Medical material inventory information management method and system

Abstract

The invention relates to the field of data processing, in particular to a medical material inventory information management method and system, and aims to extract various existing habit characteristics from complete historical material getting data. Meanwhile, the current material picking record is verified, if incomplete records such as information loss and unconfirmed handwriting exist, the material picking habit features are matched with known information of the material picking record, and if the material picking habit features can be matched with the known information, the material picking habit features are matched with the known information; and if yes, complementing the incomplete record by using the target material getting habit feature and the confidence coefficient of the target material getting habit feature. According to the method, the historical data is analyzed to obtain the material getting habit characteristics, and the missing registration data is automatically repaired by using the material getting habit characteristics, so that the confirmation workload of management personnel is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and system for managing inventory information of medical supplies. Background Art

[0002] Inventory management of medical supplies is a complex but crucial process that is directly related to the quality and efficiency of medical services. Effective inventory management can ensure the safe and effective supply of medical supplies and reduce waste. Most of the existing inventory management is based on information technology, and the information of medical supplies entering, leaving, and counting will be entered into the management system. This facilitates management such as usage registration and inventory warning.

[0003] The existing technology generally relies on manual registration for warehouse material delivery information registration. With the transformation of warehouse informatization, this information will be transcribed into the computer management system. However, for manually filled records, due to the lack of mandatory management measures, it is easy to be incomplete and the handwriting is difficult to recognize. As a result, when the information on the register is entered into the information management system, it is easy to have incomplete information. Summary of the invention

[0004] In view of this, an object of the present invention is to provide a method and system for managing inventory information of medical supplies to solve the problems in the background technology.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions: A method for managing inventory information of medical supplies of the present invention comprises the following steps: Obtain the material collection records and historical material collection data of the current management cycle, wherein the material collection records are paper materials and the historical material collection data are electronic data; Extracting a plurality of material picking habit features and confidence levels of the plurality of material picking habit features from the historical material picking data; Performing image and text recognition on the material collection record to obtain preliminary material collection data; and verifying the preliminary material collection data to obtain an incomplete record; The incomplete record is matched with the multiple material picking habit features based on the known data of the incomplete record, and when the incomplete record matches any target material picking habit feature, the incomplete record is completed based on the target material picking habit feature and the confidence of the target material picking habit feature.

[0006] In an embodiment of the present application, the historical material picking data includes multiple material picking record data, each material picking record data includes values ​​of multiple parameters, wherein multiple material picking habit features and confidence levels of multiple material picking habits are extracted from the historical material picking data, including: Remove useless parameters from historical material collection data to obtain target parameters; Combine any two parameters to get the sample data template ,in, represents the first target parameter, represents the second target parameter; Based on the sample data template Extract multiple sample data from the historical material picking data , Indicates The value of the first target parameter, Indicates the value of the second target parameter; For multiple sample data Perform clustering to obtain multiple clusters ,in, is the cluster number; Based on the multiple clusters Verify the characteristics of picking habits and the confidence of picking habits.

[0007] In one embodiment of the present application, based on the multiple clusters Verify the characteristics of picking habits and the confidence of picking habits, including: For the plurality of clusters Marking is performed to obtain the multiple clusters The class label of the first target parameter in and the class label of the second target parameter ; Statistical Clusters The number of samples within a cluster , in line with the category label The number of sample data And the category label The number of sample data ; Calculate separately The number of samples within a cluster Matches the category label The number of sample data Ratio , the number of samples within the cluster Matches the category label The number of sample data Ratio ; In the ratio When it is greater than the preset ratio threshold, the category label is determined exist The characteristics of the material picking habits; in the ratio When it is greater than the preset ratio threshold, the category label is determined exist The picking habit characteristics of the materials; and the ratio is used as the confidence of the picking habit characteristics.

[0008] In one embodiment of the present application, based on the sample data template Extract multiple sample data from the historical material picking data ,include: When the sample data template When any parameter in the sample data template is not a medical material name, Construct sample data with the values ​​of parameters belonging to the same picking record ; When the sample data template When the stored parameter is the name of a medical material, the name of the medical material is vectorized and based on the sample data template Construct sample data with the values ​​of parameters belonging to the same picking record .

