CN116484805B - Intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis - Google Patents

Abstract

The present invention provides an intelligent cleaning processing method for electric power reports that combines knowledge graphs and semantic analysis, including: constructing corresponding knowledge graphs based on triple relationships corresponding to corresponding knowledge information; determining words with data attributes in words as the first words, using other words as the second words; generate a corresponding data demand list based on the user's computing requirements configured for this power report, and select the first word with corresponding analysis attributes as the third word based on the data demand list; based on cleaning The verification strategy and historical data are used to verify the correctness of the third word; the cleaning processing strategy corresponding to the third word is determined according to the verification doubt type of the third word, and the verification doubt type includes the unreasonable doubt type or the unverified doubt type; the third word is The three words are input into the preset model to calculate the analyzed data after data cleaning, and the corresponding processed data is obtained based on the analyzed data and knowledge map.

Description

Intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis

Technical field

The invention relates to data processing technology, and in particular, to an intelligent cleaning and processing method for power reports that combines knowledge graphs and semantic analysis.

Background technique

With the digital transformation of power grid enterprises, the previous paper power reports are gradually converted into electronic power reports for data summary. Power reports generally include data statistics from multiple dimensions, such as residential power consumption data, industrial power consumption data, park power consumption data, etc. Power reports are an important basis for power grid companies to analyze data.

Due to the large number of power reports and the large amount of data, the data volume is large. When staff analyze the power reports, they often need to read a large number of power reports, and cannot effectively and quickly obtain the information in the power reports based on their needs. Relevant data, therefore, it is necessary to organize effective data for power reports. In the existing technology, the data of power reports are often extracted and organized manually. However, for power reports with a large amount of data, the workload of sorting them is huge, and due to unavoidable human reasons, This can lead to some data sorting errors.

Therefore, how to intelligently clean power reports and automatically verify the cleaned data has become an urgent problem to be solved.

Contents of the invention

Embodiments of the present invention provide an intelligent cleaning and processing method for power reports that combines knowledge graphs and semantic analysis. It can intelligently clean power reports and automatically verify the cleaned data. It can also realize automatic analysis of data based on the needs of workers. , and obtain the corresponding analysis data.

A first aspect of the embodiment of the present invention provides an intelligent cleaning processing method for power reports that combines knowledge graph and semantic analysis, including:

Receive the knowledge information configured by the user for intelligent cleaning and processing of the power report, and construct a corresponding knowledge graph based on the triplet relationship corresponding to the corresponding knowledge information, and the knowledge graph includes multiple knowledge nodes;

Perform word segmentation processing on the sentences in the power report to obtain multiple words, determine the words with data attributes in the words as the first words, use other words as the second words, and combine the second words in the sentences to the corresponding first words Perform semantic analysis to obtain the analytical attributes of the first word;

Generate a corresponding data requirement list based on the user's configured calculation requirements for this power report, and select the first word with the corresponding analysis attribute as the third word based on the data requirement list;

The cleaning verification strategy corresponding to the third word is determined according to the analysis attribute of the third word, and the correctness of the third word is verified based on the cleaning verification strategy and historical data. Each type of analysis attribute has a predetermined Designed cleaning and verification strategies;

If it is determined that the verification fails, the verification doubt type corresponding to the verification failure is determined, and the cleaning processing strategy corresponding to the third word is determined according to the verification doubt type of the third word, and the verification doubt type includes an unreasonable doubt type or an unreasonable doubt type. Verify doubtful types;

After determining that all third words corresponding to the data requirement list are obtained, the third words are input into the preset model to calculate the analyzed data after data cleaning, and the corresponding processing data is obtained based on the analyzed data and the knowledge graph. .

Optionally, in a possible implementation of the first aspect, the receiving user intelligently cleans and processes the configured knowledge information for the power report, and constructs a corresponding knowledge graph based on the triplet relationship corresponding to the corresponding knowledge information. , the knowledge graph includes multiple knowledge nodes, including:

The knowledge information includes first knowledge information corresponding to the analysis data, and second knowledge information corresponding to the processing data, and each first knowledge information or second knowledge information has a corresponding knowledge node;

The corresponding knowledge nodes are connected according to the triplet relationship configured by the user on the first knowledge information and the second knowledge information, and a corresponding knowledge graph is constructed and generated.

Optionally, in a possible implementation of the first aspect, word segmentation processing is performed on the statements in the power report to obtain multiple words, words with data attributes among the words are determined as the first words, and other words are used as the third words. The two words are combined with the second word in the sentence to perform semantic analysis on the corresponding first word to obtain the analytical attributes of the first word, including:

Perform word segmentation processing on the statements in the power report to obtain multiple words, and determine the words with data attributes among the words as the first words. The words with data attributes at least include Arabic numerals, uppercase numerals, and traditional Chinese numerals;

Use words other than the first word in all participles as second words, traverse the second words and compare them with the preset words. If it is determined that the second words correspond to the preset words, determine the second words based on the preset words. is the second word to be analyzed;

According to the positional relationship between each first word and the second word to be analyzed, the second word to be analyzed associated with the corresponding first word is determined, and the corresponding first word is semantically analyzed based on the associated second word. , get the analytical attributes of the first word;

If it is determined that the format of the first word is inconsistent with the preset format, a preset first conversion template is determined according to the format of the first word, and the first word is converted based on the first conversion template to obtain a format that meets the format requirements. First word.

Optionally, in a possible implementation of the first aspect, determining the second word associated with the corresponding first word based on the positional relationship between each first word and the second word to be analyzed, Perform semantic analysis on the corresponding first word based on the associated second word to obtain the analytical attributes of the first word, including:

If it is determined that there are multiple first words and preset merged words in a sentence, it is determined that the corresponding first words can be merged into one first word, and corresponding merge tags are added to the first words that can be merged, so that subsequent When processing the first word, merge the corresponding first word based on the merge tag;

If it is determined that a sentence contains a first word or multiple first words that can be combined into one first word, then all the second words to be analyzed in the corresponding sentence are combined with the corresponding first words or the merged first words. Associated;

If it is determined that there are multiple first words in a sentence and the multiple words cannot be combined into one first word, segment the sentence based on the position of the first word to obtain a segmentation result, and determine each first word based on the segmentation result. A second word to which one word is associated;

The analysis attributes are any one or more of subject analysis attributes, trend change analysis attributes, and concept analysis attributes included in the second word.

Optionally, in a possible implementation of the first aspect, if it is determined that there are multiple first words in a sentence and the multiple words cannot be combined into one first word, then the sentence is compared based on the position of the first word. Perform segmentation to obtain segmentation results, and determine the second words associated with each first word based on the segmentation results, including:

Determine the positions of all first words in the sentence, segment the sentence based on the positions of the first words to obtain multiple sub-segments, and determine the front and adjacent segments of each first word as associated segments;

The second word in the associated segment is used as the second word associated with the corresponding first word.

