Production process classification method and system, storage medium and terminal
By establishing a spatial vector model, converting the process text data into spatial vector coordinates, calculating the similarity and classifying the process with thresholds, solving the problem of low accuracy in process classification in the existing technology and achieving efficient process classification.
Patent Information
- Application Number
- CN202311310977.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-10
- Publication Date
- 2025-04-15
AI Technical Summary
The prior art has the problem of low accuracy when classifying production processes, especially in complex and changeable processes, which are difficult to achieve accurate classification.
By establishing a spatial vector model, the process text data is converted into n-dimensional spatial vector coordinates, and the similarity between different production processes is calculated based on these coordinates, and the process is classified in combination with the similarity threshold.
It realizes accurate classification of production processes and improves the accuracy and efficiency of process classification.
Smart Images
Figure CN119830021A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of semiconductor manufacturing technology, and in particular to a production process classification method and system, a storage medium and a terminal. Background Art
[0002] With the rapid development of artificial intelligence application technology, chips are becoming increasingly important. Chips are also called accelerators or calculators, which are modules specifically used to handle large amounts of computing tasks in artificial intelligence applications.
[0003] The birth of a chip goes through design, manufacturing and testing, which need to correspond to the integrated circuit industry chain design, manufacturing and packaging and testing industries respectively. Faced with the medical industry, unmanned driving in the security industry, smart phones, etc., each item includes complex steps and processes. For large-scale wafer foundries, how to accurately classify complex and changeable processes has become an important solution.
[0004] However, there are still many problems in the process of classifying processes in the prior art. Summary of the invention
[0005] The technical problem solved by the present invention is to provide a production process classification method and system, a storage medium and a terminal to achieve accurate classification of the production process.
[0006] To solve the above problems, the present invention provides a production process classification method, comprising: obtaining process text data of a plurality of production processes, each of the production processes comprising a plurality of production steps, and the process text data comprising step text data of each of the production steps; establishing a space vector model, the space vector model comprising n-dimensional feature parameters, and the feature parameters of each dimension are represented by digitalization; inputting the plurality of process text data into the space vector model for calculation and processing, and obtaining the n-dimensional space vector coordinates corresponding to each of the production processes; obtaining the similarity between different production processes according to the space vector coordinates of each of the production processes; providing a similarity threshold; and classifying the plurality of production processes according to the similarity between different production processes and the similarity threshold.
[0007] Optionally, before inputting the plurality of process text data into the space vector model for calculation processing, the method further includes: removing redundant data in each piece of the process text data.
[0008] Optionally, the redundant data includes: special symbol text data in each piece of the process text data, and repeated text data common to a number of the process text data.
[0009] Optionally, a number of the process text data are input into the space vector model for calculation and processing, and the method for obtaining the space vector coordinates corresponding to each of the production processes includes: obtaining the n-dimensional initial space vector coordinates in the space vector model mapped with the step text data of each production step in each of the production processes; and performing weighted addition on the n-dimensional initial space vector coordinates to obtain the n-dimensional space vector coordinates corresponding to each of the production processes.
[0010] Optionally, the method for obtaining the similarity between different production processes according to the spatial vector coordinates of each production process includes: calculating the spatial vector coordinates of different production processes by cosine similarity to obtain the similarity between different production processes.
[0011] Optionally, the method for classifying several production processes according to the similarity between different production processes and the similarity threshold includes: when the absolute value of the difference between the similarity between different production processes and the similarity threshold 1 is within a preset range, the two are classified into the same category; when the absolute value of the difference between the similarity between different production processes and the similarity threshold -1 is within a preset range, the two are classified into opposite categories; when the absolute value of the difference between the similarity between different production processes and the similarity threshold 0 is within a preset range, the two are classified into independent categories.
[0012] Correspondingly, the technical solution of the present invention also provides a production process classification system, including: a data acquisition module, used to obtain process text data of several production processes, each of the production processes includes several production steps, and the process text data includes step text data of each production step; a modeling module, used to establish a space vector model, the space vector model includes n-dimensional feature parameters, and the feature parameters of each dimension are represented by digitalization; a processing module, used to perform calculation processing after inputting several of the process text data into the space vector model, and obtain the n-dimensional space vector coordinates corresponding to each of the production processes; a similarity acquisition module, used to obtain the similarity between different production processes according to the space vector coordinates of each of the production processes; a classification module, used to set a similarity threshold, and classify several of the production processes according to the similarity between different production processes and the similarity threshold.
[0013] Optionally, the data acquisition module further includes: a redundant data removal module, configured to remove redundant data in each piece of the process text data.
