CN112101481A - Method, device and equipment for screening influence factors of target object and storage medium - Google Patents
Method, device and equipment for screening influence factors of target object and storage medium Download PDFInfo
- Publication number
- CN112101481A CN112101481A CN202011097292.7A CN202011097292A CN112101481A CN 112101481 A CN112101481 A CN 112101481A CN 202011097292 A CN202011097292 A CN 202011097292A CN 112101481 A CN112101481 A CN 112101481A
- Authority
- CN
- China
- Prior art keywords
- influence
- target
- factor
- influence factors
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012216 screening Methods 0.000 title claims abstract description 45
- 238000013145 classification model Methods 0.000 claims abstract description 78
- 230000006870 function Effects 0.000 claims description 29
- 238000012896 Statistical algorithm Methods 0.000 claims description 17
- 238000011176 pooling Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 11
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 11
- 230000004913 activation Effects 0.000 claims description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 9
- 239000013076 target substance Substances 0.000 claims description 5
- 238000012163 sequencing technique Methods 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 2
- 238000013136 deep learning model Methods 0.000 abstract description 8
- 238000005516 engineering process Methods 0.000 abstract description 3
- 238000007405 data analysis Methods 0.000 abstract description 2
- 102000003712 Complement factor B Human genes 0.000 description 10
- 108090000056 Complement factor B Proteins 0.000 description 10
- 235000013399 edible fruits Nutrition 0.000 description 10
- 241000220324 Pyrus Species 0.000 description 6
- 238000007726 management method Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 4
- 235000014443 Pyrus communis Nutrition 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000012800 visualization Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000013481 data capture Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 235000021017 pears Nutrition 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention relates to a data analysis technology, and discloses a method for screening an influence factor of a target object, which comprises the following steps: inputting the standard image into a classification model to obtain a category predicted value with the standard image as a target category; calculating the category predicted value to obtain the number of influence factors to be selected; calculating the multiple influence factors by using a classification model to obtain an influence predicted value of each influence factor, and selecting the number of influence factors from the preset multiple influence factors as a target factor set according to the influence predicted value; and calculating the label value of each influence factor in the target factor set, and selecting the influence factor with the label value as a preset label threshold value from the target factor set as a standard influence factor. The invention also provides a device for screening the influence factors of the target object and a computer readable storage medium. In addition, the invention also relates to a block chain technology, and the standard image can be stored in the block chain node. The method can obtain the influence factors in the input data of the deep learning model.
Description
Technical Field
The invention relates to the technical field of data analysis, in particular to a method and a device for screening influence factors of a target object, electronic equipment and a computer-readable storage medium.
Background
The interpretability of deep learning models for images has been a major concern for machine learning. Specifically, the interpretability of the deep learning model plays an important role in improving the model, the credibility of the model and the transparency of the model.
The feature visualization is a commonly used interpretation method for a deep learning model of an image at present, influence factors in model input data and influence of each influence factor on a model prediction result are searched in the feature visualization process and are visualized, and attention points of the model on the input data in the prediction process can be explained to a certain extent. In practice, however, the feature visualization may only show the range of interest (e.g., location coordinates) of the impact factors in the input data that the model is interested in making predictions of the data tags for the input data, but may not reflect the predictions that the model specifically makes based on which specific impact factors in the input data. For example, when generating prediction labels for image recognition, the feature visualization can only display the pixel positions of the influence factors of interest to the model in the image, but cannot reflect the label predictions that the model specifically makes according to which specific influence factors in the image.
Disclosure of Invention
The invention provides a method and a device for screening influence factors of a target object, electronic equipment and a computer readable storage medium, and mainly aims to obtain the influence factors in input data of a deep learning model.
In order to achieve the above object, the present invention provides a method for screening an influence factor of a target, comprising:
acquiring a standard image, and inputting the standard image into a pre-constructed classification model to obtain a class prediction value of the standard image as a target class;
calculating the category predicted value by using a preset number statistical algorithm to obtain the number of the influence factors to be selected;
calculating a plurality of preset influence factors by using the classification model to obtain an influence predicted value of each influence factor, and selecting the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence predicted value;
and calculating the label value of each influence factor in the target factor set, and selecting the influence factor with the label value as a preset label threshold value from the target factor set as a standard influence factor.
Optionally, the classifying model is a convolutional neural network, and the inputting the standard image into a pre-constructed classifying model to obtain a class prediction value with the standard image as a target class includes:
performing convolution on the standard image by using a main task network of the convolutional neural network to obtain a convolutional image;
pooling the convolution images to obtain characteristic images;
carrying out full-connection processing on the characteristic image to obtain a full-connection image;
and calculating the category predicted value of the fully-connected image as the target category by using an activation function.
Optionally, the selecting, according to the influence prediction value, the number of influence factors from the preset plurality of influence factors as a target factor set includes:
sequencing the plurality of influence factors according to the sequence of the influence predicted values from large to small to obtain an influence factor sequence;
and selecting the number of influence factors from the influence factor sequence as a target factor set according to the sequence from front to back from large to small.