Optionally, in a possible implementation of the first aspect, if it is determined that the format of the first word is inconsistent with the preset format, the preset first conversion template is determined according to the format of the first word, based on The first conversion template converts the first word to obtain the first word that meets the format requirements, including:

Compare the format of the first word with a preset format, the format of the first word is Arabic numeral format, uppercase numeral format or traditional Chinese numeral format, and the preset format is Arabic numeral format;

If it is determined that the format of the first word is inconsistent with the preset format, the preset first conversion template is determined according to the format of the first word, and the first word is decomposed to obtain the corresponding numerical word and position word, based on the position word Fill the numerical words into the relative empty positions in the first conversion template;

If it is determined that there are unfilled vacancies in the first conversion template, the unfilled vacancies are filled with 0;

If it is determined that all the gaps in the first conversion template have been filled, then the number composed of all the gaps will be used as the first word that meets the format requirements.

Optionally, in a possible implementation of the first aspect, if it is determined that the format of the first word is inconsistent with the preset format, the preset first conversion template is determined according to the format of the first word, and the The first word is decomposed to obtain the corresponding numerical words and positional words, and the numerical words are filled into relative vacancies in the first conversion template based on the positional words, including:

Determine a preset first conversion template according to the format of the first word. The first conversion template has a plurality of gaps. The first word is decomposed and recognized to obtain the corresponding numerical word and position word. Each numerical word and its other The adjacent position words at the rear are treated as a numerical value group;

Extract the frontmost position word in the entire first word to obtain the corresponding preset number of gaps, and each position word has a corresponding preset number of gaps;

Reserve a preset number of vacancies in the first conversion template, and add position labels to each vacancy in sequence from back to front according to the reserved vacancies;

The corresponding position labels and gaps are determined according to the position words in each numerical group, and the numerical words in the numerical group are filled into the determined gaps.

Optionally, a possible implementation of the first aspect also includes:

If it is determined that there are corresponding merge tags between multiple first words, the multiple first words are calculated according to the corresponding merge tags between the multiple first words to obtain the calculated first words, each merged Words have a preset calculation method.

Optionally, in a possible implementation of the first aspect, the corresponding data demand list is generated according to the computing demand configured by the user for this power report, and the data with corresponding analysis attributes is selected based on the data demand list. The first word as the third word includes:

Generate corresponding data statistics tables based on all the first words in the preset format, the merged first words, and the respectively associated second words;

Generate a corresponding data requirement list based on the calculation requirements configured by the user for this power report, and extract the fourth word in the data requirement list;

The fourth word is compared with the second word in the data statistics table. If the fourth word corresponds to the second word, the first word corresponding to the corresponding second word in the data statistics table is used as the third word.

Optionally, in a possible implementation of the first aspect, the cleaning verification strategy corresponding to the third word is determined based on the analysis attribute of the third word, based on the cleaning verification strategy, historical data Verify the correctness of the third word. Each type of analysis attribute has a preset cleaning verification strategy, including:

Determine the second word corresponding to the analysis attribute of the third word, and determine the corresponding cleaning verification strategy according to the corresponding second word. The second words of each type of subject analysis attribute or concept analysis attribute have a preset cleaning verification strategy;

The third word is compared with the first threshold interval included in the cleaning verification strategy. If the value of the third word is within the first threshold interval, it is determined that the corresponding third word satisfies the verification of the correctness of the cleaning verification strategy. ;

Corresponding historical data is determined according to the subject analysis attribute or concept analysis attribute of the second word, the average value is calculated according to the historical data, and the average value is interval processed according to the preset proportion value to obtain the second threshold. interval;

If the value of the third word is within the second threshold interval, it is determined that the corresponding third word meets the verification of the correctness of the historical data.

Optionally, in a possible implementation of the first aspect, if it is determined that the verification fails, the verification doubt type corresponding to the verification failure is determined, and the verification doubt type corresponding to the third word is determined according to the verification doubt type of the third word. Corresponding cleaning processing strategy, the verification doubt types include unreasonable doubt types or unverified doubt types, including:

If it is determined that the value of the third word is not within the first threshold interval, it is determined that the verification doubt type corresponding to the third word is an unreasonable doubt type;

If it is determined that the value of the third word is not within the second threshold interval, then it is determined that the verification questionable type corresponding to the third word is the questionable type to be verified;

If the third word is of an unreasonable and doubtful type, mark the corresponding third word as an error, and output the third word and the corresponding second word so that the user can directly update the third word;

If the third word is of a questionable type to be verified, the corresponding third word is marked as to be verified, and the third word and the corresponding second word are output. If it is judged that the user inputs positive verification information, the corresponding The third word serves as the final third word;

If it is determined that the user input is negative verification information, the corresponding third word is updated according to the user input.

Optionally, in a possible implementation of the first aspect, after determining that all third words corresponding to the data requirement list are obtained, the third words are input into a preset model to calculate the data after cleaning. The analysis data and the corresponding processing data are obtained based on the analysis data and knowledge graph, including:

After determining that all third words corresponding to the data requirement list are obtained, the third words are input into the preset model to calculate the analyzed data after data cleaning, and the analyzed data includes the corresponding segmented words;

Input the segmented words into the knowledge graph to determine knowledge nodes with corresponding first knowledge information, and based on the determined knowledge nodes and other knowledge nodes with second knowledge information;

The second knowledge information of other knowledge nodes with second knowledge information is counted, and the corresponding processing data is obtained and output.

A second aspect of the embodiment of the present invention provides an intelligent cleaning and processing system for power reports that combines knowledge graph and semantic analysis, including:

The construction module is used to receive the knowledge information configured by the user for intelligent cleaning and processing of power reports, and build a corresponding knowledge graph based on the triple relationship corresponding to the corresponding knowledge information. The knowledge graph includes multiple knowledge nodes;

The processing module is used to perform word segmentation processing on the statements in the power report to obtain multiple words, determine the words with data attributes in the words as the first words, use other words as the second words, and combine the second word pairs in the sentences. Perform semantic analysis on the corresponding first word to obtain the analytical attributes of the first word;

A generation module configured to generate a corresponding data requirement list based on the user's configured computing requirements for this power report, and select the first word with the corresponding analysis attribute as the third word based on the data requirement list;

A verification module, configured to determine the cleaning verification strategy corresponding to the third word based on the analytical attributes of the third word, and verify the correctness of the third word based on the cleaning verification strategy and historical data. Each type The analysis attributes have preset cleaning verification strategies;

A judgment module, configured to determine the verification doubt type corresponding to the verification failure if it is judged that the verification fails, and determine the cleaning processing strategy corresponding to the third word according to the verification doubt type of the third word, and the verification doubt type includes not Types of reasonable doubts or types of doubts to be verified;

The calculation module is used to, after determining and obtaining all third words corresponding to the data requirement list, input the third words into the preset model to calculate the analyzed data after data cleaning, which is obtained based on the analyzed data and the knowledge map. Process the data accordingly.

