Method, device and equipment for generating pellet core physical fingerprint spectrum and product

Abstract

The invention provides a method, a device, equipment and a product for generating a pellet core physical fingerprint spectrum. The method comprises the following steps: firstly, determining a physical variable for representing a pellet core physical state and a spectral variable for representing a pellet core spectral state; and further obtaining a physical variable value corresponding to the physical variable and a spectral variable value corresponding to the spectral variable of the sample pellet core, constructing a training data set, training to obtain a first ensemble learning neural network model, and finally using a predicted physical variable value output by the first ensemble learning neural network model to obtain a prediction result. The method has the advantages that the pellet core quality detection speed is high, few samples are used, and the pellet core detection result of the output physical fingerprint spectrum is more intuitive.

Description

Technical Field

[0001] The invention belongs to the field of drug production, and in particular relates to a method, a device, equipment and a product for generating a physical fingerprint spectrum of a pill core. Background Art

[0002] In the production quality evaluation of sustained-release microcapsules, dissolution properties are one of the key evaluation indicators. By changing the thickness of the sustained-release layer and the particle size of the microcapsules, different dissolution properties can be achieved. The pellet core is a key excipient for sustained-release microcapsules, and its quality will have a significant impact on the dissolution properties of the preparation. Therefore, obtaining a uniform and stable medicinal pellet core is an important prerequisite for ensuring the quality of the preparation. Physical parameters are important indicators for evaluating the quality of the pellet core. However, under existing technical conditions, the technical methods currently used to detect the physical parameters of the pellet core are relatively time-consuming and require a large amount of samples. There are problems such as low detection efficiency, high sample consumption, and non-intuitive inspection results. Summary of the invention

[0003] In view of this, the present invention aims to overcome the defects in the prior art and proposes a method, device, equipment and product for generating a physical fingerprint of a pill core.

[0004] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0005] In a first aspect, the present invention discloses a method for generating a physical fingerprint of a pill core, the method comprising:

[0006] determining a physical variable for characterizing a physical state of the pellet core and a spectral variable for characterizing a spectral state of the pellet core;

[0007] Obtaining physical variable values ​​corresponding to the physical variables of the sample pellet core and spectral variable values ​​corresponding to the spectral variables;

[0008] A training data set is constructed based on the spectral variable values ​​as input variables and the corresponding physical variable values ​​as output variables;

[0009] Using the training data set, training the initial ensemble learning neural network model to obtain the first ensemble learning neural network model;

[0010] Input the spectral variable value of the pill core for which the physical fingerprint spectrum needs to be generated into the first integrated learning neural network model to obtain the corresponding predicted physical variable value;

[0011] According to the predicted physical variable values, the corresponding physical fingerprint map is generated.

[0012] In some embodiments of the present invention, determining the physical variables used to characterize the physical state of the pill core and the spectral variables used to characterize the spectral state of the pill core includes: using the Spearman correlation analysis calculation method to calculate the correlation coefficients between all physical variables that can be obtained from the sample pill core, determining the physical variable whose absolute value of the correlation coefficient with other physical variables exceeds the first threshold the most times as the basic physical variable, and sorting the remaining physical variables in ascending order according to the absolute values ​​of the correlation coefficients with the basic physical variables, taking the physical variables corresponding to the absolute values ​​of the first N correlation coefficients as additional physical variables, where N is a positive integer; the set of basic physical variables and additional physical variables is used as the physical variable used to characterize the physical state of the pill core.

[0013] In some embodiments of the present invention, determining the physical variables used to characterize the physical state of the pill core and the spectral variables used to characterize the spectral state of the pill core includes: using the Spearman correlation analysis calculation method to calculate the correlation coefficient between each spectral variable of the sample pill core and any physical variable used to characterize the physical state of the pill core, the correlation coefficient is divided into a positive relationship array and a negative relationship array, the positive relationship array and the negative relationship array are sorted in descending order according to the absolute value of the correlation coefficient, and the top M spectral variables in the sorting are selected as the spectral variable group corresponding to this physical variable, and the set of spectral variable groups corresponding to all physical variables used to characterize the physical state of the pill core is used as the spectral variable used to characterize the spectral state of the pill core.