Optionally, the calculating the category prediction value by using a preset number statistical algorithm to obtain the number of the impact factors to be selected includes:
calculating the category predicted value by using the following quantity statistical algorithm to obtain the quantity of the influence factors to be selected:
w＝ceil(m×sk)
wherein ceil is integer arithmetic, w is the number to be selected, m is-the number of a plurality of preset influence factors, skAnd the standard image is the category predicted value of the target category k.
Optionally, the calculating a label value of each influence factor in the target factor set includes:
calculating a label value of each impact factor in the target factor set using the following label value algorithm:
wherein A isiIs the label value of the target factor, i is the label of the target factor, qiIs a preset standard label.
Optionally, after the standard image is input into a pre-constructed classification model to obtain a class prediction value of the standard image as a target class, the method further includes:
calculating a loss value between the category prediction value and the target category;
updating the parameters of the classification model according to the loss values;
and calculating the category predicted value of the target category from the standard image through the main task network of the updated classification model.
Optionally, after the target factor with the tag value being the preset tag threshold is selected as the standard impact factor, the method further includes:
acquiring a push queue task, wherein the push queue task comprises a preset push sequence;
and pushing the standard influence factors to the user according to the pushing sequence.
In order to solve the above problems, the present invention also provides an apparatus for screening an influence factor of a target, the apparatus comprising:
the class prediction module is used for acquiring a standard image, inputting the standard image into a pre-constructed classification model to obtain a class prediction value with the standard image as a target class;
the quantity calculation module is used for calculating the category predicted value by using a preset quantity statistical algorithm to obtain the quantity of the influence factors to be selected;
the target factor screening module is used for calculating a plurality of preset influence factors by using the classification model to obtain an influence predicted value of each influence factor, and selecting the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence predicted value;
and the standard influence factor selection module is used for calculating the label value of each influence factor in the target factor set and selecting the influence factor with the label value being a preset label threshold value from the target factor set as the standard influence factor.
In order to solve the above problem, the present invention also provides an electronic device, including:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a computer program executable by the at least one processor, the computer program being executable by the at least one processor to enable the at least one processor to perform the method for impact factor screening of a target object as described above.
In order to solve the above problem, the present invention further provides a computer-readable storage medium comprising a storage data area and a storage program area, wherein the storage data area stores created data, and the storage program area stores a computer program; wherein the computer program, when executed by a processor, implements the method for screening an impact factor of a target object as described above.
The method comprises the steps of calculating a category predicted value taking a standard image as a target category through a main task network of a classification model, and calculating the number to be selected of a plurality of preset influence factors according to the category predicted value and a numerical statistic calculation method, so that the number of the influence factors influencing the output of the classification model in input data is determined; the method comprises the steps of utilizing a plurality of subtask networks of a classification model to respectively calculate influence predicted values of a plurality of preset influence factors, selecting a plurality of preset influence factors of a to-be-selected number as a target factor set according to the influence predicted values, utilizing the influence predicted values to screen the plurality of preset influence factors, improving accuracy of the selected influence factors, calculating tag values of all target factors in the target factor set, selecting the target factor of which the tag value is a preset tag threshold value as a standard influence factor, and achieving the purpose of obtaining the influence factors of the classification model. Therefore, the method, the device and the computer readable storage medium for screening the influence factors of the target object can obtain the influence factors in the input data of the deep learning model.
Drawings
FIG. 1 is a schematic flow chart of a method for screening an influence factor of a target object according to an embodiment of the present invention;
FIG. 2 is a block diagram of an apparatus for screening an influence factor of a target object according to an embodiment of the present invention;
fig. 3 is a schematic diagram of an internal structure of an electronic device for implementing a method for screening an influence factor of a target object according to an embodiment of the present invention;
the implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The execution subject of the method for screening the influence factors of the target object provided by the embodiment of the present application includes, but is not limited to, at least one of electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the method for screening the influence factors of the target object may be performed by software or hardware installed in the terminal device or the server device, and the software may be a block chain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, and the like.
The invention provides a method for screening an influence factor of a target object. Fig. 1 is a schematic flow chart of a method for screening an influence factor of a target object according to an embodiment of the present invention. The method may be performed by an apparatus, which may be implemented by software and/or hardware.
In this embodiment, the method for screening an influence factor of a target substance includes:
and S1, acquiring a standard image, and inputting the standard image into a pre-constructed classification model to obtain a classification predicted value with the standard image as a target classification.
In the embodiment of the invention, the standard image is an image with a target object and a target object label. For example, with an image of a fruit (e.g., apple), the object is a fruit (apple) object and the label is the name of the fruit (e.g., apple).
In the embodiment of the invention, the standard image can be acquired from the block chain node for storing the standard image by using the python statement with the data capture function. Due to the high throughput of the block chain to data, the standard image is stored in the block chain, and then the standard image is obtained from the block chain, so that the efficiency of obtaining the standard image can be improved.
In the embodiment of the invention, the pre-constructed classification model is a convolutional neural network with an image classification function, and the convolutional neural network comprises a convolutional layer, a pooling layer and a full-connection layer.