Beneficial effects:

1. This solution can process and analyze the power report in the word dimension, obtain the first word with data attributes and other types of second words, and combine the analysis of the first word and the second word to obtain the first word analyze data. When the user has calculation requirements, a data requirement list can be generated, and the correctness of the third word can be verified based on the cleaning verification strategy and historical data. At the same time, when an abnormal situation occurs, the staff can be used to interact, so that the analyzed data is Unmistakable. Finally, this solution will also combine the knowledge information configured by users for intelligent cleaning and processing of power reports to form a knowledge graph, and use the knowledge graph to output the processing results corresponding to the analysis data. In summary, this solution can intelligently clean the power report and automatically verify the cleaned data. It can also realize automatic analysis of the data based on the needs of the staff and obtain corresponding analysis data.

2. When this program processes and analyzes the power report in the word dimension, it will classify the words based on their attributes, obtain the first word and the second word, and combine the first word with the second word to be analyzed. The positional relationship between the two words determines the second word to be analyzed that is associated with the corresponding first word, and realizes the analysis and correlation of the statements in the power report. At the same time, the format of the first word will be compared with the preset format. If it does not correspond, the format will be converted. When performing format conversion, the first conversion template will be used for conversion, and the numerical word and position will also be combined. Words realize the positioning and supplement of numbers. Through the above method, the data can be converted into the required format for subsequent processing and judgment.

3. When this solution is calculated based on the calculation requirements configured by the user for this power report, it will first verify the correctness of the third word based on the cleaning verification strategy and historical data. During verification, this plan sets up two verification methods, one is to use the first threshold interval to determine the rationality of the data, and the other is to combine historical data to determine the doubtfulness of the data. In addition, the verification results will be combined with the interaction with the staff to correct the data so that the data is accurate. Finally, this solution will also combine the knowledge graph to process the analysis data to obtain corresponding processing results, and intelligently assist staff in cleaning, analyzing and processing power reports.

Description of the drawings

Figure 1 is a schematic flow chart of an intelligent cleaning processing method for power reports that combines knowledge graph and semantic analysis provided by an embodiment of the present invention;

Figure 2 is a schematic structural diagram of an intelligent cleaning and processing system for power reports that combines knowledge graph and semantic analysis provided by an embodiment of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments These are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

The terms "first", "second", "third", "fourth", etc. (if present) in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used for Describe a specific order or sequence. It is to be understood that the data so used are interchangeable under appropriate circumstances so that the embodiments of the invention described herein are capable of being practiced in sequences other than those illustrated or described herein.

It should be understood that in various embodiments of the present invention, the size of the sequence numbers of each process does not mean the order of execution. The execution order of each process should be determined by its functions and internal logic, and should not be determined by the execution order of the embodiments of the present invention. The implementation process constitutes no limitation.

It should be understood that in the present invention, "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or equipment that includes a series of steps or units is not necessarily limited to Those steps or elements that are expressly listed may instead include other steps or elements that are not expressly listed or that are inherent to the process, method, product or apparatus.

It should be understood that in the present invention, "plurality" means two or more. "And/or" is just an association relationship that describes related objects. It means that there can be three kinds of relationships. For example, A and/or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. Condition. The character "/" generally indicates that the related objects are in an "or" relationship. "Includes A, B and C" and "includes A, B, C" means that it includes all three of A, B and C, and "includes A, B or C" means that it includes one of A, B and C. "Including A, B and/or C" means including any one, any two or three of A, B and C.

It should be understood that in the present invention, "B corresponding to A", "B corresponding to A", "A corresponding to B" or "B corresponding to A" means that B is associated with A. According to A B can be determined. Determining B based on A does not mean determining B only based on A, but can also determine B based on A and/or other information. The matching between A and B means that the similarity between A and B is greater than or equal to the preset threshold.

Depending on the context, "if" as used herein may be interpreted as "when" or "when" or "in response to determination" or "in response to detection."

The technical solution of the present invention will be described in detail below with specific examples. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

Referring to Figure 1, it is a schematic flow chart of an intelligent cleaning processing method for power reports that combines knowledge graphs and semantic analysis provided by an embodiment of the present invention. The intelligent cleaning processing method for power reports that combines knowledge graphs and semantic analysis includes S1-S6. Specifically, as follows:

S1: Receive the knowledge information configured by the user for intelligent cleaning and processing of power reports, and construct a corresponding knowledge graph based on the triplet relationship corresponding to the corresponding knowledge information. The knowledge graph includes multiple knowledge nodes.

Since this solution is combined with the knowledge graph to process power reports, this solution will receive the knowledge information configured by users for intelligent cleaning and processing of power reports, and then combine the triple relationships corresponding to the corresponding knowledge information to build the corresponding Knowledge graph.

Among them, the knowledge graph includes multiple knowledge nodes.

In some embodiments, S1 (receives the knowledge information configured by the user for intelligent cleaning and processing of power reports, and builds a corresponding knowledge graph based on the triplet relationship corresponding to the corresponding knowledge information. The knowledge graph includes multiple knowledge nodes) including S11-S12:

S11. The knowledge information includes first knowledge information corresponding to the analysis data and second knowledge information corresponding to the processing data. Each first knowledge information or second knowledge information has a corresponding knowledge node.

The knowledge information includes first knowledge information corresponding to the analysis data, and second knowledge information corresponding to the processing data. Each first knowledge information or second knowledge information has a corresponding knowledge node.

Both analyzing data and processing data can be preset. The analysis data is, for example, data on whether residential electricity consumption is too high or industrial electricity consumption is too high. The processing data is, for example, when the residential electricity consumption is too high, the residential power supply is increased, or when the industrial electricity consumption is too high, the industrial power supply is limited. deal with. The above are only examples and are not limited to the above examples.

S12: Connect the corresponding knowledge nodes according to the triplet relationship configured by the user on the first knowledge information and the second knowledge information, and construct and generate the corresponding knowledge graph.

This solution will combine the triplet relationship of the user's configuration of the first knowledge information and the second knowledge information to connect the corresponding knowledge nodes to construct and generate the corresponding knowledge graph. This process is an existing technology and will not be described again.

S2, perform word segmentation processing on the sentences in the power report to obtain multiple words, determine the words with data attributes in the words as the first words, use other words as the second words, and combine the second words in the sentences to the corresponding third words. Perform semantic analysis on a word to obtain the analytical attributes of the first word.