[0014] Optionally, the redundant data includes: special symbol text data in each piece of the process text data, and repeated text data common to a number of the process text data.
[0015] Optionally, the processing module includes: a mapping module, used to obtain the n-dimensional initial space vector coordinates in the space vector model by mapping the step text data of each production step in each production process; a weighted superposition module, used to perform weighted addition of several n-dimensional initial space vector coordinates to obtain the n-dimensional space vector coordinates corresponding to each production process.
[0016] Optionally, the similarity acquisition module calculates the spatial vector coordinates of different production processes by cosine similarity to acquire the similarity between different production processes.
[0017] Optionally, the classification module includes: an input module for inputting the similarity threshold; and a comparison module for classifying a plurality of the production processes according to the similarities between different production processes and the similarity threshold.
[0018] Optionally, the comparison module compares the similarity between different production processes with a similarity threshold of 1, and when the absolute value of the difference is within a preset range, the two are classified into the same category; the comparison module compares the similarity between different production processes with a similarity threshold of -1, and when the absolute value of the difference is within a preset range, the two are classified into opposite categories; the comparison module compares the similarity between different production processes with a similarity threshold of 0, and when the absolute value of the difference is within a preset range, the two are classified into independent categories.
[0019] Correspondingly, the technical solution of the present invention also provides a storage medium on which computer instructions are stored, characterized in that when the computer instructions are executed, the steps of the method described in any of the above technical solutions are executed.
[0020] Correspondingly, the technical solution of the present invention also provides a terminal, including a memory and a processor, wherein the memory stores computer instructions that can be run on the processor, and is characterized in that when the processor runs the computer instructions, the steps of the method described in any of the above technical solutions are executed.
[0021] Compared with the prior art, the technical solution of the present invention has the following advantages:
[0022] In the production process classification method of the technical solution of the present invention, a space vector model is established, the space vector model includes n-dimensional characteristic parameters, and the characteristic parameters of each dimension are represented by digitalization; and a number of the process text data are input into the space vector model for calculation and processing to obtain the n-dimensional space vector coordinates corresponding to each of the production processes; according to the space vector coordinates of each of the production processes, the similarity between different production processes is obtained; a similarity threshold is provided; according to the similarity between different production processes and the similarity threshold, a number of the production processes are classified. By converting the process text data that cannot be directly mathematized into the space vector coordinates that can be directly mathematized by the space vector model, the number of the production processes can be accurately classified by calculation according to the digitalized space vector coordinates.
[0023] The production process classification system of the technical solution of the present invention includes: a modeling module for establishing a space vector model, the space vector model includes n-dimensional feature parameters, and the feature parameters of each dimension are digitally represented; a processing module for performing calculation processing after inputting a number of process text data into the space vector model to obtain the n-dimensional space vector coordinates corresponding to each production process; a similarity acquisition module for obtaining the similarity between different production processes according to the space vector coordinates of each production process; a classification module for setting a similarity threshold, and classifying a number of production processes according to the similarity between different production processes and the similarity threshold. By converting the process text data that cannot be directly mathematically calculable into the space vector coordinates that can be directly mathematically calculated by the space vector model, the production processes can be accurately classified by calculating according to the digital space vector coordinates. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 It is a flow chart of a production process classification method according to an embodiment of the present invention;
[0025] Figures 2 to 4 It is a schematic diagram of each step of the production process classification method in an embodiment of the present invention;
[0026] Figure 5 It is a structural schematic diagram of a production process classification system in an embodiment of the present invention. DETAILED DESCRIPTION
[0027] As described in the background art, there are still many problems in the process of classifying processes in the prior art, which will be described in detail below.
[0028] At present, electronic technology is becoming more and more advanced, and the requirements for chips are getting higher and higher. New products are constantly being iterated. Due to the large number of products and complex steps, both the early calculation process and the later processing process require most of the time for classification. However, the current production process classification is mainly based on manual experience, which makes the classification accuracy low.
[0029] On this basis, the present invention provides a production process classification method and system, storage medium and terminal, establishes a space vector model, the space vector model includes n-dimensional feature parameters, and the feature parameters of each dimension are digitally represented; and inputs a number of the process text data into the space vector model for calculation and processing, and obtains the n-dimensional space vector coordinates corresponding to each of the production processes; according to the space vector coordinates of each of the production processes, obtains the similarity between different production processes; provides a similarity threshold; and classifies a number of the production processes according to the similarity between different production processes and the similarity threshold. By converting the process text data that cannot be directly mathematically calculated into the space vector coordinates that can be directly mathematically calculated by the space vector model, the number of the production processes can be accurately classified by calculating according to the digital space vector coordinates.