[0014] In some embodiments of the present invention, the initial ensemble learning neural network model is trained using a training data set to obtain a first ensemble learning neural network model, including: randomly extracting a specified number of samples from the training data set with replacement, constructing multiple groups of sub-training set data, each group of sub-training set data corresponding to the training of a sub-neural network model, and averaging the prediction results of all sub-neural network models as the output of the first ensemble learning neural network model.

[0015] In some embodiments of the present invention, obtaining the physical variable values ​​corresponding to the physical variables of the sample pellet core and the spectral variable values ​​corresponding to the spectral variables includes: preprocessing the spectral data matrix corresponding to the sample pellet core using a standard normal transformation method.

[0016] In some embodiments of the present invention, generating a corresponding physical fingerprint spectrum according to the predicted physical variable value includes: converting the predicted physical variable value into a set numerical range, and generating the physical fingerprint spectrum based on the converted predicted physical variable value.

[0017] In some embodiments of the present invention, the spectral variables used to characterize the spectral state of the pellet core include: the spectral variables are obtained based on near-infrared spectroscopy.

[0018] In a second aspect, the present invention discloses a device for generating a physical fingerprint of a pill core, the device comprising:

[0019] A determination module, used to determine physical variables used to characterize the physical state of the pellet core and spectral variables used to characterize the spectral state of the pellet core;

[0020] An acquisition module, used for acquiring physical variable values ​​corresponding to the physical variables of the sample pellet core and spectral variable values ​​corresponding to the spectral variables;

[0021] A training data set construction module is used to construct a training data set based on spectral variable values ​​as input variables and corresponding physical variable values ​​as output variables;

[0022] A training module, used to train the initial ensemble learning neural network model using the training data set to obtain a first ensemble learning neural network model;

[0023] A prediction module is used to input the spectral variable values ​​of the pill core for which a physical fingerprint spectrum needs to be generated into the first integrated learning neural network model to obtain the corresponding predicted physical variable values;

[0024] The generation module is used to generate the corresponding physical fingerprint spectrum according to the predicted physical variable value.

[0025] In a third aspect, the present invention discloses an electronic device, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above method.

[0026] In a fourth aspect, the present invention discloses a computer program product, including a computer program, which implements the above method when executed by a processor.

[0027] Compared with the prior art, the present invention has the following advantages:

[0028] The present invention discloses a method, device, equipment and product for generating a physical fingerprint spectrum of a pill core. The method comprises the following steps: firstly determining physical variables for characterizing the physical state of a pill core and spectral variables for characterizing the spectral state of a pill core, further obtaining physical variable values ​​corresponding to the physical variables of a sample pill core and spectral variable values ​​corresponding to the spectral variables, constructing a training data set, training to obtain a first integrated learning neural network model, and finally using the predicted physical variable values ​​output by the first integrated learning neural network model to generate a corresponding physical fingerprint spectrum. The method has the advantages of fast pill core quality detection speed, less sample usage and more intuitive pill core detection results of the output physical fingerprint spectrum. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0030] In the attached picture:

[0031] Figure 1 A schematic diagram of an application scenario of a method for generating a physical fingerprint of a pill core according to an embodiment of the present invention;

[0032] Figure 2 A schematic diagram of a method for generating a physical fingerprint of a pill core according to an embodiment of the present invention;

[0033] Figure 3 It is a schematic diagram of another embodiment of a method for generating a physical fingerprint of a pill core according to an embodiment of the present invention;

[0034] Figure 4 This is a schematic diagram of another embodiment of a method for generating a physical fingerprint of a pill core according to an embodiment of the present invention;

[0035] Figure 5 This is a schematic diagram of another embodiment of a method for generating a physical fingerprint of a pill core according to an embodiment of the present invention;

[0036] Figure 6 A schematic diagram of a device for generating a physical fingerprint of a pill core according to an embodiment of the present invention;

[0037] Figure 7 The present invention is a schematic diagram of an electronic device for generating a physical fingerprint of a pill core according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0039] Under existing technical conditions, the technical methods currently used to detect the physical parameters of pellet cores are relatively time-consuming and require the consumption of a large amount of samples. There are problems such as low detection efficiency, large sample consumption, and non-intuitive inspection results.

