A dynamic matching method for multi-semantic text representation

Abstract

The application relates to a dynamic matching method for multi-semantics text expression, which comprises the following steps: obtaining a first industrial control text input by a user currently; obtaining a second industrial control text related to the user; determining a keyword according to the first industrial control text and the second industrial control text; forming a multi-semantics vector based on the keyword; and matching the expression of the first industrial control text based on the multi-semantics vector. The method provided by the application forms a multi-semantics vector based on the industrial control text input by the user currently and the related industrial control text, matches the expression of the first industrial control text based on the multi-semantics vector, realizes accurate matching of the industrial control text expression, and enables the industrial control system to accurately identify a control instruction and realize accurate control.

Description

Technical Field

[0001] This invention relates to the field of natural speech processing technology, and in particular to a dynamic matching method for multi-semantic text expressions. Background Technology

[0002] With the rapid development of computer technology, its widespread application has brought great convenience to people. In industrial applications, people can use computer technology to control industrial systems, especially in the post-processing of text data. By processing text information, the corresponding semantic information of the text can be effectively identified, and then the text information can be directly processed to achieve industrial control.

[0003] However, current industrial control systems simply identify and process text information through keywords, which is not accurate enough and often leads to malfunctions in the industrial control system, thus affecting the industrial process. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] In view of the above-mentioned shortcomings and deficiencies of the prior art, the present invention provides a dynamic matching method for multi-semantic text expressions.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the main technical solutions adopted by the present invention include:

[0008] A dynamic matching method for multi-semantic text representations, the method comprising:

[0009] S101, Obtain the first industrial control text currently input by the user;

[0010] S102, Obtain the second industrial control text related to the user;

[0011] S103, determine keywords based on the first industrial control text and the second industrial control text;

[0012] S104, forming multi-semantic vectors based on keywords;

[0013] S105, based on multi-semantic vector matching to express the first industrial control text.

[0014] Optionally, S102 includes:

[0015] S201, determine whether the user has entered industrial control text within a preset time period;

[0016] S202, If industrial control text is input within a preset time period, all input industrial control text will be identified as the target text;

[0017] If no industrial control text is entered within the preset time period, determine whether the user has entered industrial control text.

[0018] If the user has entered industrial control text, the target text will be selected from the most recent to the oldest input time.

[0019] If the user has not entered industrial control text, the target text will be selected based on the user's attributes;

[0020] S203, Obtain the second industrial control text based on the target text.

[0021] Optionally, the target text can be selected from most recent to oldest input time, including:

[0022] Based on the input time from most recent to oldest, the industrial control text entered by the user is used as elements to form a historical text collection;

[0023] Determine the difference between the input times of two adjacent elements in the historical text set to form a difference set D, where the difference is the time of the previous element minus the time of the next element;

[0024] Calculate the standard deviation δ of all differences;

[0025] Choose the first and second elements in the difference set D;

[0026] Calculate the first value Where j is the element identifier in the selection difference set D. To select the average of the element values ​​in the difference set D, n d To determine the total number of elements in the difference set D;

[0027] If the first value Δ1 is not less than δ, then the element involved in the historical text set of the currently selected element is selected as the target text;

[0028] If the first value Δ1 is less than δ, then select the first, second and third elements of the difference set D, and re-execute the steps to calculate the first value and subsequent steps. Repeat this process, selecting one more subsequent element from the difference set D in each iteration, until the currently calculated first value Δ1 is not less than δ.

[0029] Optionally, the target text can be selected based on the user's attributes, including:

[0030] The user's attributes are formed into a user attribute set U0;

[0031] Form the attributes of other users into their respective attribute sets U. iWhere i is the identifier of other users;

[0032] Calculate the first similarity between the user and each other user. Among them, w i The weights for other users i;

[0033] According to Sim 0,i Select the target text.

[0034] Optionally, for any other user i, w i Determine this through the following steps:

[0035] S301, retrieve all third-party industrial control text entered by other user i after each recent login;

[0036] S302, for any third-party industrial control text entered by other user i after any login x.

