structural organization and time control dimension, and T represents the score of the structural organization and time control dimension.

8. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: The calling of the smart contract in the blockchain to score the contestant's speech data and obtain the contestant's eloquence competition score includes: Get user request; Downloading the contestant's speech data from the blockchain and multiple distributed storage nodes according to the user's request; Calling the smart contract from the blockchain according to the user request; The contestant's speech data is scored for multiple stages of eloquence competition according to the smart contract to obtain a total eloquence competition score.

9. An intelligent scoring method for eloquence competition as claimed in claim 8, characterized in that: The eloquence competition scoring of the contestant's speech data at multiple stages according to the smart contract to obtain the total eloquence competition score includes: According to the smart contract, the contestant's speech data is scored in multiple stages of eloquence competition to obtain a total eloquence competition score; wherein the total eloquence competition score is as follows: In the formula, S total Indicates the total score of the eloquence competition; S i represents the score of the eloquence competition at the i-th stage, n represents that there are n scoring stages in the eloquence competition, a i Represents the relative importance coefficient of the i-th stage.

10. The intelligent scoring method for eloquence competition according to claim 1, characterized in that: Also includes: Use cloud storage technology to store training resources for eloquence competitions; Get training resources for eloquence competitions; By combining document storage technology, video media storage technology and cloud storage technology, the training resources of the eloquence competition are stored in a training resource database; Use relational or non-relational databases to manage user information and perform user authentication; After the user verification is successful, the training resources for the eloquence competition are obtained through the training resource database.

11. An intelligent scoring system for eloquence competition, characterized in that: include: Speech data module, eloquence dimension module, smart contract module and competition scoring module; The speech data module is used to obtain the contestant's speech data, use blockchain technology to store the contestant's speech data on the blockchain, and use distributed storage technology to store the contestant's speech data in multiple distributed storage nodes; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data; The eloquence dimension module is used to obtain multiple eloquence dimension indicators and the weights of each eloquence dimension indicator, and automatically generate scoring rules according to the multiple eloquence dimension indicators and the weights of each eloquence dimension indicator; The eloquence dimension module includes: an eloquence dimension indicator submodule and an eloquence dimension score submodule; The eloquence dimension index submodule is used to obtain multiple eloquence dimension indicators and the weight of each eloquence dimension indicator; wherein the eloquence dimension indicators include grammatical structure indicators, sentiment analysis indicators, vocabulary richness indicators and speech speed indicators; The eloquence dimension scoring submodule is used to automatically generate scoring rules according to the multiple eloquence dimension indicators and the weights of the various eloquence dimension indicators; wherein the scoring rules include: establishing scoring models for the multiple eloquence dimension indicators respectively, obtaining multiple eloquence dimension scoring models, and the multiple eloquence dimension scoring models include a grammatical structure scoring model, a sentiment analysis scoring model, a vocabulary richness scoring model and a speech speed scoring model; using natural language processing technology to extract features of the multiple eloquence dimension indicators from the contestant's speech data, and obtaining eloquence dimension features of the multiple contestant's speech data, and the eloquence dimension features include structural features, sentiment features, vocabulary features and speech speed features; inputting the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring model corresponding to the eloquence dimension features, and outputting the scores of the contestant's various eloquence dimension indicators; calculating the total score of the contestant's first speech according to the weights of the various eloquence dimension indicators and the scores of the various eloquence dimension indicators of the contestant; The smart contract module is used to obtain the competition information and preset competition rules input by the user, convert the preset competition rules and the scoring rules into contract codes using a smart contract programming language, execute the contract code to obtain a smart contract, and store the smart contract in a contract library and a blockchain; The smart contract module includes: a competition information submodule and a contract code submodule; The competition information submodule is used to obtain the competition information input by the user and the preset competition rules; wherein the competition information input by the user includes audience participation factors, multiple speech dimensions and weights of each speech dimension; the preset competition rules include: obtaining multiple speech dimensions, the speech dimensions include content richness, language expression ability, emotional appeal, reaction speed and interactivity dimensions, and structural organization and time control dimensions; using the audience participation factor to dynamically adjust the weight of each speech dimension to obtain the adjusted weight of each speech dimension; using natural language processing technology and machine learning technology to process the contestant's speech data to obtain the contestant's scores for each speech dimension; combining the weights of each speech dimension and the contestant's scores for each speech dimension, calculating the contestant's total score for the second speech; The contract code submodule is used to convert the preset competition rules and the scoring rules into contract codes using a smart contract programming language, execute the contract codes to obtain a smart contract, and store the smart contract in a contract library and a blockchain; The competition scoring module is used to call the smart contract in the blockchain to score the contestant's speech data and obtain the contestant's eloquence competition score.