The convolution layer is used for carrying out convolution processing on the image, firstly locally perceiving each feature in the image, and then carrying out comprehensive operation on the local feature at a higher level so as to obtain global information;
the pooling layer is used for pooling the images after convolution for feature dimension reduction, so that the quantity of data and parameters can be reduced, and the fault tolerance of the model can be improved;
and the full connection layer is used for linear classification, particularly for performing linear combination on the extracted high-level feature vectors and outputting a final image classification result.
Preferably, in the embodiment of the present invention, the classification model includes a main task network and a plurality of subtask networks. The main task network is used for calculating a category predicted value of the standard image as a target category; and the subtask network is used for calculating the influence predicted value of the influence factor of the classification model.
In an embodiment of the present invention, the classifying model is a convolutional neural network, and the inputting the standard image into a pre-constructed classifying model to obtain a class prediction value with the standard image as a target class includes:
performing convolution on the standard image by using a main task network of the convolutional neural network to obtain a convolutional image;
pooling the convolution images to obtain characteristic images;
carrying out full-connection processing on the characteristic image to obtain a full-connection image;
and calculating the category predicted value of the fully-connected image as the target category by using an activation function.
In detail, the convolving the standard image with the main task network of the classification model includes multiplying pixel values in the standard image with a preset convolution kernel matrix. The activation function includes but is not limited to softmax activation function, sigmoid activation function.
Further, after the standard image is input into a pre-constructed classification model to obtain a category prediction value with the standard image as a target category, the method further includes:
calculating a loss value between the category prediction value and the target category;
updating the parameters of the classification model according to the loss values;
and calculating the category predicted value of the target category from the standard image through the main task network of the updated classification model.
In detail, in the embodiment of the present invention, a loss function is used to calculate a loss value between the category prediction value and the target category, where the loss function includes a cross entropy loss function, a mean square error loss function, and the like.
According to the embodiment of the invention, parameters of the classification model are updated according to the loss values by using an Adam optimization algorithm, and the Adam optimization algorithm can adaptively adjust the learning rate in the training process of the classification model, so that the classification detection model is more accurate, and the accuracy of the influence factors is further improved.
And S2, calculating the category predicted value by using a preset number statistical algorithm to obtain the number of the influence factors to be selected.
In an embodiment of the present invention, the preset plurality of influence factors are a plurality of influence factors that may have an influence on a classification result of the classification model. For example, when the classification model classifies an image containing fruits, the shape, color, size and other factors of the fruits in the image may affect the classification result of the classification model.
The calculating the category predicted value by using a preset number statistical algorithm to obtain the number of the influence factors to be selected comprises the following steps:
calculating the category predicted value by using the following quantity statistical algorithm to obtain the quantity of the influence factors to be selected:
w＝ceil(m×sk)
wherein ceil is integer arithmetic, w is the number to be selected, m is-the number of a plurality of preset influence factors, skAnd the standard image is the category predicted value of the target category k.
According to the embodiment of the invention, the quantity to be selected for selecting the influence factors from the preset plurality of influence factors is calculated by using a quantity statistical algorithm, so that the quantity of the influence factors which can influence the classification result of the classification model in the preset plurality of influence factors is determined, and the subsequent screening of redundant influence factors which cannot influence the classification result of the classification model is avoided.
S3, calculating a plurality of preset influence factors by using the classification model to obtain an influence predicted value of each influence factor, and selecting the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence predicted value.
In the embodiment of the present invention, the calculating, by using the classification model, a plurality of preset influence factors to obtain an influence prediction value of each influence factor includes:
performing convolution on the preset multiple influence factors by utilizing the multiple subtask networks of the classification model to obtain convolution factors;
pooling the convolution factors to obtain characteristic factors;
carrying out full connection processing on the characteristic factors to obtain full connection factors;
and calculating to obtain the influence predicted value of the full connection factor.
In detail, the classification model includes a plurality of subtask networks, and each subtask network is used for calculating an influence prediction value of an influence factor corresponding to a prediction value of a different classification in a classification result output by the main task network.
For example, the classification model includes a first subtask network and a second subtask network, the classification result output by the main task network of the classification model includes a pear and an apple, the first subtask network is used for calculating the influence prediction value of the influence factor of the pear as the classification result output by the main task network, and the second subtask network is used for calculating the influence prediction value of the influence factor of the apple as the classification result output by the main task network.
For example, the classification result output by the main task network of the classification model is that the preset multiple influence factors of the pears include: impact factor a, impact factor B, and impact factor C. The first subtask network is used for respectively calculating an influence predicted value of the influence factor A, an influence predicted value of the influence factor B and an influence predicted value of the influence factor C.
In this embodiment of the present invention, the selecting the number of impact factors from the preset multiple impact factors as a target factor set according to the impact prediction value includes:
sequencing the plurality of influence factors according to the sequence of the influence predicted values from large to small to obtain an influence factor sequence;
and selecting the number of influence factors from the influence factor sequence as a target factor set according to the sequence from front to back from large to small.