[0037] Sort all third-party industrial control texts entered after any login x in chronological order from most recent to oldest, forming the x set;

[0038] Determine the core keywords for each element in set x;

[0039] The core keywords of each element are concatenated to obtain the vector h. x ;

[0040] S303, take the vector h after each login in sequence. x Calculate its vector h after each login and other logins. y The second similarity between them;

[0041] S304, the minimum second similarity is determined as w i .

[0042] Optionally, S203 includes:

[0043] Define all target text as industrial control text; or,

[0044] The most recent preset number of target texts are identified as the second industrial control text.

[0045] Optionally, S103 includes:

[0046] S103-1, perform word segmentation on the first industrial control text and the second industrial control text;

[0047] S103-2, semantically expand the obtained word segments to obtain expanded words;

[0048] S103-3, the obtained word segments and expanded words are combined into a word segmentation set;

[0049] S103-4, for each element u in the word segmentation set,

[0050] Calculate the third similarity Sim between element u and each word in the industrial dictionary. u,a , where 'a' is the identifier of the word in the industrial dictionary;

[0051] If there is no third similarity greater than the preset similarity threshold, then element u will be removed from the word segmentation set;

[0052] If there is a third similarity greater than the preset similarity threshold, then the word corresponding to the third similarity greater than the preset similarity threshold in the industrial dictionary will be used as an attribute of element u.

[0053] S103-5, determine the keywords based on the attributes of each element in the current word segmentation set.

[0054] Optionally, S103-5 includes:

[0055] Summarize the attributes of each element in the current word segmentation set, and remove duplicate attributes to obtain the attribute set;

[0056] Determine the second value of each element in the attribute set. Where b is the element identifier in the attribute set. n is the number of elements in the current word segmentation set whose attribute includes element b. (2) n represents the total number of elements in the current word segmentation set. (3) w represents the total number of the first and second industrial control texts. b Let b be the weight of element b.

[0057] Select the attribute whose second value is greater than the preset threshold;

[0058] In the current word segmentation set, all elements corresponding to the selection attribute and the words corresponding to the selection attribute are identified as keywords.

[0059] Optionally, w b The method for determining it is as follows:

[0060] S401, Get the elements in the current word segmentation set whose attributes include element b;

[0061] S402, determine the industrial control text in which the acquired element is located, wherein the industrial control text is the first industrial control text and / or the second industrial control text;

[0062] S403, for each industrial control text.

[0063] Determine the number of word segments included that belong to the word segment set. Where 'c' represents the industrial control text identifier;

[0064] Determine the total number of word segments included.

[0065] S404, Confirm Where, n (5) For each industrial control text included, n (6) This represents the total number of industrial control texts in the relevant sector.

[0066] Optionally, S105 includes:

[0067] Semantic recognition of multiple semantic vectors;

[0068] The expression of the first industrial control text is matched based on the semantic recognition results.

[0069] (III) Beneficial Effects

[0070] The method provided by this invention obtains the first industrial control text currently input by the user; obtains the second industrial control text related to the user; determines keywords based on the first and second industrial control texts; forms a multi-semantic vector based on the keywords; and matches the expression of the first industrial control text based on the multi-semantic vector. This achieves accurate matching of the industrial control text expression, enabling the industrial control system to accurately identify control commands and achieve precise control. Attached Figure Description

[0071] Figure 1 This is a flowchart illustrating a dynamic matching method for multi-semantic text expressions provided in an embodiment of the present invention. Detailed Implementation

[0072] To better explain and facilitate understanding of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0073] Current industrial control systems simply identify and process text information through keywords, which is not accurate enough and often leads to malfunctions in the industrial control system, thus affecting industrial processes.

