US-12619895-B2 - Observation data evaluation

US12619895B2US 12619895 B2US12619895 B2US 12619895B2US-12619895-B2

Abstract

Embodiments of the present disclosure relate to methods, systems, and computer program products for observation data evaluation. In a method, a hierarchical relationship between a plurality of observation items is obtained based on a dataset including a plurality of observation samples. Here, an observation sample in the plurality of observation samples includes a group of measurements for the group of observation items, respectively. A plurality of evaluation models for evaluating an observation sample is generated based on the hierarchical relationship according to a predefined group of membership functions and a predefined group of fuzzy operators. An evaluation model is selected for a further evaluation from the plurality of evaluation models based on a plurality of confidence intervals for the plurality of evaluation models. With these embodiments, the evaluation model may be obtained in an easy and more effective way.

Inventors

Jing Li
Jing Mei
Fan Li
Ya Bin Dang

Assignees

INTERNATIONAL BUSINESS MACHINES CORPORATION

Dates

Publication Date: 20260505
Application Date: 20210713

Claims (19)

1 . A computer-implemented method comprising: determining, by one or more processors, a hierarchical relationship between a plurality of observation items based on a dataset including a plurality of observation samples, an observation sample in the plurality of observation samples including a group of measurements for the plurality of observation items, respectively; generating, by the one or more processors, a tree structure based on the plurality of observation items, a leaf node in the tree structure corresponding to an observation item in the plurality of observation items, and a root node in the tree structure corresponding to an evaluation result for the plurality of observation items; trimming, by the one or more processors, the tree structure based on the plurality of observation samples; generating, by the one or more processors, a plurality of evaluation models for evaluating an observation sample based on the hierarchical relationship according to a predefined group of membership functions and a predefined group of fuzzy operators, wherein the hierarchal relationship comprises the tree structure and a trimmed tree structure; selecting, by the one or more processors, an evaluation model for a further evaluation from the plurality of evaluation models based on a plurality of confidence intervals for the plurality of evaluation models; processing, by the one or more processors, a new sample using the evaluation model selected for further evaluation, wherein the new sample is comprised of a new group of measurements for the observation items corresponding to the trimmed tree structure of the hierarchal relationship; and generating, by the one or more processors, an evaluation result, wherein the evaluation result includes a classification according to one of the predefined group of membership functions.
2 . The method of claim 1 , wherein trimming the tree structure comprises: with respect to a subtree in the tree structure, determining, by the one or more processors, a weight between the root node of the subtree and a child node of the root node based on the plurality of observation samples; and removing, by the one or more processors, the child node in response to the weight meeting a predefined trimming rule.
3 . The method of claim 1 , wherein determining the hierarchical relationship further comprises: determining, by the one or more processors, a group of weights between the root node of the trimmed tree structure and leaf nodes in the trimmed tree structure; and representing, by the one or more processors, the hierarchical relationship based on the trimmed tree structure and the group of weights.
4 . The method of claim 3 , wherein generating the plurality of evaluation models comprises: generating the evaluation model in the plurality of evaluation models by: selecting, by the one or more processors, a membership function from the predefined group of membership functions; selecting, by the one or more processors, a fuzzy operator from the predefined group of fuzzy operators; and generating, by the one or more processors, the evaluation model based on the trimmed tree structure and the group of weights according to the selected membership function and the selected fuzzy operator.
5 . The method of claim 1 , wherein selecting the evaluation model comprises: determining, by the one or more processors, the plurality of confidence intervals for the plurality of evaluation models based on an evaluation dataset, respectively, the evaluation dataset including a plurality of evaluation samples and an evaluation sample in the plurality of evaluation samples including a group of measurements for the group of observation items, respectively; generating, by the one or more processors, a plurality of feature matrixes for the plurality of evaluation samples based on the evaluation model, respectively; obtaining, by the one or more processors, a plurality of evaluation results based on the plurality of feature matrixes and the hierarchical relationship, respectively; determining, by the one or more processors, a confidence interval in the plurality of confidence intervals for the evaluation model based on the plurality of evaluation results; and selecting, by the one or more processors, the evaluation model from the plurality of evaluation models based on the plurality of confidence intervals.
6 . The method of claim 5 , wherein determining the plurality of evaluation results, with respect to an evaluation sample in the plurality of evaluation samples, comprises: determining, by the one or more processors, a feature matrix for the evaluation sample based on the selected membership function by obtaining a count of levels that are to be included in the evaluation result for the evaluation model and determining the feature matrix based on the count, the group of measurements included in the evaluation sample, and the selected membership function; and obtaining, by the one or more processors, the evaluation result in the plurality of evaluation results for the evaluation sample based on the feature matrix and a group of weights in the hierarchical relationship.
7 . The method of claim 1 , further comprising: receiving, by the one or more processors, a target observation sample including the group of measurements for the group of observation items; and determining, by the one or more processors, the evaluation result for the target observation sample based on the selected evaluation model.
8 . The method of claim 1 , wherein the new sample is comprised of a group of skin observation measurements, wherein the evaluation result is a state of skin index of a user based on the new sample collected based on the observation measurements of the observation items of the trimmed tree structure, and wherein the state of the user's skin is within one of the predefined group of membership functions.
9 . The method of claim 2 , wherein the weight between the root node of the subtree and the child node of the root node is determined based on an Exploratory Factor Analysis (EFA) technique.
10 . The method of claim 1 , wherein the trimmed tree structure includes a group of weights for the plurality of observation items of the hierarchal relationship, wherein the group of weights are represented by a weight vector determined according to a Principal Component Analysis (PCA) algorithm.