[0040] like Figures 1 to 7As shown, the present invention discloses a method, device, equipment and product for generating a physical fingerprint spectrum of a pill core, by first determining a physical variable for characterizing the physical state of the pill core and a spectral variable for characterizing the spectral state of the pill core, further obtaining the physical variable values ​​corresponding to the physical variables of the sample pill core and the spectral variable values ​​corresponding to the spectral variables, constructing a training data set, training to obtain a first integrated learning neural network model, and finally using the predicted physical variable values ​​output by the first integrated learning neural network model to generate a corresponding physical fingerprint spectrum, which has the function of directly detecting the physical state of the pill core and directly generating a physical fingerprint spectrum, effectively improving the detection speed and detection quality, and greatly reducing the detection consumption of samples.

[0041] like Figure 1 As shown, an application scenario of a method for generating a physical fingerprint spectrum of a pill core is disclosed. The method can be installed in devices such as a desktop computer 101, a portable computer 102, and a mobile terminal 103. The spectral variable values ​​of the pill core whose quality needs to be tested are input, and the spectral variable values ​​are generated by a spectrometer. The corresponding physical fingerprint spectrum of the pill core is output to achieve quality detection of the pill core.

[0042] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0043] The present invention discloses a method for generating a physical fingerprint of a pill core, the method comprising:

[0044] Step S201, determining a physical variable for characterizing a physical state of a pill core and a spectral variable for characterizing a spectral state of a pill core;

[0045] Exemplarily, the physical variable may be one or more of bulk density, tap density, angle of repose, angle of collapse, particle size (D50), particle size span (Span), loss on drying, hygroscopicity, roundness, porosity between pellets, Carr index and Hausner ratio;

[0046] Spectral variables are obtained based on near infrared (NIR) spectroscopy.

[0047] In this embodiment, step S201 is to determine the most representative physical variable and the most representative spectral variable from all available physical variables of the pellet core.

[0048] Step S202, obtaining the physical variable values ​​corresponding to the physical variables of the sample pellet core and the spectral variable values ​​corresponding to the spectral variables;

[0049] Exemplarily, the production process of the sample pellet core is as follows:

[0050] Take 500g of microcrystalline cellulose powder for making pellet cores into a mixer, then measure 750mL of deionized water and slowly pour it into the microcrystalline cellulose. The moistened microcrystalline cellulose is stirred at a speed of 28rpm for 6min to obtain a wet material with a water content of about 60%. The above wet material preparation operation is repeated 5 times. Subsequently, an appropriate amount of wet material is successively loaded into the extrusion module of the extrusion spheronizer, and microcrystalline cellulose strips are extruded through dies with diameters of 0.3mm, 0.5mm, 0.8mm, 1.0mm, 1.5mm, 2.0mm and 2.5mm, respectively. When using dies with diameters of 0.3mm and 0.5mm, the feeding rate is set to 1.0kg / h; when using dies with diameters of 0.8mm and 1.0mm, the feeding rate is set to 1.5kg / h; when using dies with diameters of 1.5mm, 2.0mm and 2.5mm, the feeding rate is set to 2.5kg / h. After the extrudates under each mold were prepared, the extrudates of each specification were spheronized at a speed of 700 rpm for 8 min using the spheronization module of the extrusion spheronization machine to prepare micropellet cores. The batch size of the spheronization operation was set to 750 g.

[0051] After the pellet cores were made by extrusion spheronization, standard drug sieves with pore sizes of 250μm, 355μm, 450μm, 600μm, 710μm, 900μm, 1000μm, 1250μm, 1500μm and 2000μm (in accordance with the 2020 edition of the Pharmacopoeia) were used to divide the pellet cores into 10 samples with different particle size ranges. The 11th to 15th batches of samples were obtained by precisely weighing the first 10 batches of samples according to the weight ratio set in Table 1.

[0052] Table 1 Particle size distribution information of micropellet core

[0053]

[0054] By preparing 5 batches of pellet core samples with a richer range of particle sizes, the representativeness of the training set samples was improved. It should be noted that the mass of each batch of pellet cores was 200 g.

[0055] The 15 batches of samples were placed in an oven and dried at 60°C. When the drying process reached 2, 3, 4.5 and 9 hours, a quarter of each batch of samples was collected. The moisture content information of samples at different drying times is shown in Table 2.

[0056] Table 2 Moisture content of pellet core samples at different drying times

[0057]

[0058]

[0059] Finally, a total of 60 batches of sample pellets with 4 different moisture content ranges were prepared.