[0074] Based on this, the present invention provides a dynamic matching method for multi-semantic text expressions. The method includes: acquiring a first industrial control text currently input by the user; acquiring a second industrial control text related to the user; determining keywords based on the first and second industrial control texts; forming a multi-semantic vector based on the keywords; and matching the expression of the first industrial control text based on the multi-semantic vector. The method provided by the present invention forms a multi-semantic vector based on the user's currently input industrial control text and related industrial control texts, and matches the expression of the first industrial control text based on the multi-semantic vector, achieving accurate matching of industrial control text expressions. This enables industrial control systems to accurately identify control commands and achieve precise control.

[0075] See Figure 1 The implementation process of the dynamic matching method for multi-semantic text expressions provided in this embodiment is as follows:

[0076] S101, Obtain the first industrial control text currently input by the user.

[0077] The first industrial control text in this step is actually an industrial control text. In order to distinguish it from related industrial control texts, the industrial control text in this step is named the first industrial control text. The word "first" is only used for distinction and has no other substantive meaning.

[0078] Additionally, the industrial control text in this step is entered by the user.

[0079] S102, Obtain the second industrial control text related to the user.

[0080] Recognizing the first industrial control text in step S101 alone may result in inaccurate recognition. Therefore, this step will acquire other industrial control texts related to the user and use these other industrial control texts to correct the first control text, thereby improving the recognition accuracy.

[0081] The second industrial control text in this step is actually an industrial control text. In order to distinguish it from the industrial control text in step S101, the industrial control text in this step is named the second industrial control text. The "second" in the text is only used for distinction and has no other substantive meaning.

[0082] The implementation process of this step is as follows:

[0083] S201, determine whether the user has entered industrial control text within a preset time period.

[0084] For example, determining whether a user entered industrial control text within the last half hour. The closer the time, the greater the correlation between two industrial control texts, and the greater the contribution to accurate identification.

[0085] The preset time period is an empirical value, and the method for determining it is not limited in this embodiment.

[0086] S202,

[0087] 1-1 If industrial control text is entered within a preset time period, all entered industrial control text will be identified as the target text.

[0088] 1-2 If no industrial control text is entered within the preset time period, determine whether the user has entered industrial control text.

[0089] The work content of the same user is relatively stable. Based on the industrial control text that the user has previously entered, the first industrial control text obtained in step S101 can be guided in a directional manner. Therefore, if no industrial control text is entered within a preset time period, the accuracy of the final recognition will be improved based on the industrial control text that the user has entered.

[0090] 1-2-1 If the user has entered industrial control text, select the target text in order of the most recent to the oldest input time.

[0091] The process of selecting target text from most recent to oldest based on the input time is as follows:

[0092] 1. Based on the input time from most recent to oldest, the system uses the industrial control text entered by the user as elements to form a historical text collection.

[0093] For example, if the industrial control text 1 is entered at 10:20, the industrial control text 2 is entered at 10:19, the industrial control text 3 is entered at 5:19, and the industrial control text 4 is entered at 5:18, then industrial control text 1 is taken as element 1, industrial control text 2 as element 2, industrial control text 3 as element 3, and industrial control text 4 as element 4, forming a historical text set, that is, historical text set = {element 1, element 2, element 3, element 4}.

[0094] 2. Determine the difference in input time between two adjacent elements in the historical text set to form the difference set D.

[0095] The difference is the time of the previous element minus the time of the next element.

[0096] For example, the difference between element 1 and element 2 is the input time of element 1 (10:20) - the input time of element 2 (10:19) = 1 minute. The difference between element 2 and element 3 is the input time of element 2 (10:19) - the input time of element 3 (5:19) = 5 hours. The difference between element 3 and element 4 is the input time of element 3 (5:19) - the input time of element 4 (5:18) = 1 minute.

[0097] D = {1 minute, 5 hours, 1 minute}.

[0098] It should be noted that, for ease of calculation, the units can be standardized, for example, all differences can be converted to minutes, seconds, hours, or days, etc. This embodiment does not limit the standardized units.

[0099] For example, D = {1 / 60 hour, 5 hours, 1 / 60 hour}.

[0100] 3. Calculate the standard deviation δ of all differences.

[0101] The standard deviation here is the existing standard deviation calculation scheme, which will not be elaborated in detail.