12. An intelligent scoring system for eloquence competition as claimed in claim 11, characterized in that: The speech data module includes: a data acquisition submodule, a data encryption submodule, a data verification submodule and a shard storage submodule; The data acquisition submodule is used to acquire the contestant's speech data; wherein the contestant's speech data includes the contestant's speech voice data, the contestant's speech text data and the contestant's speech video data; The data encryption submodule is used to convert the contestant's speech data into encrypted speech data using an encryption algorithm, and generate a digital signature of the contestant's speech data based on the encrypted speech data; The data verification submodule is used to verify the hash value or digital signature of the contestant's speech data using blockchain technology, and after the hash value or digital signature of the contestant's speech data is successfully verified, the contestant's speech data is stored on the blockchain; The shard storage submodule is used to shard the contestant's speech data and the encrypted speech data according to preset algorithm rules, obtain multiple shards of the contestant's speech data and multiple shards of the encrypted speech data, and use distributed storage technology to store the multiple shards of the contestant's speech data and the multiple shards of the encrypted speech data in multiple distributed storage nodes.

13. The intelligent scoring system for eloquence competition according to claim 11, characterized in that: The eloquence dimension scoring submodule includes: a scoring model unit; The scoring model unit is used to input the eloquence dimension features of the contestant's speech data into the eloquence dimension scoring model corresponding to the eloquence dimension features, and output the scores of each eloquence dimension indicator of the contestant; wherein the scoring model unit includes: a grammatical structure scoring model sub-unit, a sentiment analysis scoring model sub-unit, a vocabulary richness scoring model sub-unit and a speech speed scoring model sub-unit; The grammatical structure scoring model subunit is used to input the structural features of the contestant's speech data into the grammatical structure scoring model, and calculate the contestant's grammatical structure score based on the syntax tree analysis method; wherein the grammatical structure score is specifically as follows: In the formula, represents the grammatical structure score of the contestant; g(xi) represents the grammatical structure score of the contestant's speech data; The sentiment analysis scoring model subunit is used to input the sentiment features of the contestant's speech data into the sentiment analysis scoring model and output the contestant's sentiment analysis score; wherein the sentiment analysis score is specifically as follows: In the formula, represents the sentiment analysis score of the contestant; h(xi) represents the sentiment analysis score of the contestant’s speech data; The lexical richness scoring model subunit is used to input the lexical features of the contestant's speech data into the lexical richness scoring model, calculate the lexical importance score of the contestant's speech data using the TF-IDF method and the topic modeling method, calculate the lexical diversity score of the contestant's speech data based on the lexical diversity index, and calculate the contestant's lexical richness score by combining the lexical importance score and the lexical diversity score; wherein the lexical richness score is specifically as follows: In the formula, represents the lexical richness score of the contestant; p(xi) represents the lexical richness score of the contestant's speech data; The speech speed scoring model subunit is used to input the speech speed feature of the contestant's speech data into the speech speed scoring model, and output the contestant's speech speed score; wherein the speech speed score is specifically as follows: In the formula, represents the contestant's speech speed score; q(xi) represents the speech rhythm score and speech fluency score of the contestant's speech data.

14. The intelligent scoring system for eloquence competition according to claim 11, characterized in that: The eloquence dimension scoring submodule includes: the first speech total scoring unit; The first speech total scoring unit is used to calculate the contestant's first speech total score according to the weights of the eloquence dimension indicators and the contestant's eloquence dimension indicator scores; wherein the contestant's first speech total score is specifically as follows: In the formula, S x Indicates the total score of the contestant's first speech; W j represents the weight of each eloquence dimension indicator, and j represents the number of eloquence dimension indicators; Indicates the eloquence dimension index y of the i-th paragraph of the contestant's speech data j The score on Among them, the weight of each eloquence dimension indicator W j The details are as follows: W j =f(△y i ,△x) Where W j Indicates the weight of each eloquence dimension indicator; △y i Represents the eloquence dimension index y j The degree of variation in the contestant's speech data, △x represents the complexity of the contestant's speech data.