[0060] Step S203, constructing a training data set based on the spectral variable values ​​as input variables and the corresponding physical variable values ​​as output variables;

[0061] Step S204, using the training data set to train the initial ensemble learning neural network model to obtain a first ensemble learning neural network model;

[0062] In this embodiment, the initial ensemble learning neural network model may be a Bagging ensemble learning method.

[0063] Step S205, inputting the spectral variable values ​​of the pill core for which the physical fingerprint spectrum needs to be generated into the first integrated learning neural network model to obtain the corresponding predicted physical variable values;

[0064] Step S206, generating a corresponding physical fingerprint map according to the predicted physical variable value.

[0065] In this embodiment, generating a corresponding physical fingerprint map according to the predicted physical variable value is understandable to those skilled in the art.

[0066] Based on the previous embodiment, in another embodiment of the present invention, Figure 3 As shown, determining the physical variables used to characterize the physical state of the pill core and the spectral variables used to characterize the spectral state of the pill core includes: using the Spearman correlation analysis calculation method to calculate the correlation coefficients between all physical variables that can be obtained from the sample pill core, determining the physical variable whose absolute value of the correlation coefficient with other physical variables exceeds the first threshold the most times as the basic physical variable 301, and sorting the rest in ascending order according to the absolute value of the correlation coefficient with the basic physical variable 301, taking the physical variables corresponding to the absolute values ​​of the first N correlation coefficients as the additional physical variables 302, where N is a positive integer; the set of the basic physical variables 301 and the additional physical variables 302 is used as the physical variable used to characterize the physical state of the pill core.

[0067] Exemplarily, N is set to 2; basic physical variables 301 represent key physical variables that are strongly correlated with most physical variables, and additional physical variables 302 represent physical variables that are relatively weakly correlated or have no correlation with basic physical variables 301. The set of basic physical variables 301 and additional physical variables 302 is used as physical variables for characterizing the physical state of the pill core, which makes it possible to characterize the physical state of the pill core with relatively few physical variables, greatly reduces the amount of training data for subsequent models, and effectively improves the speed of using the first integrated learning neural network model to obtain the corresponding predicted physical variable values.

[0068] In this embodiment, the basic physical variable 301 is the angle of repose, and the additional physical variable 302 is the particle size (D50) and the roundness.

[0069] Based on the previous embodiment, in another embodiment of the present invention, Figure 4 As shown, determining the physical variables used to characterize the physical state of the pill core and the spectral variables used to characterize the spectral state of the pill core includes: using the Spearman correlation analysis calculation method to calculate the correlation coefficient between each spectral variable of the sample pill core and any physical variable used to characterize the physical state of the pill core, the correlation coefficient is divided into a positive relationship array and a negative relationship array, the positive relationship array and the negative relationship array are sorted in descending order according to the absolute value of the correlation coefficient, and the top M spectral variables are selected as the spectral variable group corresponding to this physical variable, and the set of spectral variable groups corresponding to all physical variables used to characterize the physical state of the pill core is used as the spectral variable used to characterize the spectral state of the pill core.

[0070] In this embodiment, M is set to 10;

[0071] In this embodiment, there are 20 spectral variables corresponding to the angle of repose, particle size (D50) and roundness, for a total of 60.

[0072] In this embodiment, a collection of spectral variable groups corresponding to all physical variables used to characterize the physical state of the pill core is used as the spectral variable for characterizing the spectral state of the pill core. This can greatly reduce the amount of training data for subsequent models while ensuring the characterization effect of the spectral variables on the spectral state of the pill core and the correlation with the three physical variables of repose angle, particle size (D50) and roundness. It also effectively improves the speed of obtaining the corresponding predicted physical variable values ​​using the first integrated learning neural network model.

[0073] Spearman correlation analysis is a non-parametric statistical method used to evaluate the correlation between two variables. This method is particularly suitable for processing nonlinear data. The process of calculating the correlation between two variables is mainly divided into three steps, as shown below.

[0074] S1: Data sorting: Sort the data of the two variables separately and assign ranks.

[0075] S2: Calculate the rank difference: For each column of data, calculate the rank difference (di) between the two variables.