[0102] 4. Select the first and second elements in the difference set D.

[0103] For example, select element 1 and element 2.

[0104] 5. Calculate the first value

[0105] Where j is the element identifier in the selection difference set D. To select the average of the element values ​​in the difference set D, n d To determine the total number of elements in the difference set D.

[0106] For example, when choosing element 1 and element 2, the total number of elements in the difference set D is 2 (i.e., n). d =2), the average value of the elements in the difference set D is [1 / 60 (hours) + 5 (hours)] / 2 (i.e. The first value is

[0107] 6. If the first value Δ1 is not less than δ, then select the element involved in the historical text set of the currently selected element as the target text.

[0108] δ represents the dispersion of all differences, that is, the tightness between all differences, while the difference represents the relationship between input times. In other words, δ represents the tightness of the time interval between industrial control text inputs. The larger δ is, the larger the time interval between industrial control text inputs (inputs over a long period of time), and the smaller the δ is, the smaller the time interval between industrial control text inputs (inputs over a short period of time). If Δ1 is not less than δ, it means that the overall time interval of the currently selected element is greater than the overall interval. The smaller the time interval, the greater the correlation between two consecutive industrial control texts. Therefore, the continuous industrial control text with the highest correlation has been found, and the element involved in the historical text set of the currently selected element is selected as the target text.

[0109] 7. If the first value Δ1 is less than δ, select the first, second and third elements in the difference set D, and re-execute the steps to calculate the first value and subsequent steps. Repeat this process, selecting one more subsequent element from the difference set D in each iteration, until the currently calculated first value Δ1 is not less than δ.

[0110] For example, if If the value is less than δ, it means that the overall input time interval of the currently selected element is less than the overall interval. The smaller the time interval, the greater the correlation between the two consecutive industrial control texts. Therefore, the most correlated consecutive industrial control text has not been found at this time. Then, the next difference (such as the difference between element 3 and element 4, 1 / 60 (hour)) is added in and the calculation is repeated until the most correlated consecutive industrial control text is found.

[0111] 1-2-2 If the user has not entered industrial control text, select the target text based on the user's attributes.

[0112] The process of selecting target text based on user attributes is as follows:

[0113] 1. Form the user's attributes into a user attribute set U0.

[0114] The attributes of this step, such as the work content and level, can describe the attributes of the industrial control content performed by the user.

[0115] 2. Combine the attributes of other users into their respective attribute sets U. i .

[0116] Where i represents the identifier of other users.

[0117] The attribute items here are the same as the user's attribute items, but the attribute values ​​may be the same or different.

[0118] For example, if a user's attributes include job content and level, then other users' attributes will also include job content and level.

[0119] 3. Calculate the first similarity between the user and each other user.

[0120] Among them, w i Let i be the weight of other user i.

[0121] For any other user i, w i Determine this through the following steps:

[0122] S301, retrieve all third-party industrial control text entered by other user i after each recent login.

[0123] S302, for any third-party industrial control text entered by other user i after any login x.

[0124] 1) Sort all third-party industrial control texts entered after any login of x in chronological order from most recent to oldest, forming the x set.

[0125] 2) Determine the core keywords of each element in set x.

[0126] 3) Concatenate the core keywords of each element to obtain the vector h. x .

[0127] After executing step S302, the vectors of each login attempt of other user i will be obtained.

[0128] S303, take the vector h after each login in sequence. x Calculate its vector h after each login and other logins. y The second similarity between them.

[0129] The similarity calculation scheme between vectors adopts the existing scheme, which will not be elaborated here.

[0130] For example, if another user i logs in 3 times, in step S303, the vector h1 from the first login is taken, and the second similarity between h1 and h2 from the second login, and the second similarity between h1 and h3 from the third login are calculated. Then, the vector h2 from the second login is taken, and the second similarity between h2 and h1, and the second similarity between h2 and h3 are calculated. Finally, the vector h3 from the third login is taken, and the second similarity between h3 and h1, and the second similarity between h3 and h2 are calculated.