For example, the number of the impact factors to be selected is 2; presetting a plurality of influence factors including an influence factor A, an influence factor B and an influence factor C; wherein the influence predicted value of the influence factor A is 50, the influence predicted value of the influence factor B is 30, and the influence predicted value of the influence factor C is 90; then arranging a plurality of preset influence factors in the order of the influence predicted values from large to small: an influencing factor C, an influencing factor A and an influencing factor B; and selecting an influence factor C and an influence factor A as a target factor set.
According to the embodiment of the invention, the multiple influence factors to be selected are selected as the target factor set according to the influence predicted values of the preset multiple influence factors, so that the influence factors with lower influence predicted values in the preset multiple influence factors are removed, and the accuracy of selecting the influence factors is improved.
S4, calculating the label value of each influence factor in the target factor set, and selecting the influence factor with the label value being a preset label threshold value from the target factor set as a standard influence factor.
In this embodiment of the present invention, the calculating a label value of each influence factor in the target factor set includes:
calculating a label value of each impact factor in the target factor set using the following label value algorithm:
wherein A isiIs the label value of the target factor, i is the label of the target factor, qiIs a preset standard label.
Because each target factor in the target factor set may affect the classification result of the classification model, the embodiment of the present invention calculates the label value of each target factor in the target factor set, and selects the target factor with the label value as the preset label threshold as the standard impact factor.
Preferably, the preset tag threshold is 1.
For example, the target factor set includes an influence factor a and an influence factor C, where the label of the influence factor a is a color and the label of the influence factor C is a size; the label value of the influence factor A is 1, the label value of the influence factor C is 0, and the preset label threshold value is 1. Therefore, the influence factor A is selected as the standard influence factor according to the label value.
Further, after the target factor with the label value as the preset label threshold is selected as the standard influence factor, the method further includes:
acquiring a push queue task, wherein the push queue task comprises a preset push sequence;
and pushing the standard influence factors to the user according to the pushing sequence.
In detail, the push queue task may be uploaded by a user.
In detail, the pushing queue task can be uploaded by a user, and pushing the standard influence factor to the user according to the pushing sequence comprises pushing the standard influence factor to the user in a visual chart form according to the pushing sequence.
The method comprises the steps of calculating a category predicted value taking a standard image as a target category through a main task network of a classification model, and calculating the number to be selected of a plurality of preset influence factors according to the category predicted value and a numerical statistic calculation method, so that the number of the influence factors influencing the output of the classification model in input data is determined; the method comprises the steps of utilizing a plurality of subtask networks of a classification model to respectively calculate influence predicted values of a plurality of preset influence factors, selecting a plurality of preset influence factors of a to-be-selected number as a target factor set according to the influence predicted values, utilizing the influence predicted values to screen the plurality of preset influence factors, improving accuracy of the selected influence factors, calculating tag values of all target factors in the target factor set, selecting the target factor of which the tag value is a preset tag threshold value as a standard influence factor, and achieving the purpose of obtaining the influence factors of the classification model. Therefore, the method for screening the influence factors of the target object can obtain the influence factors in the input data of the deep learning model.
FIG. 2 is a schematic block diagram of an apparatus for screening an influence factor of a target substance of the present invention.
The device 100 for screening an influence factor of a target object of the present invention may be installed in an electronic device. According to the realized function, the device for screening the influence factors of the target object can comprise a category prediction module 101, a quantity calculation module 102, a target factor screening module 103 and a standard influence factor selection module 104. A module according to the present invention, which may also be referred to as a unit, refers to a series of computer program segments that can be executed by a processor of an electronic device and that can perform a fixed function, and that are stored in a memory of the electronic device.
In the present embodiment, the functions regarding the respective modules/units are as follows:
the category prediction module 101 is configured to obtain a standard image, input the standard image into a pre-constructed classification model, and obtain a category prediction value with the standard image as a target category;
the quantity calculation module 102 is configured to calculate the category prediction value by using a preset quantity statistical algorithm to obtain the quantity of the impact factors to be selected;
the target factor screening module 103 is configured to calculate a plurality of preset influence factors by using the classification model to obtain an influence prediction value of each influence factor, and select the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence prediction value;
the standard influence factor selection module 104 is configured to calculate a label value of each influence factor in the target factor set, and select an influence factor with a label value equal to a preset label threshold value from the target factor set as a standard influence factor.
In detail, the specific implementation manner of each module of the target substance influence factor screening device is as follows:
the category prediction module 101 is configured to obtain a standard image, input the standard image into a pre-constructed classification model, and obtain a category prediction value with the standard image as a target category.
In the embodiment of the invention, the standard image is an image with a target object and a target object label. For example, with an image of a fruit (e.g., apple), the object is a fruit (apple) object and the label is the name of the fruit (e.g., apple).
In the embodiment of the invention, the standard image can be acquired from the block chain node for storing the standard image by using the python statement with the data capture function. Due to the high throughput of the block chain to data, the standard image is stored in the block chain, and then the standard image is obtained from the block chain, so that the efficiency of obtaining the standard image can be improved.
In the embodiment of the invention, the pre-constructed classification model is a convolutional neural network with an image classification function, and the convolutional neural network comprises a convolutional layer, a pooling layer and a full-connection layer.