[0076] S3: Calculate the Spearman correlation coefficient:

[0077]

[0078] Among them, ρ is the Spearman correlation coefficient, di is the rank difference of the i-th pair of data, and n is the number of samples. The value range of the Spearman correlation coefficient is -1 to 1. The strength of the correlation between two variables is determined by the absolute value of the correlation coefficient between the two variables, which is mainly divided into very strong correlation (0.8 to 1.0), strong correlation (0.6 to 0.8), moderate correlation (0.4 to 0.6), weak correlation (0.2 to 0.4) and no correlation (0 to 0.2).

[0079] In this embodiment, the correlation between two variables refers to the correlation between two physical variables or between a spectral variable and a physical variable.

[0080] After obtaining a suitable training data set, the initial ensemble learning neural network model for prediction can be trained as follows:

[0081] In another embodiment of the present invention, Figure 5 As shown, the initial ensemble learning neural network model is trained using the training data set 501 to obtain the first ensemble learning neural network model, including: randomly extracting a specified number of samples from the training data set 501 with replacement, constructing multiple groups of sub-training set data 502, each group of sub-training set data 502 corresponds to training a sub-neural network model 503, and the prediction results of all sub-neural network models 503 are averaged as the output of the first ensemble learning neural network model.

[0082] In this embodiment, the sub-neural network model 503 is a GRNN neural network model.

[0083] The initial integrated learning neural network model is as follows:

[0084] The training data set 501 is D:

[0085] D={(x1,y1),(x2,y2),...,(x m ,y m )}. m is the number of training data sets 501, x n is the input variable for modeling, which is the spectral variable value; n is the output variable, is the physical variable value; use Bagging method for training:

[0086] Randomly draw from D with replacement to generate B different sub-training set data 502, represented as {D1, D2, ..., D B}, where each D n contains n samples.

[0087] For each D i , train a sub-neural network model 503, the sub-neural network model 503 is represented by hi (x).

[0088] The final prediction result is obtained by i The prediction results of (x) are averaged to determine.

[0089]

[0090] in, is the final predicted physical variable value after integration processing, x is the input spectral variable value, h i (x) is the prediction result of the i-th sub-neural network model 503.

[0091] In this embodiment, ensemble learning can improve the adaptability and stability of the model to new data, making the model more robust when facing changes in data distribution and reducing the risks caused by overfitting or underfitting; ensemble learning can better deal with the uncertainty in the data, reduce the impact of data quality issues on a single model, improve the stability and reliability of the prediction, and achieve more accurate output results of sample physical variable values.

[0092] In another embodiment of the present invention, obtaining the physical variable values ​​corresponding to the physical variables of the sample pellet core and the spectral variable values ​​corresponding to the spectral variables includes: preprocessing the spectral data matrix corresponding to the sample pellet core using a standard normal transformation method.

[0093] In this embodiment, the spectral data matrix corresponding to the sample pellet core is preprocessed using the standard normal transformation method as follows:

[0094] The spectral variables are obtained from the spectral data matrix generated by the spectrometer when testing the sample pellet core;

[0095] The spectral data matrix is ​​X, where each row represents a sample and each column represents a feature (such as wavelength). For each sample x i , its standard normal transformation can be performed as follows:

[0096] Sample x i =(x i1 ,x i2 ,...x ip ), where p is the number of features, μ i is the sample mean and σ i is the sample standard deviation, x ij is the data point before the change, x , ij is the transformed data point.

[0097]

[0098]

[0099] In another embodiment of the present invention, generating a corresponding physical fingerprint spectrum according to the predicted physical variable value includes: converting the predicted physical variable value into a set numerical range, and generating the physical fingerprint spectrum based on the converted predicted physical variable value.

[0100] In this embodiment, the numerical range is between 0 and 10.

[0101] Converting the predicted physical variable values ​​to a set numerical range can help generate a better quality physical fingerprint, making it easier for users to understand the physical state of the pellet core and evaluate the quality of the pellet core.

[0102] In some embodiments of the present invention, the spectral variables used to characterize the spectral state of the pellet core include: the spectral variables are obtained based on near-infrared spectroscopy.

[0103] In this embodiment, the average Euclidean distances of spectral data collected under multiple near-infrared spectrum collection conditions are calculated, and the near-infrared spectrum collection condition with the smallest average Euclidean distance is selected as the selected near-infrared spectrum collection condition.