This solution will segment the statements in the power report to obtain multiple words. For example, segmenting "According to this statistics, residents' electricity consumption is 20%" can obtain "according to", "this time", "Statistics", "Resident electricity consumption", "Twenty percent" and other results.

After obtaining the words, the words with data attributes in the words will be determined as the first words, other words will be used as the second words, and the corresponding first words will be semantically analyzed based on the second words in the sentence to obtain the analysis of the first words. Attributes.

In some embodiments, S2 (perform word segmentation processing on the sentences in the power report to obtain multiple words, determine the words with data attributes in the words as the first words, use other words as the second words, combine the third words in the sentences The second word performs semantic analysis on the corresponding first word to obtain the analysis attributes of the first word) including S21-S24:

S21, perform word segmentation processing on the statements in the power report to obtain multiple words, and determine the words with data attributes among the words as the first words. The words with data attributes at least include Arabic numerals, uppercase numerals, and traditional Chinese numerals.

After obtaining the words, this solution will determine the words with data attributes in the words as the first words. Among them, the words with data attributes at least include Arabic numerals, uppercase numerals, and traditional Chinese numerals.

For example, the data attribute of the word "20" is Arabic numerals, the data attribute of the word "twenty" is uppercase numerals, and the data attribute of the word "twenty" is traditional Chinese numerals.

S22, use words other than the first words in all word segments as second words, traverse the second words and compare them with the preset words, and if it is determined that the second words correspond to the preset words, then based on the preset words Determine it as the second word to be analyzed.

It can be understood that the first words are data words, and other words will be marked as second words.

This solution will traverse the second word and compare it with the preset word. If it is determined that the second word corresponds to the preset word, this solution will combine it with the preset word to determine it as the second word to be analyzed.

It is worth mentioning that since word segmentation will result in multiple second words, some second words are useful and some are useless. For example, "according to", "this time" and "statistics" are useless, and "residents" are useless. "Electricity consumption" is useful, then it can be the corresponding second word to be analyzed. When determining the second word to be analyzed, it can be determined in combination with the preset word. The default words are, for example, "residential electricity consumption", "industrial electricity consumption", etc.

S23. According to the positional relationship between each first word and the second word to be analyzed, determine the second word to be analyzed that is associated with the corresponding first word, and perform the analysis on the corresponding first word based on the associated second word. Semantic analysis to obtain the analytical attributes of the first word.

After determining the first word and the second word, this solution determines the second word to be analyzed that is associated with the corresponding first word based on the positional relationship between each first word and the second word to be analyzed.

Then, semantic analysis is performed on the corresponding first word based on the associated second word, and the analysis attribute of the first word is obtained.

It is worth mentioning that this solution needs to combine related words to obtain the meaning of the first word, thereby obtaining the analytical attributes of the first word. The analysis attributes are any one or more of subject analysis attributes, trend change analysis attributes, and concept analysis attributes included in the second word. Among them, the subject analysis attributes can be "residential electricity", "industrial electricity", etc., the trend change analysis attributes can be "too high", "too low", etc., and the concept analysis attributes can be "". For example, after correlation analysis, The analyzed attribute corresponding to the obtained "twenty percent" is "resident electricity consumption".

In some embodiments, S23 (determine the second word associated with the corresponding first word based on the positional relationship between each first word and the second word to be analyzed, and pair the corresponding third word based on the associated second word). Perform semantic analysis on a word to obtain the analysis attributes of the first word) including S231-S233:

S231, if it is determined that there are multiple first words and preset merged words in a sentence, it is determined that the corresponding first words can be merged into one first word, and corresponding merge tags are added to the first words that can be merged, so as to When subsequent processing of the first word is performed, the corresponding first word is merged and processed based on the merge tag.

If it is determined that there are multiple first words and preset merged words in a sentence, it is determined that the corresponding first words can be merged into one first word, and corresponding merge tags are added to the first words that can be merged, so that subsequent When processing the first word, the corresponding first word is merged and processed based on the merge tag.

For example, the preset merging word is "to" in "twenty percent to thirty percent". At this time, this solution will merge the corresponding first words into one first word.

On the basis of the above embodiments, it also includes:

If it is determined that there are corresponding merge tags between multiple first words, the multiple first words are calculated according to the corresponding merge tags between the multiple first words to obtain the calculated first words, each merged Words have a preset calculation method.

For example, the above-mentioned merging may be a merging method of taking an intermediate value. For example, after merging "twenty percent to thirty percent", the first word obtained is "twenty-five percent". It is worth mentioning that for the combined data, it is a rough statistic and may be slightly less accurate.

S232, if it is determined that a sentence has a first word or multiple first words that can be combined into one first word, then all the second words to be analyzed in the corresponding sentence are combined with the corresponding first words or the merged third words. associated with a word.

If it is judged that a sentence has a first word or multiple first words that can be combined into one first word.

This solution will associate all the second words to be analyzed in the corresponding sentence with the corresponding first words or the merged first words.

S233, if it is determined that there are multiple first words in a sentence and the multiple words cannot be combined into one first word, segment the sentence based on the position of the first word to obtain a segmentation result, and determine each segmentation result based on the segmentation result. a second word associated with the first word.

For example, a statement is "the statistical result is that the electricity consumption of residents is 44455 and the electricity consumption of industry is 600000", in which multiple first words "44455" and "600000" cannot be combined into one first word , at this time, this solution will segment the sentence based on the position of the first word to obtain the segmentation result, and then determine the second word associated with each first word based on the segmentation result.

Among them, S233 (if it is determined that there are multiple first words in a sentence and the multiple words cannot be combined into one first word, segment the sentence based on the position of the first word to obtain a segmented result. According to the segmented result Determining the second word associated with each first word) includes S2331-S2332:

S2331. Determine the positions of all first words in the sentence, segment the sentence based on the positions of the first words to obtain multiple sub-segments, and determine the front and adjacent segments of each first word as associated segments.

For example, for "the statistical result is that residential electricity consumption is 44455 and industrial electricity consumption is 600000", the first words are "44455" and "600000".

Based on the position of the first word, the sentence is segmented to obtain multiple sub-segments, namely "the statistical result is that the electricity consumption of residents is", "and the electricity consumption of industry is".

Then, this program will determine the segment in front of each first word and adjacent to it as the associated segment, that is, the associated segment "the statistical result is that the electricity consumption of residents is" is "44455", and the "and industrial electricity consumption is" is the associated segment of "600000".

S2332: Use the second word in the associated segment as the second word associated with the corresponding first word.

This solution will use the second word in the associated segment as the second word associated with the corresponding first word.

Through the above method, different strategies can be analyzed for different situations to obtain the second word associated with the corresponding first word.

S24, if it is determined that the format of the first word is inconsistent with the preset format, determine the preset first conversion template according to the format of the first word, and convert the first word based on the first conversion template to obtain a format that satisfies the format. The first word requested.