[0009] In one embodiment of the present application, multiple sample data Perform clustering to obtain multiple clusters ,include: When the first target parameter and the second target parameter are both numeric or vector types, construct a two-dimensional parameter , and for the two-dimensional parameters Perform density clustering to obtain multiple clusters ; When one of the first target parameter and the second target parameter is of character type and the other is of numeric type or vector type; for multiple sample data based on the character type target parameter Divide into multiple data units; in each data unit, multiple sample data are sorted based on the numerical type or vector type. Perform clustering to obtain multiple clusters .

[0010] In one embodiment of the present application, the plurality of clusters Marking is performed to obtain the multiple clusters The class label of the first target parameter in and the class label of the second target parameter ,include: When the first target parameter or the second target parameter is a numeric type, calculate the average value of the target parameter whose data type is a numeric value within the cluster and standard deviation , and based on the average and the standard deviation Constructing reference range labels ,in, is the range adjustment parameter; When the first target parameter or the second target parameter is of a vector type, a type label of the target parameter whose numerical type is a vector; When the first target parameter or the second target parameter is of character type, the character itself is used as a label of the target parameter whose data type is character.

[0011] In one embodiment of the present application, matching the incomplete record with the multiple picking habit features based on the known data of the incomplete record includes: The known data of the incomplete record is matched with the labels of the multiple material picking habit features, and the material picking habit features whose labels match the known data are used as target material picking habit features.

[0012] In one embodiment of the present application, the incomplete record is completed based on the target material picking habit feature and the confidence of the target material picking habit feature, including: Determine candidate labels for incomplete records based on the target picking habit characteristics and known data And candidate labels Confidence , ,in, For the The confidence level corresponding to the target picking habit characteristics of known data; Filter out target labels whose confidence is greater than the confidence threshold; The incomplete record is completed based on the target tag.

[0013] In one embodiment of the present application, it also includes: The multiple candidate labels with the highest confidence Sent to the corresponding object for confirmation.

[0014] The present application provides an inventory information management system for medical supplies, including: An acquisition module is used to acquire the material collection records and historical material collection data of the current management cycle, wherein the material collection records are paper materials and the historical material collection data are electronic data; A feature extraction module, used to extract multiple material picking habit features and confidence levels of the multiple material picking habit features from the historical material picking data; A verification module is used to perform image and text recognition on the material collection record to obtain preliminary material collection data; and verify the preliminary material collection data to obtain an incomplete record; A data completion module is used to match the incomplete record with the multiple material picking habit features based on the known data of the incomplete record, and when the incomplete record matches any target material picking habit feature, complete the incomplete record based on the target material picking habit feature and the confidence of the target material picking habit feature.

[0015] The beneficial effect of the present invention is that a method and system for managing inventory information of medical supplies of the present invention extracts multiple existing habit features from complete historical material collection data. At the same time, the current material collection records are verified. If there are incomplete records such as missing information and unconfirmed handwriting, the material collection habit features are used to match the known information of the material collection records. If a match can be made, the target material collection habit features and the confidence of the target material collection habit features are used to complete the incomplete records. The present application utilizes analysis of historical data to obtain material collection habit features, and utilizes the material collection habit features to automatically repair missing registration data, thereby effectively reducing the confirmation workload of management personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 is a flow chart of a method for managing inventory information of medical supplies shown in one embodiment of the present application; Figure 2 This is a flow chart for extracting the material picking habit features in one embodiment of the present application; Figure 3 It is a structural diagram of an inventory information management system for medical supplies shown in one embodiment of the present application. DETAILED DESCRIPTION

[0017] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0018] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show the layers related to the present invention rather than being drawn according to the number, shape and size of the layers in actual implementation. In actual implementation, the type, quantity and proportion of each layer may be changed arbitrarily, and the layer layout may also be more complicated.

[0019] In the following description, numerous details are discussed to provide a more thorough explanation of embodiments of the present invention; however, it is apparent to one skilled in the art that embodiments of the present invention may be practiced without these specific details.