[0104] In this embodiment, the candidate near infrared spectrum acquisition conditions are shown in Table 3:

[0105] Table 3 Factors and levels considered in the optimization of NIR spectral acquisition conditions

[0106]

[0107] In this embodiment, the selected near infrared spectrum acquisition condition is Empty 1×-16cm -1 Resolution - 64 scans.

[0108] like Figure 6 As shown, the present invention also discloses a device for generating a physical fingerprint of a pill core, comprising:

[0109] A determination module 601 is used to determine a physical variable for characterizing a physical state of a pill core and a spectral variable for characterizing a spectral state of a pill core;

[0110] An acquisition module 602 is used to acquire physical variable values ​​corresponding to the physical variables of the sample pellet core and spectral variable values ​​corresponding to the spectral variables;

[0111] A training data set construction module 603 is used to construct a training data set based on the spectral variable value as an input variable and the corresponding physical variable value as an output variable;

[0112] A training module 604 is used to train the initial ensemble learning neural network model using the training data set to obtain a first ensemble learning neural network model;

[0113] Prediction module 605, used to input the spectral variable value of the pill core for which the physical fingerprint spectrum needs to be generated into the first integrated learning neural network model to obtain the corresponding predicted physical variable value;

[0114] The generation module 606 is used to generate a corresponding physical fingerprint spectrum according to the predicted physical variable value.

[0115] The present invention also discloses an electronic device, such as Figure 7 As shown, an embodiment is disclosed, which is a block diagram of an electronic device suitable for the above-mentioned method for generating a physical fingerprint of a pill core.

[0116] The electronic device 70 of this embodiment includes a processor 701, which can perform various appropriate actions and processes according to the program stored in the ROM 702 or the program loaded from the storage part 708 to the RAM 703. The processor 701 may include, for example, a general-purpose microprocessor, an instruction set processor and / or a related chipset and / or a dedicated microprocessor, etc. The processor 701 may also include an onboard memory for caching purposes. The processor 701 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0117] In RAM703, various programs and data required for the operation of electronic device 70 are stored. Processor 701, ROM702 and RAM703 are connected to each other through bus 704, and processor 701 performs various operations of the method flow according to the embodiment of the present invention by executing the program in ROM702 and / or RAM703. It should be noted that the program can also be stored in one or more memories other than ROM702 and RAM703, and processor 701 can also perform various operations of the method flow according to the embodiment of the present invention by executing the program stored in one or more memories.

[0118] According to an embodiment of the present invention, the electronic device 70 may further include an I / O interface 705, which is also connected to the bus 704. The electronic device 70 may further include one or more of the following components connected to the I / O interface 705: an input portion 706 including a keyboard, a mouse, etc.; an output portion 707 including a cathode ray tube, a liquid crystal display, and a speaker; a storage portion 708 including a hard disk, etc.; and a communication portion 709 including a network interface card such as a LAN card, a modem, etc. The communication portion 709 performs communication processing via a network such as the Internet. A drive 7010 is also connected to the I / O interface 705 as needed. A removable medium 7011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 7010 as needed, so that a computer program read therefrom is installed into the storage portion 708 as needed.

[0119] Embodiments of the present invention also include a computer program product.

[0120] The computer program product includes a computer program, which contains program codes for executing the method provided by the embodiment of the present invention. When the computer program product runs on an electronic device, the program codes are used to enable the electronic device to implement the method provided by the embodiment of the present invention.

[0121] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of a signal on a network medium. The program code included in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0122] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written by any combination of one or more programming languages, and specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages. Programming languages ​​include but are not limited to programming languages ​​such as Java, C++, python, C language or similar. The program code can be executed completely on the user computing device, partially on the user device, partially on the remote computing device, or completely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network or a wide area network, or can be connected to an external computing device.

[0123] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments and / or claims of the present invention can be combined and / or combined in various ways, even if such a combination or combination is not explicitly recorded in the present invention. In particular, without departing from the spirit and teaching of the present invention, the features described in the various embodiments and / or claims of the present invention may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present invention.