The convolution layer is used for carrying out convolution processing on the image, firstly locally perceiving each feature in the image, and then carrying out comprehensive operation on the local feature at a higher level so as to obtain global information;
the pooling layer is used for pooling the images after convolution for feature dimension reduction, so that the quantity of data and parameters can be reduced, and the fault tolerance of the model can be improved;
and the full connection layer is used for linear classification, particularly for performing linear combination on the extracted high-level feature vectors and outputting a final image classification result.
Preferably, in the embodiment of the present invention, the classification model includes a main task network and a plurality of subtask networks. The main task network is used for calculating a category predicted value of the standard image as a target category; and the subtask network is used for calculating the influence predicted value of the influence factor of the classification model.
In an embodiment of the present invention, the category prediction module 101 is specifically configured to:
acquiring a standard image;
performing convolution on the standard image by using a main task network of the convolutional neural network to obtain a convolutional image;
pooling the convolution images to obtain characteristic images;
carrying out full-connection processing on the characteristic image to obtain a full-connection image;
and calculating the category predicted value of the fully-connected image as the target category by using an activation function.
In detail, the convolving the standard image with the main task network of the classification model includes multiplying pixel values in the standard image with a preset convolution kernel matrix. The activation function includes but is not limited to softmax activation function, sigmoid activation function.
Further, the device for screening influence factors of a target object further comprises an updating module, and the updating module is specifically configured to:
calculating a loss value between the category prediction value and the target category;
updating the parameters of the classification model according to the loss values;
and calculating the category predicted value of the target category from the standard image through the main task network of the updated classification model.
In detail, in the embodiment of the present invention, a loss function is used to calculate a loss value between the category prediction value and the target category, where the loss function includes a cross entropy loss function, a mean square error loss function, and the like.
According to the embodiment of the invention, parameters of the classification model are updated according to the loss values by using an Adam optimization algorithm, and the Adam optimization algorithm can adaptively adjust the learning rate in the training process of the classification model, so that the classification detection model is more accurate, and the accuracy of the influence factors is further improved.
The quantity calculation module 102 is configured to calculate the category prediction value by using a preset quantity statistical algorithm to obtain the quantity of the impact factors to be selected.
In an embodiment of the present invention, the preset plurality of influence factors are a plurality of influence factors that may have an influence on a classification result of the classification model. For example, when the classification model classifies an image containing fruits, the shape, color, size and other factors of the fruits in the image may affect the classification result of the classification model.
The quantity calculation module 102 is specifically configured to:
calculating the category predicted value by using the following quantity statistical algorithm to obtain the quantity of the influence factors to be selected:
w＝ceil(m×sk)
wherein ceil is integer arithmetic, w is the number to be selected, m is-the number of a plurality of preset influence factors, skAnd the standard image is the category predicted value of the target category k.
According to the embodiment of the invention, the quantity to be selected for selecting the influence factors from the preset plurality of influence factors is calculated by using a quantity statistical algorithm, so that the quantity of the influence factors which can influence the classification result of the classification model in the preset plurality of influence factors is determined, and the subsequent screening of redundant influence factors which cannot influence the classification result of the classification model is avoided.
The target factor screening module 103 is configured to calculate a plurality of preset influence factors by using the classification model to obtain an influence prediction value of each influence factor, and select the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence prediction value.
In an embodiment of the present invention, the target factor screening module 103 is specifically configured to:
performing convolution on the preset multiple influence factors by utilizing the multiple subtask networks of the classification model to obtain convolution factors;
pooling the convolution factors to obtain characteristic factors;
carrying out full connection processing on the characteristic factors to obtain full connection factors;
calculating to obtain an influence predicted value of the full connection factor;
sequencing the plurality of influence factors according to the sequence of the influence predicted values from large to small to obtain an influence factor sequence;
and selecting the number of influence factors from the influence factor sequence as a target factor set according to the sequence from front to back from large to small.
In detail, the classification model includes a plurality of subtask networks, and each subtask network is used for calculating an influence prediction value of an influence factor corresponding to a prediction value of a different classification in a classification result output by the main task network.
For example, the classification model includes a first subtask network and a second subtask network, the classification result output by the main task network of the classification model includes a pear and an apple, the first subtask network is used for calculating the influence prediction value of the influence factor of the pear as the classification result output by the main task network, and the second subtask network is used for calculating the influence prediction value of the influence factor of the apple as the classification result output by the main task network.
For example, the classification result output by the main task network of the classification model is that the preset multiple influence factors of the pears include: impact factor a, impact factor B, and impact factor C. The first subtask network is used for respectively calculating an influence predicted value of the influence factor A, an influence predicted value of the influence factor B and an influence predicted value of the influence factor C.
For example, the number of the impact factors to be selected is 2; presetting a plurality of influence factors including an influence factor A, an influence factor B and an influence factor C; wherein the influence predicted value of the influence factor A is 50, the influence predicted value of the influence factor B is 30, and the influence predicted value of the influence factor C is 90; then arranging a plurality of preset influence factors in the order of the influence predicted values from large to small: an influencing factor C, an influencing factor A and an influencing factor B; and selecting an influence factor C and an influence factor A as a target factor set.