[0020] Figure 1 is a flow chart of a method for managing inventory information of medical supplies shown in one embodiment of the present application, such as Figure 1 As shown, a method for managing inventory information of medical supplies in this embodiment may include the following steps: S110, obtaining the material picking record and historical material picking data of the current management cycle, wherein the material picking record is paper material and the historical material picking data is electronic data; Historical material collection records refer to complete material collection records that have passed verification, while material collection records for the current management cycle refer to all paper registration records that need to extract material collection information through image recognition during the current entry cycle. Material collection records are stored in the form of images.

[0021] S120, extracting a plurality of material picking habit features and confidence levels of the plurality of material picking habit features from the historical material picking data; The following table is a material collection record form in an embodiment of the present application:

[0022] The habit feature of picking materials in this application refers to the range of values ​​of any two picking parameters with a high probability of co-occurrence. For example: There is a high probability of picking up daily medical consumables such as masks, cotton swabs, and iodine tincture from 9:00-10:00 in the morning, that is, there is a habit feature of picking materials: {time: 9:00-10:00; type of picking materials: daily medical consumables}. This application is not limited to the above habit features, but also includes the picker and the time of picking, the picker and the type of picking, the type of picking and the quantity of picking, etc.

[0023] Figure 2 This is a flow chart for extracting the material picking habit features in an embodiment of the present application, such as Figure 2 As shown, specifically, the present application adopts the following process to extract multiple picking habit features and the confidence levels of multiple picking habit features, including: S121, removing useless parameters from historical material picking data to obtain target parameters; For example, remove information such as serial number, department, specification / unit, approver, etc. This information cannot reflect the customary rules, or can be directly derived from existing information (material picker-unit; material name-unit / specification).

[0024] S122, combining any two parameters to obtain a sample data template ,in, represents the first target parameter, represents the second target parameter; The sample data template refers to a template composed of two header parameters, such as material collection time / material name, material collection time / quantity, material name / quantity, etc.

[0025] S123, based on the sample data template Extract multiple sample data from the historical material picking data , Indicates The value of the first target parameter, Indicates the value of the second target parameter; After building the sample data template After that, the sample data template Compare it with the table and then extract the corresponding data samples.

[0026] However, in the subsequent processing, clustering operations are required, so the material names need to be vectorized. The specific process is as follows: (1) When the sample data template When any parameter in the sample data template is not a medical material name, Construct sample data with the values ​​of parameters belonging to the same picking record ; If the sample data template If any type of parameter is not the name of medical supplies, that is, the corresponding data type is character type or numeric type, then the original information is retained, such as the person who receives the materials, quantity, time, etc.

[0027] (2) When the sample data template When the stored parameter is the name of a medical material, the name of the medical material is vectorized and based on the sample data template Construct sample data with the values ​​of parameters belonging to the same picking record .

[0028] If the parameter is a medical material name, vectorization is performed. In this application, a pre-trained Word2vec model is used to vectorize the medical material names, so that there is a high cosine similarity between the word vectors corresponding to the medical material names with similar functions and principles. The training process of the Word2vec model is as follows: (1) Collect medical material name data: Collect the names of medical materials from open knowledge bases, official databases, or professional websites in the medical field. Ensure that the data covers a wide range of medical materials, especially those with similar functional principles.

[0029] (2) Data preprocessing: Perform word segmentation on the collected medical supplies names to remove stop words and noise. The segmented results are converted into a format suitable for Word2vec model training.

[0030] (3) Select the Word2vec model and set the model parameters: Determine the dimension of the vector (such as 100 dimensions, 200 dimensions, etc.), which will affect the vector's expressiveness and computational complexity. Set the window size (i.e., the number of context words considered). The larger the window size, the richer the context information that the model can capture. Determine the number of iterations and training algorithm to ensure that the model can fully learn the relationship between words.

[0031] (4) Use the preprocessed medical material name data to train the Word2vec model. During the training process, the model will continuously adjust parameters to learn the similarities between medical material names.

[0032] After training, the model can map each medical material name into a high-dimensional vector space. By querying the model, you can get the vector representation of each medical material name.