[0131] S304, the minimum second similarity is determined as w i .

[0132] Of all the second similarities obtained in step S303 (i.e., the second similarity between h1 and h2, the second similarity between h1 and h3, the second similarity between h2 and h1, the second similarity between h2 and h3, the second similarity between h3 and h1, and the second similarity between h3 and h2), the smallest second similarity is determined as w. i .

[0133] 4. According to Sim 0,i Select the target text.

[0134] For example, select Sim 0,i The largest or one other user will select other users to input all or part of the third control text as the target text.

[0135] This embodiment does not limit the method of determining the target text based on the third control text entered by other selected users. For example, all third control text entered by other selected users can be used as the target text, or all third industrial control text entered by other selected users in their most recent login can be used as the target text, etc.

[0136] The final number of other users selected can be determined based on experience, and this implementation does not limit the method for determining it.

[0137] S203, Obtain the second industrial control text based on the target text.

[0138] For example, all target texts can be designated as the second industrial control text. Alternatively, a preset number of target texts can be designated as the second industrial control text.

[0139] The preset quantity is an empirical value, and this embodiment does not limit the determination method.

[0140] S103, Based on the first industrial control text and the second industrial control text, determine the keywords.

[0141] The implementation process of step S103 is as follows:

[0142] S103-1, perform word segmentation on the first industrial control text and the second industrial control text.

[0143] This step is implemented using existing word segmentation schemes, which will not be elaborated here.

[0144] S103-2, semantically expand the obtained word segments to obtain expanded words.

[0145] This step will determine the synonyms and near-synonyms of each word segment, and use the obtained synonyms and near-synonyms as expansion words.

[0146] The scheme for determining synonyms and near-synonyms will still use the existing implementation scheme, which will not be elaborated here.

[0147] This step expands the vocabulary in the input industrial control text to ensure the accuracy of subsequent recognition.

[0148] S103-3, the obtained word segments and expanded words are combined into a word segment set.

[0149] S103-4, for each element u in the word segmentation set,

[0150] 1) Calculate the third similarity Sim between element u and each word in the industrial dictionary. u,a .

[0151] Here, 'a' represents the identifier of a word in the industrial dictionary. The industrial dictionary is a pre-built and continuously updated dictionary that records all industrial terms. By using the industrial dictionary, it is ensured that the final keywords are all related to industry and do not contain irrelevant words, thus guaranteeing the accuracy of the final identification.

[0152] The third similarity calculation scheme adopts the existing word-to-word similarity calculation scheme, which will not be elaborated here.

[0153] 2) If there is no third similarity greater than the preset similarity threshold, then remove element u from the word segmentation set.

[0154] The similarity threshold is a pre-set value, which can be an empirical value or a value calculated based on other methods, such as big data analysis. This embodiment does not limit the method for determining the similarity threshold.

[0155] If there is no third similarity greater than the preset similarity threshold, it means that the element is not similar to any word in the industrial dictionary, and the element is not an industrial term. Therefore, it is removed from the word segmentation set to ensure that all elements in the final word segmentation set are industrial terms, thereby ensuring the accuracy of subsequent recognition.

[0156] 3) If there is a third similarity greater than the preset similarity threshold, then the word corresponding to the third similarity greater than the preset similarity threshold in the industrial dictionary is used as the attribute of element u.

[0157] If there is a third similarity greater than the preset similarity threshold, it means that the element is similar to the word in the industrial dictionary with the third similarity greater than the preset similarity threshold. This word can represent the element. Therefore, the word corresponding to the third similarity greater than the preset similarity threshold in the industrial dictionary is used as the attribute of element u, thereby ensuring the accuracy of subsequent recognition.

[0158] S103-5, determine the keywords based on the attributes of each element in the current word segmentation set.

[0159] The implementation process of S103-5 is as follows:

[0160] 1) Summarize the attributes of each element in the current word segmentation set, and remove duplicate attributes to obtain the attribute set.

[0161] 2) Determine the second value of each element in the attribute set.