15. The intelligent scoring system for eloquence competition according to claim 11, characterized in that: The competition information submodule includes: a weight adjustment unit; The weight adjustment unit is used to dynamically adjust the weight of each speech dimension using the audience participation factor to obtain the adjusted weight of each speech dimension; wherein the adjusted weight of each speech dimension is specifically as follows: In the formula, represents the weight of each speech dimension after adjustment; w X represents the weight of each speech dimension before adjustment, f() represents the nonlinear mapping function, V X represents the audience participation factor, X represents the various speech dimensions of the contestants, X∈{C,L,E,R,T}, C represents the content richness, L represents the language expression ability, E represents the emotional appeal, R represents the reaction speed and interactivity dimension, and T represents the structural organization and time control dimension.

16. The intelligent scoring system for eloquence competition according to claim 11, characterized in that: The competition information submodule includes: a speech dimension scoring unit; The speech dimension scoring unit is used to process the contestant's speech data using natural language processing technology and machine learning technology to obtain the contestant's scores for each speech dimension; wherein the speech dimension scoring unit includes: a first dimension sub-unit, a second dimension sub-unit, a third dimension sub-unit, a fourth dimension sub-unit and a fifth dimension sub-unit; The first dimension sub-unit is used to analyze the contestant's speech data using text mining technology and NLP algorithm to obtain the contestant's content richness score; The second dimension sub-unit is used to analyze the contestant's speech data using natural language processing technology to obtain the contestant's language expression ability score; The third dimension sub-unit is used to analyze the contestant's speech data using a sentiment analysis algorithm to obtain the contestant's sentiment appeal score; The fourth dimension sub-unit is used to analyze the contestant's speech data using a machine learning model to obtain the contestant's reaction speed and interactivity dimension scores; The fifth dimension sub-unit is used to analyze the contestant's speech data using natural language processing technology to obtain the contestant's structural organization and time control dimension scores.

17. The intelligent scoring system for eloquence competition according to claim 11, characterized in that: The competition information submodule includes: a second speech overall scoring unit; The second speech total scoring unit is used to calculate the contestant's second speech total score by combining the weights of each speech dimension and the contestant's score of each speech dimension; wherein the contestant's second speech total score is specifically as follows: S=ω C ·C+ω L ·L+ω E ·E+ω R ·R+ω T ·T Where S represents the total score of the contestant’s second speech; ω C represents the weight coefficient of content richness, C represents the content richness score; ω L represents the weight coefficient of language expression ability, L represents the language expression ability score; ω E represents the weight coefficient of emotional appeal, E represents the emotional appeal score; ω R Represents the weight coefficient of the reaction speed and interactivity dimensions, R represents the reaction speed and interactivity dimension score; ω T It represents the weight coefficient of the structural organization and time control dimension, and T represents the score of the structural organization and time control dimension.

18. The intelligent scoring system for eloquence competition according to claim 11, characterized in that: The competition scoring module includes: a user request acquisition submodule, a speech data calling submodule, a smart contract calling submodule and an eloquence competition scoring submodule; The user request acquisition submodule is used to acquire user requests; The speech data calling submodule is used to download the contestant's speech data from the blockchain and multiple distributed storage nodes according to the user request; The smart contract calling submodule is used to call the smart contract from the blockchain according to the user request; The eloquence competition scoring submodule is used to perform multiple stages of eloquence competition scoring on the contestant's speech data according to the smart contract to obtain a total eloquence competition score.

19. An intelligent scoring system for eloquence competition as claimed in claim 18, characterized in that: The eloquence competition scoring submodule is used to score the contestant's speech data at multiple stages of eloquence competition according to the smart contract to obtain the total score of the eloquence competition, including: The eloquence competition scoring submodule is used to score the contestant's speech data in multiple stages of eloquence competition according to the smart contract to obtain a total score of the eloquence competition; wherein the total score of the eloquence competition is as follows: In the formula, S total Indicates the total score of the eloquence competition; S i represents the score of the eloquence competition at the i-th stage, n represents that there are n scoring stages in the eloquence competition, a i Represents the relative importance coefficient of the i-th stage.

20. An intelligent scoring system for eloquence competition as claimed in claim 19, characterized in that: Also includes: Training resource modules; The training resource module is used to store the training resources for the eloquence competition using cloud storage technology; wherein the training resource module includes: a resource acquisition submodule, a resource storage submodule, a user verification submodule and a resource download submodule; The resource acquisition submodule is used to acquire training resources for eloquence competitions; The resource storage submodule is used to store the training resources of the eloquence competition in a training resource database by combining document storage technology, video media storage technology and cloud storage technology; The user verification submodule is used to manage user information and perform user verification using a relational database or a non-relational database; The resource downloading submodule is used to obtain the training resources for the eloquence competition through the training resource database after the user verification is successful.

Images