[0033] S124, for multiple sample data Perform clustering to obtain multiple clusters ,in, is the cluster number; After the sample data is extracted, the present application clusters the multiple sample data, thereby clustering the sample data with similar values ​​of the first target parameter and the second target parameter. The clustering method of the present application is as follows: (1) When the first target parameter and the second target parameter are both numeric or vector types, construct a two-dimensional parameter , and for the two-dimensional parameters Perform density clustering to obtain multiple clusters ; If the first target parameter and the second target parameter are both of numerical type and vector type, the DBSCAN clustering algorithm can be directly used for clustering.

[0034] If one of the first target parameter and the second target parameter is a numeric type and the other is a vector type, the first target parameter and the second target parameter are first combined to generate a multidimensional vector, and then the multidimensional vector is clustered. After clustering is completed, the numeric type target parameter can be separated.

[0035] For example, the word vector converted in the previous text is 100-dimensional, and the values ​​of the two target parameters are combined into a 101-dimensional vector for clustering. After clustering is completed, the parameters of the 101th dimension of all data samples in the cluster are separated to complete the restoration.

[0036] (2) When one of the first target parameter and the second target parameter is of character type and the other is of numeric type or vector type; Divide into multiple data units; in each data unit, multiple sample data are sorted based on the numerical type or vector type. Perform clustering to obtain multiple clusters .

[0037] If one of the first target parameter and the second target parameter is of character type, the target parameter of character type is first divided, and then the other target parameter is clustered.

[0038] For example, the first target parameter is the person who receives the materials, and the second target parameter is the word vector of the material name. In this case, multiple data units are first divided according to the person who receives the materials. Then the data samples in each data unit are clustered according to the word vector. Since the number of people authorized to receive the materials is valid, dividing the data according to the person who receives the materials will not over-divide the data.

[0039] The reason why clustering is used is that some parameters have too many values, and it is impossible to get a representative probability value when performing probability calculations later. For example, there may be hundreds of values ​​for material names, time, and quantity. Therefore, data samples with similar values ​​are clustered together through clustering as data samples of a certain habitual pattern. The probability value calculated based on the samples within the cluster can reflect the characteristics of the habitual pattern.

[0040] S125, for the plurality of clusters Marking is performed to obtain the multiple clusters The class label of the first target parameter in and the class label of the second target parameter ; Since the data samples are clustered in the previous article, in order to summarize the clustering results, this application labels the clusters. Specifically, it includes: (1) When the first target parameter or the second target parameter is a numeric type, calculate the average value of the target parameter whose data type is a numeric value within the cluster. and standard deviation , and based on the average and the standard deviation Constructing reference range labels ,in, is the range adjustment parameter; If the target parameter type within the cluster is a numeric type, then the range label can be constructed using the mean and standard deviation of the target parameter within the cluster. , for example, the time range is 9:00-10:00, and the quantity range is 200-300; Specifically, in general The value is 3, which means that the range label meets the three standard deviation principle.

[0041] (2) When the first target parameter or the second target parameter is of a vector type, a type label of the target parameter whose numerical type is a vector; If the target parameter in the cluster is a vector type, that is, the name of the material, since the names of materials with similar functions and principles are clustered in the previous article, they can be labeled according to their functions and principles. For example, disinfection materials, antibiotics, surgical instruments, etc.

[0042] (3) When the first target parameter or the second target parameter is of character type, the character itself is used as a label of the target parameter whose data type is character.

[0043] If it is a character type, use the original application as the label directly and use the name of the person who picked the materials.

[0044] S126, statistical cluster The number of samples within a cluster , in line with the category label The number of sample data And the category label The number of sample data ; S127, calculated separately The number of samples within a cluster Matches the category label The number of sample data Ratio , the number of samples within the cluster Matches the category label The number of sample data Ratio ; The meaning of the ratio is that when the value of the target parameter matches the label, the probability that the value of another target parameter matches the corresponding label. For example, when the picker is XX, there is an X% probability that the pick type will be surgical instruments. Therefore, by calculating the ratio, the application can find various candidate pick habit features and the probability of candidate pick habit features.