In some cases, the format of the first word may be inconsistent with a preset format, which may be an Arabic numeral format.

If it is determined that the format of the first word is inconsistent with the preset format, a preset first conversion template is determined according to the format of the first word, and the first word is converted based on the first conversion template to obtain a first word that meets the format requirements. Through format conversion, subsequent data organization in the corresponding format can be carried out.

In some embodiments, S24 (if it is determined that the format of the first word is inconsistent with the preset format, determine a preset first conversion template based on the format of the first word, and convert the first word based on the first conversion template Convert to obtain the first word that meets the format requirements) including S241-S244:

S241. Compare the format of the first word with a preset format. The format of the first word is Arabic numeral format, uppercase numeral format or traditional Chinese numeral format. The preset format is Arabic numeral format.

This solution will compare the format of the first word with the Arabic numeral format. If the format of the first word does not correspond to the Arabic numeral format, then the format of the first word needs to be converted.

S242, if it is determined that the format of the first word is inconsistent with the preset format, determine the preset first conversion template according to the format of the first word, decompose the first word to obtain the corresponding numerical word and position word, based on The position word fills the numerical word into the relative empty position in the first conversion template.

It is worth mentioning that this solution will set up a preset first conversion template, and then use the preset first conversion template to decompose the first word to obtain the corresponding numerical word and positional word, and convert the numerical word based on the positional word. Fill to the relative empty space in the first conversion template;

Among them, S242 (if it is determined that the format of the first word is inconsistent with the preset format, determine the preset first conversion template according to the format of the first word, decompose the first word to obtain the corresponding numerical word and position word , filling the numerical words into relative vacancies in the first conversion template based on the position words) including S2421-S2424:

S2421. Determine the preset first conversion template according to the format of the first word. The first conversion template has multiple gaps. Decompose and identify the first word to obtain the corresponding numerical word and position word. Each numerical word is The position words adjacent to its rear are treated as a numerical group.

The first conversion template has a plurality of empty spaces, and the empty spaces are used to fill in corresponding data.

First, this solution will decompose and identify the first word to obtain the corresponding numerical characters and positional characters, and then use each numerical character and its adjacent positional character as a numerical value group.

It is worth mentioning that the numerical word is the data corresponding to the digital bit, and the position word is the data corresponding to the position bit. For example, for "two hundred and thirty", "2" and "3" corresponding to "two" and "three" are numerical characters, and "hundred" and "ten" are positional characters. Among them, "2" and "hundred" are a numerical value group, and "3" and "tens" are a numerical value group.

S2422: Extract the frontmost position word in the entire first word to obtain the corresponding preset number of gaps. Each position word has a corresponding preset number of gaps.

This solution extracts the first positional word in the entire first word to obtain the corresponding preset number of slots, where each positional word has a corresponding preset number of slots.

For example, if the front position word is "thousand", then the corresponding preset number of slots is 4; if the front position word is "hundred", then the corresponding preset number of slots is 3; If the first position word is "十", then its corresponding default number of vacancies is 2.

S2423: Reserve a preset number of slots in the first conversion template, and add position labels to each slot in sequence from back to front according to the reserved slots.

This solution will reserve a preset number of vacancies in the first conversion template, and add position labels to each vacancy sequentially from back to front according to the reserved vacancies.

For example, if the first position word is "hundred", then the corresponding preset number of slots is 3. This solution will add position labels to each slot in sequence from back to front. For example, the three empty position labels correspond to "one", "ten", and "hundred" respectively.

S2424: Determine the corresponding position label and empty space according to the position word in each numerical value group, and fill the numerical words in the numerical value group into the determined empty position.

Finally, this solution determines the corresponding position labels and gaps based on the position words in each numerical group, and fills the numerical words in the numerical group into the determined gaps.

For example, fill "2" into the empty space corresponding to "hundred" and fill "3" into the empty space corresponding to "tens".

S243, if it is determined that there are unfilled vacancies in the first conversion template, fill the unfilled vacancies with 0s.

For example, during the above filling process, the ones bit has not been filled. At this time, this solution will fill the unfilled empty bits with 0, and finally obtain the data of 230.

S244, if it is determined that all the gaps in the first conversion template have been filled, then use the number composed of all the gaps as the first word that meets the format requirements.

If it is determined that all the gaps in the first conversion template have been filled, then the number composed of all the gaps will be used as the first word that meets the format requirements.

It should be noted that for the rare data with a position word at the end, this solution can convert the position word at the end accordingly. For example, for "two million and three hundred thousand", the "ten thousand" at the end can be converted into four zeros, and finally the conversion data of 2,300,000 is obtained.

S3: Generate a corresponding data requirement list based on the user's configured computing requirements for this power report, and select the first word with the corresponding analysis attribute as the third word based on the data requirement list.

Because users have different needs, the data they need to analyze is also different. This solution will generate a corresponding data requirement list based on the calculation requirements configured by the user for this power report.

After obtaining the data requirement list, the first word with the corresponding analysis attribute can be selected as the third word in combination with the data requirement list.

In some embodiments, S3 (generating a corresponding data requirement list based on the computing requirements configured by the user for this power report, and selecting the first word with the corresponding analysis attribute as the third word based on the data requirement list) includes S31 -S33:

S31: Generate a corresponding data statistics table based on all the first words in the preset format, the merged first words, and the respectively associated second words.

First, this solution will count all the first words in the preset format, the merged first words, and the respectively associated second words, and generate a corresponding data statistics table.

S32: Generate a corresponding data requirement list based on the calculation requirements configured by the user for this power report, and extract the fourth word in the data requirement list.

In some cases, the user may have multiple computing requirements. In this case, corresponding data statistics tables can be generated for all the first words in the preset format, the merged first words, and the respectively associated second words.

The calculation requirement is, for example, the calculation of residential electricity consumption, industrial electricity consumption, etc., and the corresponding fourth word may be residential electricity consumption, industrial electricity consumption, etc.

S33: Compare the fourth word with the second word in the data statistics table. If the fourth word corresponds to the second word, use the first word corresponding to the corresponding second word in the data statistics table as the third word.

This solution will compare the fourth word with the second word in the data statistics table. If the fourth word corresponds to the second word, it means that there is corresponding data in the data statistics table. At this time, this solution will use the first word corresponding to the corresponding second word in the data statistics table as the third word.

S4: Determine the cleaning verification strategy corresponding to the third word according to the analysis attribute of the third word, and verify the correctness of the third word based on the cleaning verification strategy and historical data. The analysis attributes of each type Has preset cleaning verification strategies.

After obtaining the third word, this solution will determine the cleaning and verification strategy corresponding to the third word based on the analysis attribute of the third word. Each type of analysis attribute has a preset cleaning and verification strategy.