[0030] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.
[0031] Figure 1 1 is a flow chart of a production process classification method according to an embodiment of the present invention, comprising:
[0032] Step S101, obtaining process text data of a plurality of production processes, each of the production processes comprising a plurality of production steps, and the process text data comprising step text data of each of the production steps;
[0033] Step S102, establishing a space vector model, wherein the space vector model includes characteristic parameters of n dimensions, and the characteristic parameters of each dimension are represented by digitalization;
[0034] Step S103, inputting a number of the process text data into the space vector model for calculation processing, and obtaining the n-dimensional space vector coordinates corresponding to each of the production processes;
[0035] Step S104, obtaining similarities between different production processes according to the spatial vector coordinates of each production process;
[0036] Step S105, providing a similarity threshold;
[0037] Step S106: classify the production processes according to the similarities between the different production processes and the similarity threshold.
[0038] The following is a detailed description of each step of the production process classification method in conjunction with the accompanying drawings.
[0039] Figures 2 to 4 It is a schematic diagram of each step of the production process classification method in an embodiment of the present invention.
[0040] Please refer to Figure 2 , obtaining process text data of a plurality of production processes, each of the production processes comprising a plurality of production steps, and the process text data comprising step text data of each of the production steps.
[0041] In this embodiment, the four production processes are taken as examples.
[0042] In this embodiment, the process text data mainly includes numbers, English letters and Chinese characters.
[0043] In this embodiment, after obtaining the process text data of a plurality of production processes, the method further includes: removing redundant data in each piece of the process text data.
[0044] In this embodiment, the redundant data includes: special symbol text data in each piece of the process text data, and repeated text data common to a number of the process text data.
[0045] In this embodiment, the data independence in each piece of the process text data is ensured by removing the special symbol text data and the duplicate text data common to a plurality of the process text data.
[0046] Please refer to Figure 3 , a space vector model 100 is established, wherein the space vector model 100 includes characteristic parameters of n dimensions, and the characteristic parameters of each dimension are represented by digitalization.
[0047] In this embodiment, the production process is an ion implantation process, and the space vector model 100 includes three-dimensional characteristic parameters as an example, wherein the three-dimensional characteristic parameters are: ion element, dose of implanted ions, and angle of implanted ions.
[0048] It should be noted that, in this embodiment, when the characteristic parameter is an ionic element, it is digitally represented by the nuclear charge number of the element, such as when the ionic element is hydrogen, it is represented by the number 1, and when the ionic element is oxygen, it is represented by the number 8.
[0049] In other embodiments, the dimension of the characteristic parameters in the space vector model may be greater than 3.
[0050] Please refer to Figure 4 , inputting some of the process text data into the space vector model 100 for calculation and processing, and obtaining the n-dimensional space vector coordinates corresponding to each of the production processes.
[0051] In this embodiment, a number of the process text data are input into the space vector model 100 for calculation and processing, and the method for obtaining the space vector coordinates corresponding to each of the production processes includes: obtaining the n-dimensional initial space vector coordinates in the space vector model 100 of the step text data of each of the production steps in each of the production processes; and performing weighted addition on the n-dimensional initial space vector coordinates to obtain the n-dimensional space vector coordinates corresponding to each of the production processes.
[0052] In this embodiment, the weights of the initial space vectors are all assigned the same weight.
[0053] In other embodiments, if some of the production steps are very important, a larger weight will be given to the corresponding initial space vector.
[0054] Please continue to refer to Figure 4 , according to the spatial vector coordinates of each production process, the similarity between different production processes is obtained.
[0055] In this embodiment, according to the spatial vector coordinates of each production process, the method for obtaining the similarity between different production processes includes: calculating the spatial vector coordinates of different production processes by cosine similarity to obtain the similarity between different production processes.
[0056] Please continue to refer to Figure 4 , providing a similarity threshold; classifying several of the production processes according to the similarities between different production processes and the similarity threshold.
[0057] In this embodiment, the process text data that cannot be directly quantified is converted into the space vector coordinates that can be directly quantified by the space vector model 100, and the several production processes can be accurately classified by calculation based on the digitized space vector coordinates.
[0058] In this embodiment, according to the similarity between different production processes and the similarity threshold, the method for classifying several production processes includes: when the absolute value of the difference between the similarity between different production processes and the similarity threshold 1 is within a preset range, the two are classified into the same category; when the absolute value of the difference between the similarity between different production processes and the similarity threshold -1 is within the preset range, the two are classified into opposite categories; when the absolute value of the difference between the similarity between different production processes and the similarity threshold 0 is within the preset range, the two are classified into independent categories.