According to the embodiment of the invention, the multiple influence factors to be selected are selected as the target factor set according to the influence predicted values of the preset multiple influence factors, so that the influence factors with lower influence predicted values in the preset multiple influence factors are removed, and the accuracy of selecting the influence factors is improved.
The standard influence factor selection module 104 is configured to calculate a label value of each influence factor in the target factor set, and select an influence factor with a label value equal to a preset label threshold value from the target factor set as a standard influence factor.
In the embodiment of the present invention, the standard impact factor selecting module 104 is specifically configured to:
calculating a label value of each impact factor in the target factor set using the following label value algorithm:
wherein A isiIs the label value of the target factor, i is the label of the target factor, qiIs a preset standard label.
Because each target factor in the target factor set may affect the classification result of the classification model, the embodiment of the present invention calculates the label value of each target factor in the target factor set, and selects the target factor with the label value as the preset label threshold as the standard impact factor.
Preferably, the preset tag threshold is 1.
For example, the target factor set includes an influence factor a and an influence factor C, where the label of the influence factor a is a color and the label of the influence factor C is a size; the label value of the influence factor A is 1, the label value of the influence factor C is 0, and the preset label threshold value is 1. Therefore, the influence factor A is selected as the standard influence factor according to the label value.
Further, after the target factor with the label value as the preset label threshold is selected as the standard influence factor, the method further includes:
acquiring a push queue task, wherein the push queue task comprises a preset push sequence;
and pushing the standard influence factors to the user according to the pushing sequence.
In detail, the push queue task may be uploaded by a user.
In detail, the pushing queue task can be uploaded by a user, and pushing the standard influence factor to the user according to the pushing sequence comprises pushing the standard influence factor to the user in a visual chart form according to the pushing sequence.
The method comprises the steps of calculating a category predicted value taking a standard image as a target category through a main task network of a classification model, and calculating the number to be selected of a plurality of preset influence factors according to the category predicted value and a numerical statistic calculation method, so that the number of the influence factors influencing the output of the classification model in input data is determined; the method comprises the steps of utilizing a plurality of subtask networks of a classification model to respectively calculate influence predicted values of a plurality of preset influence factors, selecting a plurality of preset influence factors of a to-be-selected number as a target factor set according to the influence predicted values, utilizing the influence predicted values to screen the plurality of preset influence factors, improving accuracy of the selected influence factors, calculating tag values of all target factors in the target factor set, selecting the target factor of which the tag value is a preset tag threshold value as a standard influence factor, and achieving the purpose of obtaining the influence factors of the classification model. Therefore, the influence factor screening device for the target object can obtain the influence factors in the input data of the deep learning model.
Fig. 3 is a schematic structural diagram of an electronic device for implementing the method for screening an influence factor of a target object according to the present invention.
The electronic device 1 may include a processor 10, a memory 11 and a bus, and may further include a computer program stored in the memory 11 and executable on the processor 10, such as an influence factor screening program 12 for an object.
The memory 11 includes at least one type of readable storage medium, which includes flash memory, removable hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, magnetic disk, optical disk, etc. The memory 11 may in some embodiments be an internal storage unit of the electronic device 1, such as a removable hard disk of the electronic device 1. The memory 11 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in mobile hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), and the like, which are provided on the electronic device 1. Further, the memory 11 may also include both an internal storage unit and an external storage device of the electronic device 1. The memory 11 may be used to store not only application software installed in the electronic device 1 and various types of data, such as codes of the influence factor filter 12 of the object, but also temporarily store data that has been output or is to be output.
The processor 10 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips. The processor 10 is a Control Unit (Control Unit) of the electronic device, connects various components of the electronic device by using various interfaces and lines, and executes various functions and processes data of the electronic device 1 by running or executing programs or modules (for example, an influence factor screening program for executing a target object, etc.) stored in the memory 11 and calling data stored in the memory 11.
The bus may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is arranged to enable connection communication between the memory 11 and at least one processor 10 or the like.
Fig. 3 shows only an electronic device with components, and it will be understood by those skilled in the art that the structure shown in fig. 3 does not constitute a limitation of the electronic device 1, and may comprise fewer or more components than those shown, or some components may be combined, or a different arrangement of components.
For example, although not shown, the electronic device 1 may further include a power supply (such as a battery) for supplying power to each component, and preferably, the power supply may be logically connected to the at least one processor 10 through a power management device, so as to implement functions of charge management, discharge management, power consumption management, and the like through the power management device. The power supply may also include any component of one or more dc or ac power sources, recharging devices, power failure detection circuitry, power converters or inverters, power status indicators, and the like. The electronic device 1 may further include various sensors, a bluetooth module, a Wi-Fi module, and the like, which are not described herein again.
Further, the electronic device 1 may further include a network interface, and optionally, the network interface may include a wired interface and/or a wireless interface (such as a WI-FI interface, a bluetooth interface, etc.), which are generally used for establishing a communication connection between the electronic device 1 and other electronic devices.