[0045] S128, in ratio When it is greater than the preset ratio threshold, the category label is determined exist The characteristics of the material picking habits; in the ratio When it is greater than the preset ratio threshold, the category label is determined exist The picking habit characteristics of the materials; and the ratio is used as the confidence of the picking habit characteristics.

[0046] If the probability is greater than a preset ratio threshold, such as 10%, then it can be determined that this type of material picking habit feature exists in past material picking records.

[0047] In addition, for the confidence level, even if the same collection habit features are used, the confidence level is different from the perspective of different category labels. For example, for the collection person XX, there is a 30% probability that he will collect medical supplies between 9:00 and 10:00 in the morning. However, during the time period between 9:00 and 10:00 in the morning, there are many people who collect medical supplies, and the probability that they are the collection person XX is only 10%. Therefore, it is necessary to calculate the collection habit features of different labels separately. This is also very important for the subsequent matching process.

[0048] S130, performing image and text recognition on the material picking record to obtain preliminary material picking data; and verifying the preliminary material picking data to obtain an incomplete record; In this application, a scanner or multi-function printer is used to scan paper forms into digital image files in PDF or common image formats such as JPEG, PNG, etc. Then the PDF or image format files are converted into text using the OCR service. At this time, the text is not 100% correct, so it is necessary to manually check and correct some common errors to obtain preliminary collection data.

[0049] During the manual verification process, some errors that cannot be modified can be marked or deleted, and marked incomplete records can be obtained by taking advantage of missing, unclear handwriting, smears, etc.

[0050] S140, matching the incomplete record with the multiple material picking habit features based on the known data of the incomplete record, and when the incomplete record matches any target material picking habit feature, completing the incomplete record based on the target material picking habit feature and the confidence of the target material picking habit feature.

[0051] In the present application, a variety of material picking habit features are matched based on known data of incomplete records, and any material picking habit feature whose label matches the known data is used as the target material picking habit feature.

[0052] For example, the incomplete record is: As can be seen, the picker information is missing;

[0053] At this time, multiple material picking habit features are matched with the known data "Material picking time: 9:30", "Material name: Mask", and "Quantity: 200" to obtain multiple target material picking habit features that match the above known data.

[0054] For example, the target picking habit characteristics include: {Material collection time: 9:00-10:00; Material collection person: A}; The confidence level from the material collection time perspective is: 15%; {Material collection time: 9:10-10:00; Material collection person: B}; The confidence level from the perspective of material collection time is: 18%; {Material name: daily consumables; Material picker: A}; The confidence level from the material name perspective is: 23%; {Material name: daily consumables; Material picker: B}; The confidence level from the material name perspective is: 11%; {Data volume: 150-300; Material picker: A}; Confidence level from the material name perspective: 15%; {Data volume: 100-250; Material picker: B}; Confidence level from the material name perspective: 28%; Then, the total confidence of multiple candidate results can be calculated based on the confidence of the target picking habit characteristics, that is, the candidate labels of the incomplete records can be determined based on the target picking habit characteristics and known data. And candidate labels Confidence , ,in, For the The confidence level corresponding to the target picking habit characteristics of known data; Take the above target picking habit characteristics as an example; The confidence level of the incomplete parameter A in the incomplete record is: 15%+23%+15%=53%; The confidence level of the incomplete parameter B in the incomplete record is: 18%+11%+28%=57%; Then, the target tags with a confidence greater than the confidence threshold are selected; the confidence threshold in this embodiment is set to 50%, which means that both A and B can be used as target tags; Finally, the incomplete record is completed based on the target label.

[0055] Both A and B are automatically filled in to the missing parts and the confidence level is marked; For example, the completed picking record is:

[0056] Since the real result cannot be directly inferred in this application, it is just to find the most likely result based on the existing clues. Therefore, this application will use the most confident candidate tags. Send to the corresponding object for confirmation. The target object is the picker in the record. If the incomplete part is the picker, it will be sent to multiple inferred candidate pickers for confirmation.

[0057] If confirmation information is received from the candidate picker, the real record can be restored to the table.