[0059] As shown in Table 1, Table 1 shows the three-dimensional space vector coordinates corresponding to the four different production processes in the space vector model 100, which are as follows:
[0060] Coordinate Axes Production process 1 Production process 2 Production process 3 Production process 4 X 132.5 278.6 92.9 -45 Y ﹣68 144.3 41 101.5 Z 242 56 185.4 77.8
[0061] Table 1
[0062] As shown in Table 2, the similarities between the different production processes are calculated based on the cosine similarity, as follows:
[0063] Production process 1 Production process 2 Production process 3 Production process 4 Production process 1 1.0 1 / 0.4 1 / 0.9 1 / 0.2 Production process 2 0.4 1.0 1 / 0.63 1 / 0.15 Production process 3 0.9 0.63 1.0 1 / 0.5 Production process 4 0.2 0.15 0.5 1.0
[0064] Table 2
[0065] It should be noted that the "1 / " in Table 2 represents the reverse comparison between the different production processes, and does not mean "1 divided by", such as the final similarities of "0.4" and "1 / 0.4" are both 0.4.
[0066] As shown in Table 2, the similarity between the production process 2 and the production process 4 is 0.15, and the difference between the two is close to 0, so the two can be classified as independent categories; the similarity between the production process 1 and the production process 3 is 0.91, and the similarity between the difference between the two is close to 1, so the two can be classified as the same category.
[0067] Figure 5 It is a structural schematic diagram of a production process classification system in an embodiment of the present invention.
[0068] Correspondingly, the present invention also provides a production process classification system, please refer to Figure 5 And continue to combine reference Figures 2 to 4, including: a data acquisition module 200, used to acquire process text data of a plurality of production processes, each of the production processes includes a plurality of production steps, and the process text data includes step text data of each of the production steps; a modeling module 201, used to establish a space vector model 100, the space vector model 100 includes n-dimensional feature parameters, and the feature parameters of each dimension are represented by digitalization; a processing module 202, used to perform calculation processing after inputting the plurality of process text data into the space vector model 100, and acquire the n-dimensional space vector coordinates corresponding to each of the production processes; a similarity acquisition module 203, used to acquire the similarity between different production processes according to the space vector coordinates of each of the production processes; a classification module 204, used to set a similarity threshold, and classify the plurality of production processes according to the similarity between different production processes and the similarity threshold.
[0069] In this embodiment, the process text data that cannot be directly quantified is converted into the space vector coordinates that can be directly quantified by the space vector model 100, and the several production processes can be accurately classified by calculation based on the digitized space vector coordinates.
[0070] Please continue to refer to Figure 5 In this embodiment, the data acquisition module 200 further includes: a redundant data removal module 2001, which is used to remove redundant data in each of the process text data.
[0071] In this embodiment, the redundant data includes: special symbol text data in each piece of the process text data, and repeated text data common to a number of the process text data.
[0072] Please continue to refer to Figure 5 In this embodiment, the processing module 202 includes: a mapping module 2021, which is used to obtain the n-dimensional initial space vector coordinates in the space vector model 100 mapped with the step text data of each production step in each production process; a weighted superposition module 2022, which is used to perform weighted addition on several n-dimensional initial space vector coordinates to obtain the n-dimensional space vector coordinates corresponding to each production process.
[0073] In this embodiment, the similarity acquisition module calculates the space vector coordinates of different production processes by cosine similarity to acquire the similarity between different production processes.
[0074] Please continue to refer to Figure 5In this embodiment, the classification module 204 includes: an input module 2041, used to input the similarity threshold; a comparison module 2042, used to classify several production processes according to the similarities between different production processes and the similarity threshold.
[0075] In this embodiment, the comparison module compares the similarity between different production processes with the similarity threshold 1, and when the absolute value of the difference is within the preset range, the two are classified into the same category; the comparison module compares the similarity between different production processes with the similarity threshold -1, and when the absolute value of the difference is within the preset range, the two are classified into opposite categories; the comparison module compares the similarity between different production processes with the similarity threshold 0, and when the absolute value of the difference is within the preset range, the two are classified into independent categories.
[0076] Correspondingly, an embodiment of the present invention further provides a storage medium on which computer instructions are stored, characterized in that when the computer instructions are executed, the steps of the method described in any of the above embodiments are executed.
[0077] Correspondingly, an embodiment of the present invention further provides a terminal, including a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, and when the processor executes the computer instructions, the steps of the method described in any one of the above embodiments are executed.
[0078] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.
