View source for 17857222. METHOD, ELECTRONIC DEVICE, AND COMPUTER PROGRAM PRODUCT FOR DETECTING MODEL DRIFT simplified abstract (Dell Products L.P.)

=METHOD, ELECTRONIC DEVICE, AND COMPUTER PROGRAM PRODUCT FOR DETECTING MODEL DRIFT=


==Organization Name==

[[:Category:Dell Products L.P.|Dell Products L.P.]]


[[Category:Dell Products L.P.]]

==Inventor(s)==

[[:Category:Jiacheng Ni of Shanghai (CN)|Jiacheng Ni of Shanghai (CN)]][[Category:Jiacheng Ni of Shanghai (CN)]]

[[:Category:Zijia Wang of WeiFang (CN)|Zijia Wang of WeiFang (CN)]][[Category:Zijia Wang of WeiFang (CN)]]

[[:Category:Sanping Li of Beijing (CN)|Sanping Li of Beijing (CN)]][[Category:Sanping Li of Beijing (CN)]]

[[:Category:Zhen Jia of Shanghai (CN)|Zhen Jia of Shanghai (CN)]][[Category:Zhen Jia of Shanghai (CN)]]

==METHOD, ELECTRONIC DEVICE, AND COMPUTER PROGRAM PRODUCT FOR DETECTING MODEL DRIFT - A simplified explanation of the abstract==


This abstract first appeared for US patent application 17857222 titled 'METHOD, ELECTRONIC DEVICE, AND COMPUTER PROGRAM PRODUCT FOR DETECTING MODEL DRIFT

==Simplified Explanation==

The patent application describes a method for detecting model drift in a decision tree model used for image classification, text classification, or data mining. The method involves converting training data into an input vector represented by Shapley values, where each dimension of the vector represents an input feature of the model. The training data is then clustered based on the input vector to obtain data clusters. 

* The method converts training data into an input vector represented by Shapley values.
* The input vector represents the input features of a decision tree model used for image classification, text classification, or data mining.
* The training data is clustered based on the input vector to obtain data clusters.
* The method can detect the drift degree of the decision tree model by comparing a new input vector with the existing data clusters.

Potential applications of this technology:

* Monitoring and detecting model drift in image classification systems.
* Monitoring and detecting model drift in text classification systems.
* Monitoring and detecting model drift in data mining systems.

Problems solved by this technology:

* Model drift can occur over time due to changes in the input data, leading to degraded performance of the decision tree model.
* Detecting model drift manually can be time-consuming and subjective.
* This technology provides an automated method to detect model drift and assess its degree.

Benefits of this technology:

* Improved accuracy and performance of decision tree models by detecting and addressing model drift.
* Time and cost savings by automating the detection of model drift.
* Objective and consistent assessment of model drift degree.


==Original Abstract Submitted==

In a method for detecting a model drift provided in an illustrative embodiment of the present disclosure, training data in a training data set is converted into an input vector represented by Shapley values. A plurality of dimensions of the input vector indicates a plurality of input features of a decision tree model. The decision tree model has been trained for performing at least one of image classification, text classification, or data mining. The method also includes: clustering, on the basis of the input vector, the training data set, so as to obtain a plurality of data clusters. The method also includes: in response to receiving a first input, converting the first input into a first input vector represented by Shapley values. The method also includes: detecting a drift degree of the decision tree model on the basis of the first input vector and the plurality of data clusters.



[[Category:G06K9/62]]
[[Category:G06N5/00]]