11 . The method of claim 10 , wherein the weight vector is determined based on a number of the plurality of observation items and wherein the dataset is filtered to remove one or more measurements corresponding to one or more removed observation items of the plurality of observation items, wherein the observation items corresponding to the trimmed tree structure are represented in a matrix X.
12 . The method of claim 11 , wherein the matrix X is normalized to a normalized matrix Z, and wherein a correlation coefficient matrix C is determined according to the normalized matrix Z.
13 . The method of claim 1 , further comprising: applying one or more clustering algorithms to the new group of measurements for the observation items to obtain a strain energy storage index, wherein the strain energy storage index includes a plurality of levels, wherein each of the plurality of levels corresponds to one of the classifications further corresponding to one of the predefined group of membership functions; and generating the plurality of evaluation models based on the predefined group of membership functions, wherein each of the plurality of evaluation models corresponds to one of the plurality of levels.
14 . A computer-implemented system, comprising a computer processor coupled to a computer-readable memory unit, the memory unit comprising instructions that when executed by the computer processor implements a method comprising: determining a hierarchical relationship between a plurality of observation items based on a dataset including a plurality of observation samples, an observation sample in the plurality of observation samples including a group of measurements for the plurality of observation items, respectively; generating a tree structure based on the plurality of observation items, a leaf node in the tree structure corresponding to an observation item in the plurality of observation items, and a root node in the tree structure corresponding to an evaluation result for the plurality of observation items; trimming the tree structure based on the plurality of observation samples; generating a plurality of evaluation models for evaluating an observation sample based on the hierarchical relationship according to a predefined group of membership functions and a predefined group of fuzzy operators, wherein the hierarchal relationship comprises the tree structure and a trimmed tree structure; selecting an evaluation model for a further evaluation from the plurality of evaluation models based on a plurality of confidence intervals for the plurality of evaluation models; processing a new sample using the evaluation model selected for further evaluation, wherein the new sample is comprised of a new group of measurements for the observation items corresponding to the trimmed tree structure of the hierarchal relationship; and generating an evaluation result, wherein the evaluation result includes a classification according to one of the predefined group of membership functions.
15 . The computer-implemented system of claim 14 , wherein selecting the evaluation model comprises: determining the plurality of confidence intervals for the plurality of evaluation models based on an evaluation dataset, respectively, the evaluation dataset including a plurality of samples and a sample in the plurality of samples including the group of measurements for the group of observation items, respectively; generating a plurality of feature matrixes for the plurality of evaluation samples based on the evaluation model, respectively; obtaining a plurality of evaluation results based on the plurality of feature matrixes and the hierarchical relationship, respectively; determining a confidence interval in the plurality of confidence intervals for the evaluation model based on the plurality of evaluation results; and selecting the evaluation model from the plurality of evaluation models based on the plurality of confidence intervals.
16 . The computer-implemented system of claim 14 , wherein the new sample is comprised of a group of skin observation measurements, wherein the evaluation result is a state of skin index of a user based on the new sample collected based on the observation measurements of the observation items of the trimmed tree structure, and wherein the state of the user's skin is within one of the predefined group of membership functions.
17 . A computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by an electronic device to cause the electronic device to perform a method, the method comprising: determining a hierarchical relationship between a plurality of observation items based on a dataset including a plurality of observation samples, an observation sample in the plurality of observation samples including a group of measurements for the plurality of observation items, respectively; generating a tree structure based on the plurality of observation items, a leaf node in the tree structure corresponding to an observation item in the plurality of observation items, and a root node in the tree structure corresponding to an evaluation result for the plurality of observation items; trimming the tree structure based on the plurality of observation samples; generating a plurality of evaluation models for evaluating an observation sample based on the hierarchical relationship according to a predefined group of membership functions and a predefined group of fuzzy operators, wherein the hierarchal relationship comprises the tree structure and a trimmed tree structure; selecting an evaluation model for a further evaluation from the plurality of evaluation models based on a plurality of confidence intervals for the plurality of evaluation models; processing a new sample using the evaluation model selected for further evaluation, wherein the new sample is comprised of a new group of measurements for the observation items corresponding to the trimmed tree structure of the hierarchal relationship; and generating an evaluation result, wherein the evaluation result includes a classification according to one of the predefined group of membership functions.
18 . The computer program product of claim 17 , wherein selecting the evaluation model comprises: determining the plurality of confidence intervals for the plurality of evaluation models based on an evaluation dataset, respectively, the evaluation dataset including a plurality of samples and a sample in the plurality of samples including the group of measurements for the group of observation items, respectively; generating a plurality of feature matrixes for the plurality of evaluation samples based on the evaluation model, respectively; obtaining a plurality of evaluation results based on the plurality of feature matrixes and the hierarchical relationship, respectively; determining a confidence interval in the plurality of confidence intervals for the evaluation model based on the plurality of evaluation results; and selecting the evaluation model from the plurality of evaluation models based on the plurality of confidence intervals.
19 . The computer program product of claim 17 , wherein the new sample is comprised of a group of skin observation measurements, wherein the evaluation result is a state of skin index of a user based on the new sample collected based on the observation measurements of the observation items of the trimmed tree structure, and wherein the state of the user's skin is within one of the predefined group of membership functions.