[0162] Where b is the element identifier in the attribute set. n is the number of elements in the current word segmentation set whose attribute includes element b. (2) n represents the total number of elements in the current word segmentation set. (3)w represents the total number of the first and second industrial control texts. b Let b be the weight of element b.

[0163] Among them, w b The method for determining it is as follows:

[0164] S401, retrieve the elements in the current word segmentation set whose attributes include element b.

[0165] S402, determine the industrial control text in which the acquired element is located, wherein the industrial control text is the first industrial control text and / or the second industrial control text.

[0166] S403, for each industrial control text.

[0167] (1) Determine the number of word segments included that belong to the word segment set. Where 'c' represents the industrial control text identifier.

[0168] (2) Determine the total number of word segments included.

[0169] S404, Confirm

[0170] Where, n (5) For each industrial control text included, n (6) This represents the total number of industrial control texts in the relevant sector.

[0171] 3) Select the attribute whose second value is greater than the preset threshold.

[0172] 4) In the current word segmentation set, all elements corresponding to the selection attribute and words corresponding to the selection attribute are identified as keywords.

[0173] At this point, the keywords include not only the most relevant elements but also the most relevant attributes, ensuring that the keywords accurately represent the user's industrial semantics and thus guaranteeing the accuracy of the final recognition results.

[0174] S104, forming a multi-semantic vector based on keywords.

[0175] This step employs existing solutions, such as inputting all keywords into a transformer encoder to obtain corresponding multidimensional vectors. This embodiment does not limit the implementation scheme for forming multi-semantic vectors.

[0176] S105, based on multi-semantic vector matching to express the first industrial control text.

[0177] This step can be achieved using existing semantic recognition methods, such as performing semantic recognition on multiple semantic vectors. The semantic recognition results are then used to match the expression of the first industrial control text.

[0178] This embodiment does not limit the method of semantic recognition, nor the specific details of obtaining the final recognition result (i.e., the expression of matching the first industrial control text), and all are implemented using existing semantic recognition methods.

[0179] This embodiment provides a dynamic matching method for multi-semantic text expressions. The method involves acquiring a first industrial control text currently input by the user; acquiring a second industrial control text related to the user; determining keywords based on the first and second industrial control texts; forming a multi-semantic vector based on the keywords; and matching the expression of the first industrial control text based on the multi-semantic vector. The method provided in this embodiment forms a multi-semantic vector based on the user's currently input industrial control text and related industrial control texts, and then matches the expression of the first industrial control text based on the multi-semantic vector. This achieves accurate matching of industrial control text expressions, enabling the industrial control system to accurately identify control commands and achieve precise control.

[0180] To better understand the above technical solutions, exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that the present invention can be understood more clearly and thoroughly, and that the scope of the present invention can be fully conveyed to those skilled in the art.

[0181] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0182] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions.

[0183] It should be noted that any reference numerals placed between parentheses in the claims should not be construed as limiting the claims. The word "comprising" does not exclude the presence of components or steps not listed in the claims. The word "a" or "an" preceding a component does not exclude the presence of a plurality of such components. The invention can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In claims that enumerate several means, several of these means may be embodied by the same hardware. The use of the terms first, second, third, etc., is merely for convenience of expression and does not indicate any order. These terms can be understood as part of the component names.

[0184] Furthermore, it should be noted that in the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0185] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the claims should be interpreted to include both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0186] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, then this invention should also include these modifications and variations.