Then, combine the cleaning verification strategy and historical data to verify the correctness of the third word. It is understandable that the statistical data in the power report may be inaccurate. For example, the data obtained is that residents’ electricity consumption is 1 kilowatt hour. This data must be wrong. Therefore, before analyzing it, it is still necessary to The data correctness will be verified.

In some embodiments, S4 (determine the cleaning verification strategy corresponding to the third word based on the analysis attribute of the third word, and verify the correctness of the third word based on the cleaning verification strategy and historical data, Each type of analysis attribute has preset cleaning validation strategies) including S41-S44:

S41. Determine the second word with the analysis attribute corresponding to the third word, and determine the corresponding cleaning verification strategy according to the corresponding second word. The second word with each type of subject analysis attribute or concept analysis attribute has a preset cleaning verification Strategy.

For example, if the subject analysis attribute is residential electricity consumption, the preset cleaning verification strategy may be to set an electricity consumption interval, that is, a first threshold interval. Subsequently, it is determined whether the residents' electricity consumption is within the first threshold interval and verified.

S42: Compare the third word with the first threshold interval included in the cleaning verification strategy. If the value of the third word is within the first threshold interval, determine whether the corresponding third word satisfies the correctness of the cleaning verification strategy. verification.

First, this solution will compare the third word with the first threshold interval included in the cleaning verification strategy. It can be understood that if the value of the third word is within the first threshold interval, it is determined that the corresponding third word meets the verification of the correctness of the cleaning verification strategy.

It is worth mentioning that the above steps are to verify whether the data is reasonable and are definitely wrong.

S43: Determine corresponding historical data according to the subject analysis attribute or concept analysis attribute of the second word, calculate an average value based on the historical data, and perform interval processing on the average value according to a preset proportion value to obtain the third Two threshold intervals.

It can be understood that this step is to obtain the corresponding historical data and judge it based on the historical data.

After the average value is obtained by combining the historical data, the interval processing can be performed by combining the preset proportion value to obtain the second threshold interval. For example, if the preset proportion value is 0.5 and the average value is 1000, then the second threshold interval may be 500-1500.

It is worth mentioning that the above steps are to verify whether the data is doubtful. If it is outside the second threshold interval, it is not necessarily wrong and is doubtful.

S44: If the value of the third word is within the second threshold interval, it is determined that the corresponding third word meets the verification of the correctness of the historical data.

If the value of the third word is within the second threshold interval, this solution will determine that the corresponding third word meets the verification of the correctness of historical data.

S5, if it is determined that the verification fails, determine the verification doubt type corresponding to the verification failure, and determine the cleaning processing strategy corresponding to the third word according to the verification doubt type of the third word. The verification doubt type includes an unreasonable doubt type. or questionable types to be verified.

In some embodiments, S5 (if it is determined that the verification fails, determine the verification question type corresponding to the verification failure, and determine the cleaning processing strategy corresponding to the third word according to the verification question type of the third word, and the verification question type Types include unreasonable doubt types or doubtful types to be verified) including S51-S55:

S51, if it is determined that the value of the third word is not within the first threshold interval, determine that the verification doubt type corresponding to the third word is an unreasonable doubt type.

It can be understood that if it is determined that the value of the third word is not within the first threshold interval, it means that the data must be wrong, and this solution will determine that the verification doubt type corresponding to the third word is an unreasonable doubt type.

S52: If it is determined that the value of the third word is not within the second threshold interval, determine that the verification questionable type corresponding to the third word is the questionable type to be verified.

It can be understood that if it is determined that the value of the third word is not within the second threshold interval, it means that the data is doubtful, and this solution will determine that the verification doubtful type corresponding to the third word is the doubtful type to be verified.

S53, if the third word is of an unreasonable and doubtful type, mark the corresponding third word as an error, and output the third word and the corresponding second word so that the user can directly update the third word.

If the third word is of an unreasonable and doubtful type, the corresponding third word is marked as an error, and the third word and the corresponding second word are output, so that the user can directly update the third word.

S54, if the third word is of a questionable type to be verified, mark the corresponding third word as to be verified, and output the third word and the corresponding second word. If it is judged that the user inputs positive verification information, the corresponding The corresponding third word serves as the final third word.

If the third word is a questionable type to be verified, mark the corresponding third word as to be verified, and output the third word and the corresponding second word to assist the user in further judgment. If it is judged that the user inputs a positive verification The information uses the corresponding third word as the final third word.

S55, if it is determined that the user input is negative verification information, update the corresponding third word according to the user input.

If it is determined that the user input is negative verification information, the corresponding third word is updated according to the user input.

S6: After determining all the third words corresponding to the data requirement list, input the third words into the preset model to calculate the analyzed data after data cleaning, and obtain the corresponding analysis data based on the analyzed data and the knowledge map. Data processing.

This solution is equipped with a preset model. The third words can be input into the preset model to calculate the analyzed data after data cleaning, and then the corresponding processed data can be obtained by combining the analyzed data and the knowledge graph.

In some embodiments, S6 (after determining that all third words corresponding to the data requirement list are obtained, input the third words into the preset model to calculate the analyzed data after data cleaning, based on the analyzed data and The knowledge graph obtains the corresponding processing data) including S61-S63:

S61: After determining that all third words corresponding to the data requirement list are obtained, the third words are input into the preset model to calculate analyzed data after data cleaning, and the analyzed data includes corresponding segmented words.

After this solution determines and obtains all the third words corresponding to the data requirement list, the third words can be input into the preset model to calculate the analyzed data after data cleaning, where the analyzed data includes the corresponding segmented words.

For example, a preset model can be used to compare residential electricity consumption data and industrial electricity consumption data to obtain analytical data. For example, the analytical data can be that industrial electricity consumption accounts for a relatively high proportion. Finally, segmented words with a high proportion of industrial electricity consumption are obtained.

S62: Input the segmented words into the knowledge graph to determine knowledge nodes with corresponding first knowledge information, and based on the determined knowledge nodes and other knowledge nodes with second knowledge information.

This solution will input segmented words into the knowledge graph to determine knowledge nodes with corresponding first knowledge information, and based on the determined knowledge nodes and other knowledge nodes with second knowledge information.

The second knowledge node may be a processing node corresponding to the first knowledge node. For example, for the first knowledge node with a high proportion of industrial electricity consumption, the corresponding second knowledge node may be "limit industrial electricity consumption", "It is recommended that factories increase the proportion of clean energy power supply" and other nodes.

S63: Count the second knowledge information of other knowledge nodes with second knowledge information, obtain corresponding processing data and output it.