Description

BACKGROUND The present invention relates to data processing, and more specifically, to methods, systems, and computer program products for evaluating observation data. Nowadays, evaluation systems are provided for evaluating states of various types of objects. For example, in a skincare product industry, skin states of multiple persons should be evaluated as one from four levels (“excellent,” “good,” “moderate,” and “poor”) before product developments. In another example, in a manufacturing industry, various aspects of a product should be observed to determining a quality level of the product. There have been proposed solutions for evaluating various objects. However, these solutions heavily depend on expert knowledge and manual operations. SUMMARY According to one embodiment of the present disclosure, there is provided a computer-implemented method that may be implemented by one or more processors. In the method, one or more processors obtain a hierarchical relationship between a plurality of observation items based on a dataset including a plurality of observation samples. Here, an observation sample in the plurality of observation samples includes a group of measurements for the group of observation items, respectively. One or more processors generate a plurality of evaluation models for evaluating an observation sample based on the hierarchical relationship according to a predefined group of membership functions and a predefined group of fuzzy operators. One or more processors select an evaluation model from the plurality of evaluation models based on a plurality of confidence intervals for the plurality of evaluation models. According to another embodiment of the present disclosure, there is provided a computer-implemented system. The computer-implemented system comprises a computer processor coupled to a computer-readable memory unit, where the memory unit comprises instructions that when executed by the computer processor implements the above method. According to another embodiment of the present disclosure, there is provided a computer program product. The computer program product comprises a computer readable storage medium having program instructions embodied therewith. The program instructions are executable by an electronic device to cause the electronic device to perform actions of the above method. BRIEF DESCRIPTION OF THE DRAWINGS Through the more detailed description of some embodiments of the present disclosure in the accompanying drawings, the above and other objects, features and advantages of the present disclosure will become more apparent, wherein the same reference generally refers to the same components in the embodiments of the present disclosure. FIG. 1 depicts a cloud computing node according to an embodiment of the present disclosure. FIG. 2 depicts a cloud computing environment according to an embodiment of the present disclosure. FIG. 3 depicts abstraction model layers according to an embodiment of the present disclosure. FIG. 4 depicts an example procedure for obtaining an evaluation model according to an embodiment of the present disclosure. FIG. 5 depicts a flowchart of an example method for obtaining an evaluation model according to an embodiment of the present disclosure. FIG. 6A depicts an example tree structure associated with a plurality of observation items according to an embodiment of the present disclosure. FIG. 6B depicts an example subtree structure associated with a portion of the plurality of observation items according to an embodiment of the present disclosure. FIG. 7 depicts an example tree structure after a trimming procedure according to an embodiment of the present disclosure. FIG. 8 depicts an example membership function according to an embodiment of the present disclosure. FIG. 9 depicts an example procedure for determining an evaluation result according to an embodiment of the present disclosure. FIG. 10 depicts an example procedure for determining an evaluation result for a target sample based on a selected evaluation model according to an embodiment of the present disclosure. DETAILED DESCRIPTION Some embodiments will be described in more detail with reference to the accompanying drawings, in which the embodiments of the present disclosure have been illustrated. However, the present disclosure can be implemented in various manners, and thus should not be construed to be limited to the embodiments disclosed herein. It is to be understood that although this disclosure includes a detailed description on cloud computing, implementation of the teachings recited herein are not limited to a cloud computing environment. Rather, embodiments of the present disclosure are capable of being implemented in conjunction with any other type of computing environment now known or later developed. Cloud computing is a model of service delivery for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g. networks, network ba