Claims

1. A method for dynamic matching of multi-semantic text expressions, characterized in that, The method comprises: S101, obtaining first industrial control text input by a user currently; S102, obtaining second industrial control text related to the user; S103, determining a keyword according to the first industrial control text and the second industrial control text; S104, forming a multi-semantics vector based on the keyword; S105, matching expression of the first industrial control text based on the multi-semantics vector; The S102 comprises: S201, determining whether the user inputs industrial control text within a preset time period; S202, if industrial control text is input within the preset time period, determining all the input industrial control text as target text; If no industrial control text is input within the preset time period, determining whether the user has input industrial control text; If the user has input industrial control text, selecting target text from near to far according to input time; If the user has not input industrial control text, selecting target text according to attributes of the user; S203, obtaining the second industrial control text according to the target text; The selecting target text from near to far according to input time comprises: forming a historical text set by taking the industrial control text input by the user as elements from near to far according to input time; determining a difference between input time of adjacent two elements in the historical text set to form a difference set D, wherein the difference is time of a previous element-time of a next element; calculating the standard deviation of all the differences ; selecting a first element and a second element in the difference set D; computing a first value ; wherein j is an element identification in the set of selection differences D, is an average of element values in the set of selection differences D, n d is a total number of elements in the set of selection differences D; if the first value is not less than then selecting the element involved in the historical text set by the currently selected element as the target text; If the first value is less than , the first element, the second element and the third element of the difference set D are selected, the step of calculating the first value and the subsequent steps are re-executed, and the cycle is repeated, and one subsequent element of the difference set D is selected each time, until the currently calculated first value is not less than .

2. The method of claim 1, wherein, The selecting target text according to attributes of the user comprises: forming a user attribute set U0 by taking the attributes of the user as elements; forming respective attribute sets U of attributes of other users i where i is an identification of the other user. computing a first similarity between the user and each other user where w i is a weight for other user i; According to Sim 0,i , select the target text.

3. The method of claim 2, wherein, For any other user i, the w i is determined by the steps of: S301, obtaining all third industrial control text input by other users i after each login recently; S302, for all the third industrial control text input by other users i after each login x, ordering all the third industrial control text input by other users i after each login x from near to far according to time sequence to form an x set; determining core keywords of each element in the x set; Splicing the core keywords of each element to obtain vector h x ; S303, sequentially taking the vector h after each login x , calculating the second similarity between it and the vector h after each login of other times y ; S304, determine the smallest second similarity as w i .

4. The method of claim 1, wherein, The S203 comprises: determining all the target text as second industrial control text; or determining a preset number of target text recently as second industrial control text.

5. The method of claim 1, wherein, The S103 comprises: S103-1, performing word segmentation on the first industrial control text and the second industrial control text; S103-2, performing semantic extension on the segmented words to obtain extended words; S103-3, forming a segmented word set by taking the segmented words and the extended words as elements; S103-4, for each element u in the segmented word set, a third similarity Sim between the computational element u and each word in the industrial dictionary u,a wherein a is an identification of a word in the industrial dictionary; if there is no third similarity greater than a preset similarity threshold, removing the element u from the segmented word set; if there is a third similarity greater than a preset similarity threshold, taking a word corresponding to the third similarity greater than the preset similarity threshold in an industrial dictionary as an attribute of the element u; S103-5, determining a keyword according to the attributes of each element in the current segmented word set.

6. The method of claim 5, wherein, The S103-5 comprises: summarizing the attributes of each element in the current segmented word set and removing duplicate attributes to obtain an attribute set; determining second values of elements in the attribute set wherein b is an element identifier in the attribute set, n is the number of elements in the current segmentation set whose attributes include element b (2) n is the total number of elements in the current segmentation set (3) w is the total number of first and second industrial control texts b is the weight of element b; selecting an attribute with a second value greater than a preset threshold; The element corresponding to the selection attribute and the word corresponding to the selection attribute in the current segmentation set are both determined as keywords.

7. The method of claim 6, wherein, The w b The determination method is: S401, obtaining an element b included in an attribute in a current segmentation set; S402, determining an industrial control text where the obtained element is located, wherein the industrial control text where the element is located is a first industrial control text and / or a second industrial control text; S403, for each industrial control text where the element is located, determining the number of word segmentation included in the word segmentation set wherein c is the industrial control text identifier; determining the total number of segmented words included therein ; S404, determining wherein n (6) is the total number of industrial control texts.

8. The method of claim 1, wherein, The S105 includes: performing semantic recognition on the multi-semantics vector; matching the expression of the first industrial control text according to the semantic recognition result.

Images