[0124] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although the embodiments are described above, this does not mean that the measures in the various embodiments cannot be used in combination. The scope of the present invention is limited by the attached claims and their equivalents. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. A method for generating a physical fingerprint of a pill core, characterized in that: The method comprises: determining a physical variable for characterizing a physical state of the pellet core and a spectral variable for characterizing a spectral state of the pellet core; Obtaining the physical variable value corresponding to the physical variable of the sample pellet core and the spectral variable value corresponding to the spectral variable; Constructing a training data set based on the spectral variable value as an input variable and the corresponding physical variable value as an output variable; Using the training data set, training an initial ensemble learning neural network model to obtain a first ensemble learning neural network model; Inputting the spectral variable value of the pill core for which a physical fingerprint spectrum needs to be generated into the first integrated learning neural network model to obtain the corresponding predicted physical variable value; According to the predicted physical variable value, a corresponding physical fingerprint spectrum is generated.

2. The method for generating a physical fingerprint of a pill core according to claim 1, characterized in that: The method of determining the physical variables used to characterize the physical state of the pill core and the spectral variables used to characterize the spectral state of the pill core includes: using the Spearman correlation analysis calculation method to calculate the correlation coefficients between all physical variables that can be obtained from the sample pill core, determining the physical variable whose absolute value of the correlation coefficient with other physical variables exceeds the first threshold the most times as the basic physical variable, and sorting the remaining physical variables in ascending order according to the absolute values ​​of the correlation coefficients with the basic physical variables, taking the physical variables corresponding to the first N absolute values ​​of the correlation coefficients as additional physical variables, where N is a positive integer; and the set of the basic physical variables and the additional physical variables as the physical variables used to characterize the physical state of the pill core.

3. A method for generating a physical fingerprint of a pill core according to claim 2, characterized in that: The method of determining the physical variables used to characterize the physical state of the pill core and the spectral variables used to characterize the spectral state of the pill core includes: using the Spearman correlation analysis calculation method to calculate the correlation coefficient between each spectral variable of the sample pill core and any physical variable used to characterize the physical state of the pill core, the correlation coefficient is divided into a positive relationship array and a negative relationship array, the positive relationship array and the negative relationship array are sorted in descending order according to the absolute value of the correlation coefficient, and the first M spectral variables in the sorting are selected as the spectral variable group corresponding to this physical variable, and the set of the spectral variable groups corresponding to all the physical variables used to characterize the physical state of the pill core is used as the spectral variable used to characterize the spectral state of the pill core.

4. The method for generating a physical fingerprint of a pill core according to claim 1, characterized in that: The method of using the training data set to train the initial ensemble learning neural network model to obtain the first ensemble learning neural network model includes: randomly extracting a specified number of samples from the training data set with replacement, constructing multiple groups of sub-training set data, each group of sub-training set data corresponding to the training of a sub-neural network model, and averaging the prediction results of all the sub-neural network models as the output of the first ensemble learning neural network model.

5. The method for generating a physical fingerprint of a pill core according to claim 1, characterized in that: Before obtaining the physical variable values ​​corresponding to the physical variables of the sample pellet core and the spectral variable values ​​corresponding to the spectral variables, the method includes: preprocessing the spectral data matrix corresponding to the sample pellet core using a standard normal transformation method.

6. The method for generating a physical fingerprint of a pill core according to claim 1, characterized in that: Generating the corresponding physical fingerprint spectrum according to the predicted physical variable value includes: converting the predicted physical variable value into a set numerical range, and generating the physical fingerprint spectrum based on the converted predicted physical variable value.

7. The method for generating a physical fingerprint of a pill core according to claim 1, characterized in that: The spectral variables used to characterize the spectral state of the pellet core include: the spectral variables are obtained based on near-infrared spectroscopy.

8. A device for generating a physical fingerprint of a pill core, characterized in that: The device comprises: A determination module, used to determine physical variables used to characterize the physical state of the pellet core and spectral variables used to characterize the spectral state of the pellet core; An acquisition module, used for acquiring the physical variable values ​​corresponding to the physical variables of the sample pellet core and the spectral variable values ​​corresponding to the spectral variables; A training data set construction module is used to construct a training data set based on the spectral variable value as an input variable and the corresponding physical variable value as an output variable; A training module, used to train an initial ensemble learning neural network model using the training data set to obtain a first ensemble learning neural network model; A prediction module, used for inputting the spectral variable value of the pill core for which a physical fingerprint spectrum needs to be generated into the first integrated learning neural network model to obtain the corresponding predicted physical variable value; A generation module is used to generate a corresponding physical fingerprint spectrum according to the predicted physical variable value.

9. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to perform the method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images