Optionally, the electronic device 1 may further comprise a user interface, which may be a Display (Display), an input unit (such as a Keyboard), and optionally a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, or the like. The display, which may also be referred to as a display screen or display unit, is suitable for displaying information processed in the electronic device 1 and for displaying a visualized user interface, among other things.
It is to be understood that the described embodiments are for purposes of illustration only and that the scope of the appended claims is not limited to such structures.
The influence factor filter 12 of the target object stored in the memory 11 of the electronic device 1 is a combination of a plurality of computer programs, and when running in the processor 10, can realize:
acquiring a standard image, and inputting the standard image into a pre-constructed classification model to obtain a class prediction value of the standard image as a target class;
calculating the category predicted value by using a preset number statistical algorithm to obtain the number of the influence factors to be selected;
calculating a plurality of preset influence factors by using the classification model to obtain an influence predicted value of each influence factor, and selecting the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence predicted value;
and calculating the label value of each influence factor in the target factor set, and selecting the influence factor with the label value as a preset label threshold value from the target factor set as a standard influence factor.
Further, the integrated modules/units of the electronic device 1, if implemented in the form of software functional units and sold or used as separate products, may be stored in a computer readable storage medium. The computer-readable medium may include: any entity or device capable of carrying said computer program code, recording medium, U-disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM).
Further, the computer usable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus, device and method can be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
In addition, functional modules in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof.
The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any accompanying claims should not be construed as limiting the claim concerned.
The block chain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the system claims may also be implemented by one unit or means in software or hardware. The terms second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (10)
1. A method for screening an influence factor of a target, comprising:
acquiring a standard image, and inputting the standard image into a pre-constructed classification model to obtain a class prediction value of the standard image as a target class;
calculating the category predicted value by using a preset number statistical algorithm to obtain the number of the influence factors to be selected;
calculating a plurality of preset influence factors by using the classification model to obtain an influence predicted value of each influence factor, and selecting the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence predicted value;
and calculating the label value of each influence factor in the target factor set, and selecting the influence factor with the label value as a preset label threshold value from the target factor set as a standard influence factor.
2. The method for screening the influence factors of the target object according to claim 1, wherein the classification model is a convolutional neural network, and the inputting the standard image into the pre-constructed classification model to obtain the predicted class value of the standard image as the target class includes:
performing convolution on the standard image by using a main task network of the convolutional neural network to obtain a convolutional image;
pooling the convolution images to obtain characteristic images;
carrying out full-connection processing on the characteristic image to obtain a full-connection image;
and calculating the category predicted value of the fully-connected image as the target category by using an activation function.
3. The method for screening influence factors of a target object according to claim 1, wherein the selecting the number of influence factors from the preset plurality of influence factors as a target factor set according to the influence prediction value comprises:
sequencing the plurality of influence factors according to the sequence of the influence predicted values from large to small to obtain an influence factor sequence;
and selecting the number of influence factors from the influence factor sequence as a target factor set according to the sequence from front to back from large to small.
4. The method for screening the influence factors of the target object according to claim 1, wherein the step of calculating the category predicted value by using a preset number statistical algorithm to obtain the number of the influence factors to be selected comprises the steps of:
calculating the category predicted value by using the following quantity statistical algorithm to obtain the quantity of the influence factors to be selected:
w＝ceil(m×sk)
wherein ceil is integer arithmetic, w is the number to be selected, m is-the number of a plurality of preset influence factors, skAnd the standard image is the category predicted value of the target category k.
5. The method for screening influence factors of a target substance according to claim 1, wherein the calculating the label value of each influence factor in the target factor set comprises:
calculating a label value of each impact factor in the target factor set using the following label value algorithm:
wherein A isiIs the label value of the target factor, i is the label of the target factor, qiIs a preset standard label.
6. The method for screening the influence factors of the target object according to any one of claims 1 to 5, wherein after the standard image is input into a pre-constructed classification model to obtain the predicted value of the class of the standard image as the target class, the method further comprises:
calculating a loss value between the category prediction value and the target category;
updating the parameters of the classification model according to the loss values;
and calculating the category predicted value of the target category from the standard image through the main task network of the updated classification model.
7. The method for screening influence factors of target substances according to claim 1, wherein after the target factors with the tag values being the preset tag threshold values are selected as standard influence factors, the method further comprises:
acquiring a push queue task, wherein the push queue task comprises a preset push sequence;
and pushing the standard influence factors to the user according to the pushing sequence.
8. An apparatus for screening an influence factor of a target, comprising:
the class prediction module is used for acquiring a standard image, inputting the standard image into a pre-constructed classification model to obtain a class prediction value with the standard image as a target class;
the quantity calculation module is used for calculating the category predicted value by using a preset quantity statistical algorithm to obtain the quantity of the influence factors to be selected;
the target factor screening module is used for calculating a plurality of preset influence factors by using the classification model to obtain an influence predicted value of each influence factor, and selecting the number of influence factors from the plurality of preset influence factors as a target factor set according to the influence predicted value;
and the standard influence factor selection module is used for calculating the label value of each influence factor in the target factor set and selecting the influence factor with the label value being a preset label threshold value from the target factor set as the standard influence factor.