Finally, the second knowledge information of other knowledge nodes with second knowledge information is counted, and the corresponding processed data is obtained and output. It can be understood that, for a first knowledge node, it may have multiple second knowledge nodes, and this solution will output the corresponding multiple second knowledge information as processing data.

Refer to Figure 2, which is a schematic structural diagram of an intelligent cleaning and processing system for power reports that combines knowledge graphs and semantic analysis provided by an embodiment of the present invention. The intelligent cleaning and processing system for power reports that combines knowledge graphs and semantic analysis includes:

The construction module is used to receive the knowledge information configured by the user for intelligent cleaning and processing of power reports, and build a corresponding knowledge graph based on the triple relationship corresponding to the corresponding knowledge information. The knowledge graph includes multiple knowledge nodes;

The processing module is used to perform word segmentation processing on the statements in the power report to obtain multiple words, determine the words with data attributes in the words as the first words, use other words as the second words, and combine the second word pairs in the sentences. Perform semantic analysis on the corresponding first word to obtain the analytical attributes of the first word;

A generation module configured to generate a corresponding data requirement list based on the user's configured computing requirements for this power report, and select the first word with the corresponding analysis attribute as the third word based on the data requirement list;

A verification module, configured to determine the cleaning verification strategy corresponding to the third word based on the analytical attributes of the third word, and verify the correctness of the third word based on the cleaning verification strategy and historical data. Each type The analysis attributes have preset cleaning verification strategies;

A judgment module, configured to determine the verification doubt type corresponding to the verification failure if it is judged that the verification fails, and determine the cleaning processing strategy corresponding to the third word according to the verification doubt type of the third word, and the verification doubt type includes not Types of reasonable doubts or types of doubts to be verified;

The calculation module is used to, after determining and obtaining all third words corresponding to the data requirement list, input the third words into the preset model to calculate the analyzed data after data cleaning, which is obtained based on the analyzed data and the knowledge map. Process the data accordingly.

The present invention also provides a storage medium. A computer program is stored in the storage medium. When the computer program is executed by a processor, the computer program is used to implement the methods provided by the above-mentioned various embodiments.

The storage medium may be a computer storage medium or a communication medium. Communication media includes any medium that facilitates transfer of a computer program from one place to another. Computer storage media can be any available media that can be accessed by a general purpose or special purpose computer. For example, a storage medium is coupled to a processor such that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be an integral part of the processor. The processor and storage medium may be located in application specific integrated circuits (Application Specific Integrated Circuits, ASIC for short). Additionally, the ASIC can be located in the user equipment. Of course, the processor and the storage medium may also exist as discrete components in the communication device. Storage media can be read-only memory (ROM), random access memory (RAM), CD-ROM, tapes, floppy disks, optical data storage devices, etc.

The present invention also provides a program product, the program product includes execution instructions, and the execution instructions are stored in a storage medium. At least one processor of the device can read the execution instruction from the storage medium, and at least one processor executes the execution instruction so that the device implements the methods provided by the various embodiments mentioned above.

In the above embodiments of the terminal or server, it should be understood that the processor may be a central processing unit (English: Central Processing Unit, referred to as: CPU), or other general-purpose processors, digital signal processors (English: Digital Signal Processor (abbreviation: DSP), application specific integrated circuit (English: Application Specific Integrated Circuit, abbreviation: ASIC), etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The steps of the method disclosed in the present invention can be directly implemented by a hardware processor, or executed by a combination of hardware and software modules in the processor.

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, but not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions described in the foregoing embodiments can still be modified, or some or all of the technical features can be equivalently replaced; and these modifications or substitutions do not deviate from the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention. scope.

Claims (12)