[0058] The above process can greatly improve the information repair efficiency of information management personnel, help information management personnel narrow the confirmation scope of incomplete information to a smaller range, and reduce the communication costs caused by confirmation and supplementation work.

[0059] The present invention provides a method for managing inventory information of medical supplies, which extracts a variety of existing habit features from complete historical material collection data. At the same time, the current material collection records are verified. If there are incomplete records such as missing information and unidentifiable handwriting, the material collection habit features are used to match the known information of the material collection records. If a match can be made, the target material collection habit features and the confidence of the target material collection habit features are used to complete the incomplete records. The present application uses the analysis of historical data to obtain material collection habit features, and uses the material collection habit features to automatically repair the missing registration data, thereby effectively reducing the confirmation workload of management personnel.

[0060] like Figure 3 As shown, the present application provides an inventory information management system for medical supplies, including: An acquisition module is used to acquire the material collection records and historical material collection data of the current management cycle, wherein the material collection records are paper materials and the historical material collection data are electronic data; A feature extraction module, used to extract multiple material picking habit features and confidence levels of the multiple material picking habit features from the historical material picking data; A verification module is used to perform image and text recognition on the material collection record to obtain preliminary material collection data; and verify the preliminary material collection data to obtain an incomplete record; A data completion module is used to match the incomplete record with the multiple material picking habit features based on the known data of the incomplete record, and when the incomplete record matches any target material picking habit feature, complete the incomplete record based on the target material picking habit feature and the confidence of the target material picking habit feature.

[0061] The inventory information management system of medical supplies of the present invention extracts various existing habit features from the complete historical material collection data. At the same time, the current material collection records are verified. If there are incomplete records such as missing information and unconfirmed handwriting, the material collection habit features are used to match the known information of the material collection records. If a match can be made, the target material collection habit features and the confidence of the target material collection habit features are used to complete the incomplete records. The present application uses the analysis of historical data to obtain the material collection habit features, and uses the material collection habit features to automatically repair the missing registration data, thereby effectively reducing the confirmation workload of management personnel.

[0062] This embodiment also provides an electronic terminal, including: a processor and a memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes any one of the methods in this embodiment.

[0063] The computer-readable storage medium in this embodiment can be understood by ordinary technicians in this field: all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to the computer program. The aforementioned computer program can be stored in a computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other media that can store program codes.

[0064] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run computer programs so that the electronic terminal executes each step of the above method.

[0065] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0066] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0067] In the above-mentioned embodiments, although the present invention has been described in conjunction with the specific embodiments of the present invention, many replacements, modifications and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. The embodiments of the present invention are intended to cover all such replacements, modifications and variations falling within the broad scope of the appended claims.

[0068] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for managing inventory information of medical supplies, characterized in that: Includes steps: Obtain the material collection records and historical material collection data of the current management cycle, wherein the material collection records are paper materials and the historical material collection data are electronic data; Extracting a plurality of material picking habit features and confidence levels of the plurality of material picking habit features from the historical material picking data; Performing image and text recognition on the material collection record to obtain preliminary material collection data; and verifying the preliminary material collection data to obtain an incomplete record; The incomplete record is matched with the multiple material picking habit features based on the known data of the incomplete record, and when the incomplete record matches any target material picking habit feature, the incomplete record is completed based on the target material picking habit feature and the confidence of the target material picking habit feature.

2. A method for managing inventory information of medical supplies according to claim 1, characterized in that: The historical material picking data includes a plurality of material picking record data, each of which includes values ​​of a plurality of parameters, wherein a plurality of material picking habit features and a plurality of confidence levels of the material picking habits are extracted from the historical material picking data, including: Remove useless parameters from historical material collection data to obtain target parameters; Combine any two parameters to get the sample data template ,in, represents the first target parameter, represents the second target parameter; Based on the sample data template Extract multiple sample data from the historical material picking data , Indicates The value of the first target parameter, Indicates the value of the second target parameter; For multiple sample data Perform clustering to obtain multiple clusters ,in, is the cluster number; Based on the multiple clusters Verify the characteristics of picking habits and the confidence of picking habits.