Claims
1. A production process classification method, characterized in that: include: Acquire process text data of a plurality of production processes, each of the production processes comprising a plurality of production steps, and the process text data comprising step text data of each of the production steps; Establishing a space vector model, wherein the space vector model includes characteristic parameters of n dimensions, and the characteristic parameters of each dimension are represented by digitalization; Inputting a number of the process text data into the space vector model for calculation and processing, and obtaining the n-dimensional space vector coordinates corresponding to each of the production processes; According to the spatial vector coordinates of each production process, obtaining the similarity between different production processes; Provide a similarity threshold; According to the similarities between different production processes and the similarity threshold, several production processes are classified.
2. The production process classification method according to claim 1, characterized in that: Before inputting the plurality of process text data into the space vector model calculation process, the method further includes: removing redundant data in each piece of the process text data.
3. The production process classification method according to claim 2, characterized in that: The redundant data includes: special symbol text data in each of the process text data, and repeated text data common to a number of the process text data.
4. The production process classification method according to claim 1, characterized in that: A method for inputting a number of the process text data into the space vector model for calculation and processing to obtain the space vector coordinates corresponding to each of the production processes includes: obtaining the n-dimensional initial space vector coordinates mapped in the space vector model by the step text data of each production step in each of the production processes; and performing weighted addition on the n-dimensional initial space vector coordinates to obtain the n-dimensional space vector coordinates corresponding to each of the production processes.
5. The production process classification method according to claim 1, characterized in that: According to the spatial vector coordinates of each of the production processes, a method for obtaining the similarity between different production processes includes: calculating the spatial vector coordinates of different production processes by cosine similarity to obtain the similarity between different production processes.
6. The production process classification method according to claim 5, characterized in that: The method for classifying several production processes according to the similarity between different production processes and the similarity threshold includes: when the absolute value of the difference between the similarity between different production processes and the similarity threshold 1 is within a preset range, the two are classified into the same category; when the absolute value of the difference between the similarity between different production processes and the similarity threshold -1 is within a preset range, the two are classified into opposite categories; when the absolute value of the difference between the similarity between different production processes and the similarity threshold 0 is within a preset range, the two are classified into independent categories.
7. A production process classification system, characterized in that: include: A data acquisition module, used to acquire process text data of a plurality of production processes, each of the production processes comprising a plurality of production steps, and the process text data comprising step text data of each of the production steps; A modeling module, used for establishing a space vector model, wherein the space vector model includes characteristic parameters of n dimensions, and the characteristic parameters of each dimension are represented by digitalization; A processing module, used for performing calculation processing after inputting a number of the process text data into the space vector model, and obtaining the n-dimensional space vector coordinates corresponding to each of the production processes; A similarity acquisition module, used for acquiring the similarity between different production processes according to the spatial vector coordinates of each production process; The classification module is used to set a similarity threshold and classify a plurality of the production processes according to the similarities between different production processes and the similarity threshold.
8. The production process classification system according to claim 7, characterized in that: The data acquisition module further includes: a redundant data removal module, which is used to remove redundant data in each piece of process text data.
9. The production process classification system according to claim 8, characterized in that: The redundant data includes: special symbol text data in each of the process text data, and repeated text data common to a number of the process text data.
10. The production process classification system according to claim 7, characterized in that: The processing module includes: a mapping module, which is used to obtain the n-dimensional initial space vector coordinates in the space vector model by mapping the step text data of each production step in each production process; and a weighted superposition module, which is used to perform weighted addition on several n-dimensional initial space vector coordinates to obtain the n-dimensional space vector coordinates corresponding to each production process.
11. The production process classification system according to claim 7, characterized in that: The similarity acquisition module calculates the spatial vector coordinates of different production processes through cosine similarity to acquire the similarity between different production processes.
12. The production process classification system according to claim 7, characterized in that: The classification module includes: an input module for inputting the similarity threshold; and a comparison module for classifying a plurality of the production processes according to the similarities between different production processes and the similarity threshold.
13. The production process classification system according to claim 12, characterized in that: The comparison module compares the similarity between different production processes with the similarity threshold 1, and when the absolute value of the difference is within the preset range, the two are classified into the same category; the comparison module compares the similarity between different production processes with the similarity threshold -1, and when the absolute value of the difference is within the preset range, the two are classified into opposite categories; the comparison module compares the similarity between different production processes with the similarity threshold 0, and when the absolute value of the difference is within the preset range, the two are classified into independent categories.
14. A storage medium having computer instructions stored thereon, characterized in that: When the computer instructions are executed, the steps of the method according to any one of claims 1 to 6 are executed.
15. A terminal comprising a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, characterized in that: When the processor runs the computer instructions, the steps of the method according to any one of claims 1 to 6 are performed.