1. An intelligent cleaning processing method for power reports that combines knowledge graph and semantic analysis, which is characterized by: Receive the knowledge information configured by the user for intelligent cleaning and processing of the power report, and construct a corresponding knowledge graph based on the triplet relationship corresponding to the corresponding knowledge information, and the knowledge graph includes multiple knowledge nodes; Perform word segmentation processing on the sentences in the power report to obtain multiple words, determine the words with data attributes in the words as the first words, use other words as the second words, and combine the second words in the sentences to the corresponding first words Perform semantic analysis to obtain the analytical attributes of the first word; Generate a corresponding data requirement list based on the user's configured calculation requirements for this power report, and select the first word with the corresponding analysis attribute as the third word based on the data requirement list; The cleaning verification strategy corresponding to the third word is determined according to the analysis attribute of the third word, and the correctness of the third word is verified based on the cleaning verification strategy and historical data. Each type of analysis attribute has a predetermined Designed cleaning and verification strategies; If it is determined that the verification fails, the verification doubt type corresponding to the verification failure is determined, and the cleaning processing strategy corresponding to the third word is determined according to the verification doubt type of the third word, and the verification doubt type includes an unreasonable doubt type or an unreasonable doubt type. Verify doubtful types; After determining that all third words corresponding to the data requirement list are obtained, the third words are input into the preset model to calculate the analyzed data after data cleaning, and the corresponding processing data is obtained based on the analyzed data and the knowledge graph. . 2. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 1, characterized in that, The receiving user intelligently cleans and processes the configured knowledge information for power reports, and constructs a corresponding knowledge graph based on the triplet relationship corresponding to the corresponding knowledge information. The knowledge graph includes multiple knowledge nodes, including: The knowledge information includes first knowledge information corresponding to the analysis data, and second knowledge information corresponding to the processing data, and each first knowledge information or second knowledge information has a corresponding knowledge node; The corresponding knowledge nodes are connected according to the triplet relationship configured by the user on the first knowledge information and the second knowledge information, and a corresponding knowledge graph is constructed and generated. 3. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 2, characterized in that, Perform word segmentation processing on the sentences in the power report to obtain multiple words, determine the words with data attributes in the words as the first words, use other words as the second words, and combine the second words in the sentences to the corresponding third words. Perform semantic analysis on a word to obtain the analytical attributes of the first word, including: Perform word segmentation processing on the statements in the power report to obtain multiple words, and determine the words with data attributes among the words as the first words. The words with data attributes at least include Arabic numerals, uppercase numerals, and traditional Chinese numerals; Use words other than the first word in all participles as second words, traverse the second words and compare them with the preset words. If it is determined that the second words correspond to the preset words, determine the second words based on the preset words. is the second word to be analyzed; According to the positional relationship between each first word and the second word to be analyzed, the second word to be analyzed associated with the corresponding first word is determined, and the corresponding first word is semantically analyzed based on the associated second word. , get the analytical attributes of the first word; If it is determined that the format of the first word is inconsistent with the preset format, a preset first conversion template is determined according to the format of the first word, and the first word is converted based on the first conversion template to obtain a format that meets the format requirements. First word. 4. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 3, characterized in that, Determining the second word associated with the corresponding first word based on the positional relationship between each first word and the second word to be analyzed, and performing semantic analysis on the corresponding first word based on the associated second word, Get the analytical attributes of the first word, including: If it is determined that there are multiple first words and preset merged words in a sentence, it is determined that the corresponding first words can be merged into one first word, and corresponding merge tags are added to the first words that can be merged, so that subsequent When processing the first word, merge the corresponding first word based on the merge tag; If it is determined that a sentence contains a first word or multiple first words that can be combined into one first word, then all the second words to be analyzed in the corresponding sentence are combined with the corresponding first words or the merged first words. Associated; If it is determined that there are multiple first words in a sentence and the multiple words cannot be combined into one first word, segment the sentence based on the position of the first word to obtain a segmentation result, and determine each first word based on the segmentation result. A second word to which one word is associated; The analysis attributes are any one or more of subject analysis attributes, trend change analysis attributes, and concept analysis attributes included in the second word. 5. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 4, characterized in that, If it is determined that there are multiple first words in a sentence and the multiple words cannot be combined into one first word, segment the sentence based on the position of the first word to obtain a segmentation result, and determine each segmentation result based on the segmentation result. second words associated with the first word, including: Determine the positions of all first words in the sentence, segment the sentence based on the positions of the first words to obtain multiple sub-segments, and determine the front and adjacent segments of each first word as associated segments; The second word in the associated segment is used as the second word associated with the corresponding first word. 6. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 4, characterized in that, If it is determined that the format of the first word is inconsistent with the preset format, a preset first conversion template is determined according to the format of the first word, and the first word is converted based on the first conversion template to obtain a format that satisfies the format. The first word required includes: Compare the format of the first word with a preset format, the format of the first word is Arabic numeral format, uppercase numeral format or traditional Chinese numeral format, and the preset format is Arabic numeral format; If it is determined that the format of the first word is inconsistent with the preset format, the preset first conversion template is determined according to the format of the first word, and the first word is decomposed to obtain the corresponding numerical word and position word, based on the position word Fill the numerical words into the relative empty positions in the first conversion template; If it is determined that there are unfilled vacancies in the first conversion template, the unfilled vacancies are filled with 0; If it is determined that all the gaps in the first conversion template have been filled, then the number composed of all the gaps will be used as the first word that meets the format requirements. 7. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 6, characterized in that, If it is determined that the format of the first word is inconsistent with the preset format, the preset first conversion template is determined according to the format of the first word, and the first word is decomposed to obtain the corresponding numerical word and position word, based on The position word fills the numerical word into the relative empty position in the first conversion template, including: Determine a preset first conversion template according to the format of the first word. The first conversion template has a plurality of gaps. The first word is decomposed and identified to obtain the corresponding numerical word and position word. Each numerical word and its other The adjacent position words at the rear are treated as a numerical value group; Extract the frontmost position word in the entire first word to obtain the corresponding preset number of gaps, and each position word has a corresponding preset number of gaps; Reserve a preset number of vacancies in the first conversion template, and add position labels to each vacancy in sequence from back to front according to the reserved vacancies; The corresponding position labels and gaps are determined according to the position words in each numerical group, and the numerical words in the numerical group are filled into the determined gaps. 8. The intelligent cleaning processing method for electric power reports combining knowledge graph and semantic analysis according to claim 6, characterized in that it further includes: If it is determined that there are corresponding merge tags between multiple first words, the multiple first words are calculated according to the corresponding merge tags between the multiple first words to obtain the calculated first words, each merged Words have a preset calculation method. 9. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 6, characterized in that, The method generates a corresponding data demand list based on the user's configured computing requirements for this power report, and selects the first word with corresponding analysis attributes as the third word based on the data demand list, including: Generate corresponding data statistics tables based on all the first words in the preset format, the merged first words, and the respectively associated second words; Generate a corresponding data requirement list based on the calculation requirements configured by the user for this power report, and extract the fourth word in the data requirement list; The fourth word is compared with the second word in the data statistics table. If the fourth word corresponds to the second word, the first word corresponding to the corresponding second word in the data statistics table is used as the third word. 10. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 9, characterized in that, The cleaning verification strategy corresponding to the third word is determined based on the analysis attribute of the third word, and the correctness of the third word is verified based on the cleaning verification strategy and historical data. The analysis attributes of each type are Has preset cleaning verification strategies, including: Determine the second word corresponding to the analysis attribute of the third word, and determine the corresponding cleaning verification strategy according to the corresponding second word. The second words of each type of subject analysis attribute or concept analysis attribute have a preset cleaning verification strategy; The third word is compared with the first threshold interval included in the cleaning verification strategy. If the value of the third word is within the first threshold interval, it is determined that the corresponding third word satisfies the verification of the correctness of the cleaning verification strategy. ; Corresponding historical data is determined according to the subject analysis attribute or concept analysis attribute of the second word, the average value is calculated according to the historical data, and the average value is interval processed according to the preset proportion value to obtain the second threshold. interval; If the value of the third word is within the second threshold interval, it is determined that the corresponding third word meets the verification of the correctness of the historical data. 11. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 10, characterized in that, If it is determined that the verification fails, the verification doubt type corresponding to the verification failure is determined, and the cleaning processing strategy corresponding to the third word is determined according to the verification doubt type of the third word, and the verification doubt type includes an unreasonable doubt type. or questionable types to be verified, including: If it is determined that the value of the third word is not within the first threshold interval, it is determined that the verification doubt type corresponding to the third word is an unreasonable doubt type; If it is determined that the value of the third word is not within the second threshold interval, then it is determined that the verification questionable type corresponding to the third word is the questionable type to be verified; If the third word is of an unreasonable and doubtful type, mark the corresponding third word as an error, and output the third word and the corresponding second word so that the user can directly update the third word; If the third word is of a questionable type to be verified, the corresponding third word is marked as to be verified, and the third word and the corresponding second word are output. If it is judged that the user inputs positive verification information, the corresponding The third word serves as the final third word; If it is determined that the user input is negative verification information, the corresponding third word is updated according to the user input. 12. The intelligent cleaning processing method for power reports combining knowledge graph and semantic analysis according to claim 11, characterized in that, After it is determined that all third words corresponding to the data requirement list are obtained, the third words are input into the preset model to calculate the analyzed data after data cleaning, and the corresponding analysis data is obtained based on the analyzed data and the knowledge map. Process data including: After determining that all third words corresponding to the data requirement list are obtained, the third words are input into the preset model to calculate the analyzed data after data cleaning, and the analyzed data includes the corresponding segmented words; Input the segmented words into the knowledge graph to determine knowledge nodes with corresponding first knowledge information, and based on the determined knowledge nodes and other knowledge nodes with second knowledge information; The second knowledge information of other knowledge nodes with second knowledge information is counted, and the corresponding processing data is obtained and output.