3. A method for managing inventory information of medical supplies according to claim 2, characterized in that: Based on the multiple clusters Verify the characteristics of picking habits and the confidence of picking habits, including: For the plurality of clusters Marking is performed to obtain the multiple clusters The class label of the first target parameter in and the class label of the second target parameter ; Statistical Clusters The number of samples within a cluster , in line with the category label The number of sample data And the category label The number of sample data ; Calculate separately The number of samples within a cluster Matches the category label The number of sample data Ratio , the number of samples within the cluster Matches the category label The number of sample data Ratio ; In the ratio When it is greater than the preset ratio threshold, the category label is determined exist The characteristics of the material picking habits; in the ratio When it is greater than the preset ratio threshold, the category label is determined exist The picking habit characteristics of the materials; and the ratio is used as the confidence of the picking habit characteristics.

4. A method for managing inventory information of medical supplies according to claim 2, characterized in that: Based on the sample data template Extract multiple sample data from the historical material picking data ,include: When the sample data template When any parameter in the sample data template is not a medical material name, Construct sample data with the values ​​of parameters belonging to the same picking record ; When the sample data template When the stored parameter is the name of a medical material, the name of the medical material is vectorized and based on the sample data template Construct sample data with the values ​​of parameters belonging to the same picking record .

5. A method for managing inventory information of medical supplies according to claim 2, characterized in that: For multiple sample data Perform clustering to obtain multiple clusters ,include: When the first target parameter and the second target parameter are both numeric or vector types, construct a two-dimensional parameter , and for the two-dimensional parameters Perform density clustering to obtain multiple clusters ; When one of the first target parameter and the second target parameter is of character type and the other is of numeric type or vector type; for multiple sample data based on the character type target parameter Divide into multiple data units; in each data unit, multiple sample data are sorted based on the numerical type or vector type. Perform clustering to obtain multiple clusters .

6. A method for managing inventory information of medical supplies according to claim 3, characterized in that: For the plurality of clusters Marking is performed to obtain the multiple clusters The class label of the first target parameter in and the class label of the second target parameter ,include: When the first target parameter or the second target parameter is a numeric type, calculate the average value of the target parameter whose data type is a numeric value within the cluster and standard deviation , and based on the average and the standard deviation Constructing reference range labels ,in, is the range adjustment parameter; When the first target parameter or the second target parameter is of a vector type, a type label of the target parameter whose numerical type is a vector; When the first target parameter or the second target parameter is of character type, the character itself is used as a label of the target parameter whose data type is character.

7. A method for managing inventory information of medical supplies according to claim 3, characterized in that: Matching the incomplete record with the multiple picking habit features based on the known data of the incomplete record includes: The known data of the incomplete record is matched with the labels of the multiple material picking habit features, and the material picking habit features whose labels match the known data are used as target material picking habit features.

8. A method for managing inventory information of medical supplies according to claim 7, characterized in that: The incomplete record is completed based on the target picking habit feature and the confidence of the target picking habit feature, including: Determine candidate labels for incomplete records based on the target picking habit characteristics and known data And candidate labels Confidence , ,in , for The confidence level corresponding to the target picking habit characteristics of known data; Filter out target labels whose confidence is greater than the confidence threshold; The incomplete record is completed based on the target tag.

9. A method for managing inventory information of medical supplies according to claim 8, characterized in that: Also includes: The multiple candidate labels with the highest confidence Sent to the corresponding object for confirmation.

10. A medical supplies inventory information management system, characterized in that: include: An acquisition module is used to acquire the material collection records and historical material collection data of the current management cycle, wherein the material collection records are paper materials and the historical material collection data are electronic data; A feature extraction module, used to extract multiple material picking habit features and confidence levels of the multiple material picking habit features from the historical material picking data; A verification module is used to perform image and text recognition on the material collection record to obtain preliminary material collection data; and verify the preliminary material collection data to obtain an incomplete record; A data completion module is used to match the incomplete record with the multiple material picking habit features based on the known data of the incomplete record, and when the incomplete record matches any target material picking habit feature, complete the incomplete record based on the target material picking habit feature and the confidence of the target material picking habit feature.

Images