9. An electronic device, characterized in that the electronic device comprises:
at least one processor; and the number of the first and second groups,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of impact factor screening for a target object according to any one of claims 1 to 7.
10. A computer-readable storage medium comprising a storage data area and a storage program area, wherein the storage data area stores created data, and the storage program area stores a computer program; wherein the computer program, when executed by a processor, implements a method for impact factor screening of an object according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011097292.7A CN112101481B (en) | 2020-10-14 | 2020-10-14 | Method, device, equipment and storage medium for screening influence factors of target object |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011097292.7A CN112101481B (en) | 2020-10-14 | 2020-10-14 | Method, device, equipment and storage medium for screening influence factors of target object |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112101481A true CN112101481A (en) | 2020-12-18 |
| CN112101481B CN112101481B (en) | 2024-03-19 |
Family
ID=73783359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011097292.7A Active CN112101481B (en) | 2020-10-14 | 2020-10-14 | Method, device, equipment and storage medium for screening influence factors of target object |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112101481B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114819590A (en) * | 2022-04-20 | 2022-07-29 | 平安科技（深圳）有限公司 | Intelligent strategy recommendation method, device, equipment and storage medium |
| CN116664016A (en) * | 2023-07-27 | 2023-08-29 | 北京中关村科金技术有限公司 | Screening method and device of ESG (electronic service guide) sub-topics, electronic equipment and readable storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180137338A1 (en) * | 2016-11-16 | 2018-05-17 | The Governing Council Of The University Of Toronto | System and method for classifying and segmenting microscopy images with deep multiple instance learning |
| CN110147878A (en) * | 2018-11-28 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Data processing method, device and equipment |
-
2020
- 2020-10-14 CN CN202011097292.7A patent/CN112101481B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180137338A1 (en) * | 2016-11-16 | 2018-05-17 | The Governing Council Of The University Of Toronto | System and method for classifying and segmenting microscopy images with deep multiple instance learning |
| CN110147878A (en) * | 2018-11-28 | 2019-08-20 | 腾讯科技（深圳）有限公司 | Data processing method, device and equipment |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114819590A (en) * | 2022-04-20 | 2022-07-29 | 平安科技（深圳）有限公司 | Intelligent strategy recommendation method, device, equipment and storage medium |
| CN114819590B (en) * | 2022-04-20 | 2023-07-18 | 平安科技（深圳）有限公司 | Policy intelligent recommendation method, device, equipment and storage medium |
| CN116664016A (en) * | 2023-07-27 | 2023-08-29 | 北京中关村科金技术有限公司 | Screening method and device of ESG (electronic service guide) sub-topics, electronic equipment and readable storage medium |
| CN116664016B (en) * | 2023-07-27 | 2023-09-26 | 北京中关村科金技术有限公司 | Screening method and device of ESG (electronic service guide) sub-topics, electronic equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112101481B (en) | 2024-03-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111814962A (en) | Method and device for acquiring parameters of recognition model, electronic equipment and storage medium | |
| CN112465071A (en) | Image multi-label classification method and device, electronic equipment and medium | |
| CN112380859A (en) | Public opinion information recommendation method and device, electronic equipment and computer storage medium | |
| CN112541745A (en) | User behavior data analysis method and device, electronic equipment and readable storage medium | |
| CN112465060A (en) | Method and device for detecting target object in image, electronic equipment and readable storage medium | |
| CN111914939A (en) | Method, device and equipment for identifying blurred image and computer readable storage medium | |
| CN111783982A (en) | Attack sample acquisition method, device, equipment and medium | |
| CN113298159A (en) | Target detection method and device, electronic equipment and storage medium | |
| CN112101481B (en) | Method, device, equipment and storage medium for screening influence factors of target object | |
| CN112579621A (en) | Data display method and device, electronic equipment and computer storage medium | |
| CN113516417A (en) | Service evaluation method and device based on intelligent modeling, electronic equipment and medium | |
| CN112885423A (en) | Disease label detection method and device, electronic equipment and storage medium | |
| CN111985449A (en) | Rescue scene image identification method, device, equipment and computer medium | |
| CN114491047A (en) | Multi-label text classification method and device, electronic equipment and storage medium | |
| CN113657385B (en) | Data detection method and device of electronic metering device and electronic equipment | |
| CN114708461A (en) | Multi-modal learning model-based classification method, device, equipment and storage medium | |
| CN113869456A (en) | Sampling monitoring method and device, electronic equipment and storage medium | |
| CN113268665A (en) | Information recommendation method, device and equipment based on random forest and storage medium | |
| CN112990374A (en) | Image classification method, device, electronic equipment and medium | |
| CN113487621A (en) | Medical image grading method and device, electronic equipment and readable storage medium | |
| CN113157739A (en) | Cross-modal retrieval method and device, electronic equipment and storage medium | |
| CN113705686B (en) | Image classification method, device, electronic equipment and readable storage medium | |
| CN113822379B (en) | Process process anomaly analysis method and device, electronic equipment and storage medium | |
| CN113706019A (en) | Service capability analysis method, device, equipment and medium based on multidimensional data | |
| CN113343882A (en) | Crowd counting method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |