A data-knowledge multi-level collaborative fusion-driven method and system for assessing children's growth and development

Abstract

A data-knowledge multi-level collaborative fusion-driven method and system for assessing children's growth and development includes: acquiring and preprocessing a multi-source dataset of children's growth and development; constructing a knowledge base representation and a data probability representation based on standard deviation scores for data elements of each data attribute in the multi-source dataset; constructing a quantum-inspired coupled representation of the data elements based on the knowledge base representation and the data probability representation; grouping the quantum-inspired coupled representation by gender, age, and genetic height in multiple levels, and performing multi-level collaborative fusion inspired by the principle of quantum interference; constructing assessment indicators based on the results of the multi-level collaborative fusion; verifying the general and specific patterns of children's growth and development based on the assessment indicators; and for the child to be assessed, comparing the corresponding multi-level collaborative fusion results and assessment indicators to output an individual deviation assessment result.

Description

Technical Field

[0001] This invention relates to the field of intelligent medical assessment technology, and in particular to a method and system for assessing children's growth and development status driven by a multi-level collaborative fusion of data and knowledge under quantum inspiration. Background Technology

[0002] Child growth and development assessment is a fundamental task in pediatric clinical practice and public health. Accurately assessing the growth status of individual children, promptly identifying cases deviating from the norm, and uncovering population growth and development patterns are crucial for early intervention, personalized health management, and clinical decision support. Currently, child growth and development assessment mainly relies on statistical reference tools such as international and domestic child growth standards, as well as the experience and judgment of clinicians. With the development of medical informatization, multi-source child growth and development data (such as height, weight, bone age, genetic height, BMI, etc.) are constantly accumulating, making data-driven automated assessment possible. However, existing assessment technologies still have the following unresolved technical problems:

[0003] First, there is a lack of integration of clinical knowledge. Traditional data-driven assessment methods (such as percentile curve-based assessments) directly compare source data with statistical distributions, failing to incorporate clinically significant standard deviation scores (SDS) into the assessment process. SDS quantifies the degree to which an individual deviates from the normal population and is an important tool for pediatricians to assess growth and development. However, most existing methods treat SDS as a post-processing calculation result rather than as an integral part of the assessment framework, resulting in assessment results lacking a structured representation of clinical knowledge.

[0004] Second, the hierarchical heterogeneity of multi-source data is difficult to represent uniformly. Children's growth and development data exhibit significant hierarchical heterogeneity: firstly, physiological stratification by sex and age; secondly, genetic background stratification based on genetic height (or predicted height); and thirdly, heterogeneous associations among multiple attributes (height, weight, bone age, etc.). Existing fusion methods (such as DS evidence reasoning, maximum likelihood estimation, and deep neural networks) fail to simultaneously consider this multi-layered grouping structure, directly mixing all data for processing, leading to the disruption of consistency within the same group.

[0005] Third, the assessment process lacks interpretable physical mechanisms. While black-box models such as deep neural networks can output assessment results, their internal decision-making processes are difficult for clinicians to understand and verify. Reasoning methods (such as evidence-based reasoning) offer some interpretability, but their parameter settings and conflict coefficient calculations lack intuitive physical meaning related to the inherent distribution characteristics of the data. Although quantum-inspired methods have been attempted for data fusion, current work largely focuses on the definition of quantum ground states and quantum probability calculations, and has not yet integrated the principles of quantum computing with the needs of tiered assessment of children's growth and development.

[0006] Fourth, there is a lack of pattern verification capabilities. Existing assessment techniques primarily serve to determine the growth status of individual children, rarely extracting clinically significant population patterns from the fusion results (e.g., typical distributions of weight SDS and bone age SDS in different genetic height groups). Pattern verification requires high consistency and stability in the fusion results, which current methods struggle to guarantee. Summary of the Invention

[0007] To address the aforementioned issues, this invention proposes a data-knowledge multi-level collaborative fusion-driven method and system for assessing children's growth and development. By using quantum-inspired coupling representation to unify the modeling of the SDS knowledge base and the data probability representation, and by using multi-level collaborative fusion (intra-height group fusion → intra-age group fusion) to obtain highly consistent assessment results, and by verifying general and specific patterns based on the assessment results, it provides interpretable and quantifiable decision support for clinical assessment.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0009] Firstly, a data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development includes the following steps:

[0010] S1. Acquire and preprocess a multi-source dataset of children's growth and development; construct a knowledge base representation and a data probability representation for each data attribute in the multi-source dataset of children's growth and development based on the standard deviation score; construct a quantum-inspired coupled representation for the data elements based on the knowledge base representation and the data probability representation.

[0011] S2. The quantum-inspired coupling representation is grouped into multiple layers according to gender, age and genetic height, and a multi-level collaborative fusion inspired by the principle of quantum interference is performed;

[0012] S3. Based on the results of multi-level collaborative integration, construct evaluation indicators; based on the evaluation indicators, identify the general and specific laws of children's growth and development; for the children to be evaluated, compare their corresponding multi-level collaborative integration results and evaluation indicators, and output the individual deviation evaluation results.

[0013] Preferably, S1 includes:

[0014] Multiplying the data probability representation of a data element by the quantized value of its corresponding knowledge base representation yields the quantum-inspired coupled representation of the data element; the quantum-inspired coupled representation of each attribute vector under each age subset is the superposition of the quantum-inspired coupled representations corresponding to all knowledge base parts therein.

[0015] Preferably, S1 includes:

[0016] The multi-source dataset of children's growth and development is grouped using a three-level grouping dimension including gender, age, and genetic height; genetic height is divided into N continuous intervals; each age subset is further subdivided into up to N height groups based on the genetic height interval.

[0017] S2 includes:

[0018] For the set of quantum-inspired coupled representations of data elements within the same sex, age, and genetic height group, the first level of quantum interference fusion is performed within the genetic height group to obtain the height group fusion result; then, the height group fusion result is performed within the age group to obtain the age group fusion result; wherein, the quantum interference fusion uses the average probability representation part of the set of quantum-inspired coupled representation data elements as the interference baseline, and the deviations between the data probability representation parts of each data element and the average probability representation parts are algebraically superimposed to simulate the quantum interference effect.

[0019] Preferably, in S2, the quantum-inspired coupling representation set of data elements within the same sex, age, and genetic height group performs first-level quantum interference fusion within the genetic height group, including:

[0020] For a subset of data elements that share the same knowledge base representation, the quantum heuristic coupling after interference is represented as follows:

[0021] ;

[0022] in, For data elements The data probability representation part, This represents the average probability. This refers to the quantized value of the corresponding knowledge base representation. The probabilistic representation of a subset of data elements having the same knowledge base representation is the result of interference; the interference result of the attribute values ​​corresponding to this knowledge base representation is:

[0023] ;

[0024] in, The arithmetic mean of the attribute values ​​for this subset; after calculating the attribute value interference results for each of the 7 knowledge base representations, the knowledge base representation with the largest probability representation after interference is selected as the fusion result for the height group:

[0025] .

[0026] Preferably, in S2, the height group fusion results are subjected to a second-level quantum interference fusion within the age group, resulting in age group fusion results including:

[0027] The fusion result of all height groups within the same age subset is regarded as a new set of data elements, and each new data element has a probability representation part. and knowledge base representation part The second-order average probability representation of this set is calculated as the reference baseline for the second-level interference.

[0028] For parts representing the same knowledge base Given a set of height groups, calculate their quantum-inspired coupling representation after interference:

[0029] ;

[0030] Select the largest height group for fusion results As a result of age group integration.

[0031] Preferably, S2 includes:

[0032] When the dataset lacks genetic height information, the height group division is omitted, and single-level quantum interference-inspired fusion is performed directly within the age subset.

[0033] Preferably, S3 includes:

[0034] For each height group fusion result or age group fusion result, the following indicators are defined:

[0035] Most frequently occurring SDS: The SDS value corresponding to the knowledge base representation part with the highest probability in the fusion result, taking the representative value of the interval;

[0036] Probability of the most frequently occurring SDS: This knowledge base represents the output probability of a portion of the corresponding data.

[0037] SDS weighted average: A probabilities-weighted average of each SDS value;

[0038] Attribute value deviation rate: The ratio of attribute value deviation to the average value of the source data;

[0039] General rule verification includes:

[0040] Authentication of the concentrated range of genetic height: Statistically analyze the most frequent SDS probability and SDS weighted average of each genetic height group to identify the height range in which most children are concentrated;

[0041] SDS Normal Range Authentication: Calculate the weighted average SDS for each age group. If the weighted average SDS falls within the preset normal range, the child's overall growth and development is authenticated as normal.

[0042] High-determinism indicator pattern verification: Compare the SDS probability-weighted average of different attributes to identify attribute indicators with higher certainty;

[0043] Specific pattern authentication includes:

[0044] Specific patterns within height groups: For height groups where the probability of most frequent SDS exceeds a preset threshold, record the most frequent SDS value to identify typical patterns of weight SDS and bone age SDS within a specific height group.

[0045] Certification of key age window patterns: Observe the trend of the weighted average SDS of each age group with age, identify the age intervals where SDS increases or decreases significantly, and certify them as rapid growth initiation windows or growth deceleration windows, respectively.

[0046] Window pattern authentication for entering normal state: For weight SDS or bone age SDS indicators, identify the age at which they first enter the preset normal range from a state that is above or below the normal range, and authenticate it as the window period for entering normal growth.

[0047] Preferably, the output of individual deviation assessment results includes:

[0048] The results of age group integration are based on the gender and age of the child to be evaluated;

[0049] Based on the height group fusion results corresponding to the genetic height location of the child to be evaluated;

[0050] Calculate the current height SDS, weight SDS, and bone age SDS values ​​of the child to be evaluated;

[0051] The SDS value of the child to be evaluated is compared with the most frequent SDS and the weighted average of SDS in the corresponding fusion results. If the absolute difference between the individual SDS and the most frequent SDS is greater than the preset threshold, or the individual SDS exceeds the preset confidence interval of the weighted average of SDS, it is judged as a significant deviation.

[0052] The output includes an assessment report showing the degree of deviation, and provides corresponding recommendations based on the growth patterns of certification.

[0053] Preferably, S1 includes:

[0054] Obtain a multi-source dataset of children's growth and development, including height, weight, BMI, bone age, and genetic height attributes, and divide it into multiple age subsets;

[0055] The standard deviation score range is divided into multiple standard intervals, each standard interval corresponding to a knowledge base representation part; each knowledge base representation part uses binary encoding, and different encoding values ​​correspond to different standard intervals.

[0056] In each age subset, the proportion of data elements in each standard deviation score interval to the total number of data elements of that attribute is calculated, and the square root of the proportion is used to obtain the probability representation of the corresponding knowledge base representation. For a single data element, its probability representation is calculated based on the proportion of attribute values ​​in its standard deviation score subset and the overall probability of the standard deviation score subset.

[0057] Secondly, a data-knowledge multi-level collaborative fusion-driven child growth and development assessment system includes:

[0058] The data processing and coupled representation module is used to acquire and preprocess multi-source datasets of children's growth and development. Based on the standard deviation score, it constructs a knowledge base representation part and a data probability representation part for the data elements of each data attribute in the multi-source dataset of children's growth and development. Based on the knowledge base representation part and the data probability representation part, it constructs a quantum-inspired coupled representation of the data elements.

[0059] The multi-level collaborative fusion module is used to group quantum-inspired coupled representations into multiple levels according to gender, age and genetic height, and perform multi-level collaborative fusion inspired by the principle of quantum interference.

[0060] The assessment and certification module is used to construct assessment indicators based on the results of multi-level collaborative fusion; to certify the general and specific patterns of children's growth and development based on the assessment indicators; and to compare the corresponding multi-level collaborative fusion results and assessment indicators for the children to be assessed, and output the individual deviation assessment results.

[0061] The aforementioned data-knowledge multi-level collaborative fusion driven child growth and development assessment system is used to implement the data-knowledge multi-level collaborative fusion driven child growth and development assessment method and its steps as described in the first aspect.

[0062] Compared with the prior art, the beneficial effects of the present invention are reflected in:

[0063] (1) Synergistic effect of knowledge-data unified representation, multi-level quantum interference fusion and pattern authentication: This invention uses quantum-inspired coupling representation to unify the modeling of SDS clinical knowledge base and data probability representation, enabling clinical knowledge to directly participate in fusion calculation; on this basis, a multi-level quantum interference fusion strategy is used to fully adapt to the hierarchical heterogeneous structure of children's growth and development data (gender → age → genetic height), obtaining highly consistent and stable fusion results; then, based on the fusion results, a complete evaluation index system is constructed to realize a complete reasoning chain from group pattern authentication to individual deviation assessment. The synergistic effect of the three fundamentally solves the systemic technical problems in the existing technology, such as the lack of clinical knowledge integration, the difficulty in unifying the hierarchical heterogeneity of multi-source data, the lack of interpretable physical mechanisms in the evaluation process, and the lack of pattern authentication function, thus forming a closed-loop and verifiable intelligent evaluation framework.

[0064] (2) Quantum-inspired unified representation of knowledge and data: This invention constructs SDS clinical knowledge as the knowledge base representation and source data as the data probability representation, achieving unification of the two through coupled representation. This mechanism enables knowledge to directly participate in fusion computation, rather than merely serving as a post-processing rule.

[0065] (3) Multi-level quantum interference fusion strategy: Inspired by the principle of quantum interference, this invention designs a first-level (within height group) and second-level (within age group) collaborative fusion. This strategy makes full use of the hierarchical structure of children's growth and development data (gender → age → genetic height). Each level simulates the interference effect by superimposing the deviation between the probability representation part and the mean value. The fusion process has clear physical intuition.

[0066] (4) Pattern Authentication Function: This invention defines a complete set of evaluation indicators such as MCSDS, PMCSDS, WASDS, and attribute value deviation rate, and realizes the authentication of general patterns (such as the concentrated range of genetic height, the normal range of SDS, and high certainty indicators) and specific patterns (such as the specific pattern of height group, the key age window, and the window to enter the normal state) based on these indicators.

[0067] (5) High interpretability of evaluation results: None of the fusion steps in this invention involve black-box parameter learning; the inputs and outputs of each step have clear mathematical and physical meanings. Clinicians can directly understand the "average probability representation" as the baseline, the "interference result" as the degree of deviation, and the "maximum probability state" as the most typical state of the group. This is in stark contrast to deep learning methods.

[0068] (6) Deep integration of individual deviation assessment and group pattern: This invention not only outputs the group fusion result, but also realizes the quantitative assessment of individual deviation by comparing the individual SDS with the MCSDS and WASDS in the fusion result. More importantly, the reference standards used for deviation assessment (such as key age window and normal SDS interval) are certified from the same fusion result, realizing a consistent logical chain "from group to individual". Attached Figure Description

[0069] Figure 1 This is an overall flowchart of the method in Embodiment 1 of the present invention;

[0070] Figure 2 This is a schematic diagram of multi-level grouping of data elements based on quantum-inspired coupling representation in Embodiment 1 of the present invention (divided into 16 height groups according to gender, age, and genetic height).

[0071] Figure 3 is a schematic diagram of the interference result of the probability representation part based on the quantum interference principle in Embodiment 1 of the present invention, wherein... Figure 3 In the figure, A represents the interference results for all height groups. Figure 3 In the diagram, B represents the interference results of two probabilities within the height group, indicating a subset of groups. Detailed Implementation

[0072] To make the technical means, inventive features, objectives, and effects of the invention readily understandable, the invention is further described below with reference to specific illustrations. However, the invention is not limited to the embodiments described below.

[0073] It should be noted that the structures, proportions, sizes, etc., illustrated in the accompanying drawings of this specification are only used to complement the content disclosed in the specification for those skilled in the art to understand and read, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0074] To address the shortcomings of existing technologies, such as the lack of clinical knowledge integration, the difficulty in unifying the representation of heterogeneous multi-source data, the lack of interpretable physical mechanisms in the assessment process, and the absence of pattern verification functions, this invention provides a data-knowledge fusion-driven method and system for assessing children's growth and development. The core innovation of this invention lies in: constructing SDS clinical knowledge as the knowledge base representation and source data as the data probability representation, with the two forming a unified expression through quantum-inspired coupling; based on this, inspired by the principle of quantum interference, a first interference fusion is performed within the genetic height group, followed by a second interference fusion within the age group, obtaining multi-level, highly consistent assessment results; finally, by defining a series of assessment indicators such as the most frequently occurring SDS (MCSDS), MCSDS probability (PMCSDS), SDS weighted average (WASDS), and attribute value deviation rate, the general and specific patterns of children's growth and development are verified from the fusion results.

[0075] Example 1:

[0076] like Figure 1 The method for assessing children's growth and development, driven by a multi-level collaborative fusion of data and knowledge, includes the following steps:

[0077] Step 1: Quantum-inspired coupled representation of data and knowledge

[0078] Step 1.1: Multi-source data acquisition and preprocessing

[0079] Multi-source datasets for children's growth and development typically include a basic growth dataset and a bone age dataset. The basic growth dataset contains the following attribute vectors: sex, age, height, weight, body mass index (BMI), BMI percentile, and genetic height. The bone age dataset contains the following attribute vectors: height percentile, predicted height, and bone age. Additionally, SDS knowledge sets associated with these attributes are incorporated, including height SDS, weight SDS, and bone age SDS.

[0080] Suppose that dataset DS consists of L attribute vectors Composition, that is Each data unit Data units were generated from the same child at the same age. All data units were divided into age subsets based on age (2-18 years old, with 1.0-year intervals), and the total number of data units was [number missing]. And the first Attribute vectors in age subsets The number of data elements is denoted as Gender (male, female) is treated as a separate grouping dimension throughout all subsequent steps.

[0081] Step 1.2: Define the SDS knowledge base representation part

[0082] SDS values ​​are clinically divided into 7 standard intervals: (-∞,-3], (-3,-2], (-2,-1], (-1,1), [1,2), [2,3), [3,+∞). These 7 intervals correspond to 7 knowledge base representation parts. , , , , , , Each knowledge base representation can be encoded using 3 bits of binary code: , , , , , , (state (The probability is set to zero and ignored). The knowledge base representation does not carry numerical value information, but only indicates the SDS category it belongs to.

[0083] For any data element Based on the range of its associated SDS values, determine its knowledge base representation. For example, if a boy's height SDS value is -1.5, falling within the interval (-2, -1), then the knowledge base representation of his height attribute is: The knowledge base representation of BMI and BMI percentile is jointly defined by the knowledge base representation of height and weight, and the influence of height SDS on BMI (empirical value of 67%) is greater than that of weight SDS on BMI (empirical value of 33%).

[0084] Step 1.3: Define the data probability representation part

[0085] Age Attribute vectors In a subset of the data, statistics have a knowledge base representation. Number of data elements ,satisfy Then the knowledge base representation part The probability representation is as follows: This definition makes This satisfies the normalization requirements. For data elements... Let it belong to the knowledge base representation part. The subset in which it is located, and the subset in which it is located. The number of equal data elements is ,but The data probability representation is as follows:

[0086] (1)

[0087] This definition takes into account both the proportion of attribute values ​​in the subset and the overall probability of the knowledge base representation, so that the data probability representation reflects both the relative magnitude of individual values ​​and the representativeness of the SDS group to which it belongs in the population.

[0088] Step 1.4: Quantum-inspired coupling representation

[0089] Data elements Quantum-inspired coupling representation Its knowledge base representation part quantization value With the data probability representation part The product of and has the physical meaning of . With probability In state:

[0090] (2)

[0091] age Lower attribute vector The quantum-inspired coupling is represented as the superposition of the coupling representations of all knowledge base parts therein:

[0092] (3)

[0093] Data Unit The quantum-inspired coupling is represented as a combination of data elements coupled together, but the subsequent fusion process is handled separately by attribute to ensure the clinical independence of different attributes.

[0094] Step 1.5: Construction of the genetic height group

[0095] To eliminate the interference of genetic background on the assessment results, this invention further introduces genetic height (or predicted height) as a third-level grouping dimension in addition to gender and age grouping. Genetic height (or predicted height) is divided into 16 continuous intervals (unit: cm):

[0096] (-∞,130), [130,135), [135,140), [140,145), [145,150), [150,155), [155,160), [160,165), [165,170), [170,175), [175,180), [180,185), [185,190), [190,195), [195,200), [200,+∞). Each age subset is further subdivided into up to 16 height groups based on genetic height intervals. Data elements within each height group have similar genetic backgrounds, and their SDS distributions are more consistent, providing high-quality input for subsequent quantum interference-inspired fusion.

[0097] Step 2: Multi-level synergistic fusion inspired by the principle of quantum interference

[0098] Inspired by the principle of quantum interference (where the probability amplitudes of quantum states enhance or destructively interfere during superposition), this invention superimposes the deviations between the probability representation and the average probability representation of data within the same group to simulate the interference effect, thereby obtaining highly consistent fusion results. The fusion is divided into two levels: the first level is fusion within height groups; the second level is fusion within age groups.

[0099] Step 2.1: Quantum Interference-Inspired Intragroup Fusion of Height

[0100] For a set of quantum-inspired coupling representation data elements within the same sex, age, and genetic height group { First, calculate the average probability representation of the set. :

[0101] (4)

[0102] The average probability representation serves as the reference baseline for the interferometry. For representations with the same knowledge base... subset { The coupling of its interference results is represented as the algebraic sum of the differences between the data probability representation and the average probability representation of each data element, multiplied by the quantized value of the knowledge base representation:

[0103] (5)

[0104] in This represents the probability representation after interference. This formula simulates the effect of the superposition of in-phase states enhancing each other and the superposition of out-of-phase states canceling each other out in quantum interference: when the probability representations of multiple data elements are all higher than the average value, the interference result is enhanced; when there are both positive and negative values, the interference result is weakened.

[0105] Accordingly, the interference result of the attribute values ​​corresponding to the knowledge base representation part is as follows:

[0106] (6)

[0107] in This is the arithmetic mean of the attribute values ​​of this subset. This definition guarantees that the offset of the interfering attribute values ​​relative to the mean is equal to the algebraic sum of the offsets of each data element. The interferometry results are calculated for each of the seven knowledge base representations. After that, the first ( The coupling representation of the height group fusion results depends on the largest interference state in the probability representation part:

[0108] (7)

[0109] That is, the part representing the maximum probability after selection of the interference. Knowledge base representation This represents the typical height group. Therefore, the corresponding attribute value interference result is:

[0110] (8)

[0111] To visually illustrate the above interference fusion process, Figure 3 A schematic diagram of the interference result based on the probability representation of the quantum interference principle is given. Among them, Figure 3 B in the example illustrates the interference process between two different subsets of knowledge base representations (such as the solid line group and the dashed line group) within the same height group: the probability representation of each data element relative to the average probability representation. The deviations, after algebraic superposition, produce either enhancing or destructive interference effects, resulting in the final probability representation of the interference. This refers to the output probability of the knowledge base representation. Figure 3 A in the figure further provides the probability representation partial distribution of all 16 height groups within this age group after interference. By comparing the maximum probability representation partial distribution of each height group... This allows us to determine the representative state of the age group, where the height group corresponding to the peak is the most typical genetic height range of the age group. This result is consistent with the logic of selecting the highest probability state as the fusion output in formulas (7) and (8).

[0112] Step 2.2: Quantum Interference-Inspired Intra-Age Group Fusion

[0113] The age group fusion takes the height group fusion result as input and employs the same quantum interference principle as the first level. The fusion results of the 16 height groups are treated as 16 "new data elements," each with a probability representation. and knowledge base representation part First, calculate the second-order average probability representation of this age subset:

[0114] (9)

[0115] For parts representing the same knowledge base Given a set of height groups, calculate the coupling representation of their interference results:

[0116] (10)

[0117] The final fusion result for the age group depends on the fusion result of the tallest group. .

[0118] Step 2.3: Alternative Fusion Methods

[0119] When the dataset lacks genetic height (or predicted height) information, the height group division can be omitted, and single-level quantum interference-inspired fusion can be performed directly within the age subset. In this case, the fusion process is the same as in step 2.1, but height grouping is not performed; instead, interference is directly applied to all data elements within the age subset according to the seven knowledge base representations. This method can be considered a simplified implementation of the present invention, but its precision in pattern verification is lower than that of a complete multi-level fusion.

[0120] Step 3: Evaluation of the indicator system and verification of patterns

[0121] Based on the above multi-level collaborative integration results, this invention defines a set of evaluation indicators to quantitatively assess children's growth and development status and verify group patterns.

[0122] Step 3.1: Definition of Evaluation Indicators

[0123] For the fusion results of each age group (or height group), the following indicators are defined:

[0124] (1) Most frequently occurring SDS (MCSDS): The SDS value corresponding to the knowledge base representation part with the highest probability in the fusion result, taking the interval representative value (-3, -2, -1, 0, 1, 2, 3).

[0125] (2) Probability of most frequently occurring SDS (PMCSDS): This knowledge base represents the output probability of the corresponding part. (or ).

[0126] (3) SDS weighted average (WASDS): The SDS values ​​are weighted according to their probabilities.

[0127] (4) Attribute value deviation rate: The ratio of attribute value deviation to the average value of source data, expressed as a percentage.

[0128] Step 3.2: General Pattern Verification

[0129] By comparing the above indicators across different height groups and different age groups, the following general patterns can be verified:

[0130] (1) Pattern of concentrated range of genetic height: Statistical analysis of PMCSDS and WASDS of each genetic height group to identify the height range in which most children are concentrated. For example, if the PMCSDS of the boys' genetic height group [155, 185) is generally higher than that of other groups, then this range is identified as the concentrated range of boys' genetic height.

[0131] (2) Normal range of SDS: Calculate the WASDS of each age group. If the value falls within the range of [-1.0, 1.0], the overall normality of the child's growth and development is confirmed.

[0132] (3) High-certainty index pattern: Compare the probability weighted average of SDS for different attributes to identify which attributes (such as genetic height SDS and bone age SDS) have higher certainty.

[0133] Step 3.3: Specific Pattern Authentication

[0134] (1) Specific patterns within height groups: For height groups with PMCSDS exceeding 0.5 (i.e., probability greater than 50%), record their MCSDS to verify patterns such as "typical values ​​of weight SDS and bone age SDS in specific height groups".

[0135] (2) Key age window pattern: Observe the trend of WASDS changes with age in each age group and identify the age intervals in which SDS increases or decreases significantly. The increasing interval is identified as the "rapid growth start window" and the decreasing interval is identified as the "growth deceleration or end window".

[0136] (3) Normal state window pattern: For indicators such as body weight SDS and bone age SDS, identify the age at which they first enter the normal state from a high or low state (WASDS∈[-0.5,0.5]) and identify it as the "window period for entering normal growth".

[0137] Step 3.4: Individual Deviation Assessment

[0138] For individual children, input their gender, age, height, weight, bone age, genetic height (or predicted height), and perform a deviation assessment following the procedure below:

[0139] (1) The age group fusion results are based on gender and age.

[0140] (2) The fusion results of the height group corresponding to the genetic height location.

[0141] (3) Calculate the current SDS value of the individual and compare it with the MCSDS and WASDS in the fusion result.

[0142] (4) If the absolute difference between the individual SDS and the MCSDS is greater than 1.0, or if the individual SDS exceeds the 95% confidence interval of the WASDS (fitted by the probability distribution of the fusion result), it is judged as “significant deviation”.

[0143] (5) Output deviation report, including: individual SDS, group MCSDS, group WASDS, degree of deviation (mild / moderate / significant), and recommendations based on certification patterns (such as "in the rapid growth window, it is recommended to strengthen nutrition").

[0144] Example 2:

[0145] This embodiment uses clinically collected pediatric growth and development data, including a baseline growth dataset and a bone age dataset. The baseline growth dataset contains seven attributes: sex, age, height, genetic height, weight, BMI, and BMI percentile. The bone age dataset contains three attributes: height percentile, predicted height, and bone age. Associated SDS knowledge includes height SDS, weight SDS, and bone age SDS. All data has been anonymized. Given the often insufficient data units for children under 3 years old and over 16 years old, the evaluation range is limited to 3.0 to 16.9 years.

[0146] Step 1: Construction of Quantum-Inspired Coupled Representations

[0147] Step 1.1: Multi-level grouping

[0148] The first-level groups are constructed based on gender (male / female) and age (3-16 years old, in 1.0-year intervals). Taking the 3-year-old male group as an example, this group has... Each data unit is further divided into two levels of groups based on its genetic height range. Each age subset contains a maximum of 16 height groups. For example, the genetic height range [175, 180] (unit: cm) constitutes one height group. Figure 2 The grouping structure is shown.

[0149] Step 1.2: Calculation of the knowledge base representation and the data probability representation

[0150] Taking a 3-year-old male with height as an example, we statistically analyze the distribution of height SDS across different intervals within this age subset, calculate the probability representation corresponding to each knowledge base representation, and further calculate the quantum-inspired coupling representation of the height data elements. Similarly, we perform the same calculations for attributes such as weight, bone age, and BMI.

[0151] Step 2: Multi-level Collaborative Integration

[0152] Step 2.1: Intra-group synergistic integration

[0153] Taking a 3-year-old male in the genetic height group [175, 180) as an example, this height group contains K height data elements. First, the average probability representation is calculated, then the interference result of each knowledge base representation is calculated, and the knowledge base representation corresponding to the maximum value of the probability representation is taken as the fusion result of this height group. The above fusion is performed on 16 height groups in sequence to obtain 16 fusion results.

[0154] Step 2.2: Intra-age group synergistic integration

[0155] Taking a 3-year-old male as an example, the fusion results of 16 height groups are used as input to calculate the second-order average probability representation. For a set of height groups with the same knowledge base representation, the probability representation after interference is calculated, and the largest one is taken as the final fusion result for that age group.

[0156] Step 2.3 Alternative Fusion Methods

[0157] If the data lacks genetic height, the height group division can be omitted, and all height data elements within the 3-year-old subset for males can be directly interferentially fused according to the seven knowledge base representations. This simplified method is also effective, but the fusion result cannot distinguish differences between different genetic backgrounds. This embodiment uses a complete multi-level fusion.

[0158] Step 3: Evaluation Indicators and Pattern Verification

[0159] Step 3.1: Calculation of Evaluation Indicators

[0160] Calculate the MCSDS, PMCSDS, WASDS, and attribute value deviation rate for the main genetic height group.

[0161] Step 3.2: General Pattern Verification

[0162] Based on the above indicators, we can verify general patterns such as the concentration range of genetic height, the normal range of SDS, and the patterns of highly deterministic indicators.

[0163] Step 3.3: Specific Pattern Authentication

[0164] Based on the above indicators, we further verified specific patterns such as height-specific patterns, critical age window patterns, patterns of entering the normal state window, and patterns of ending the normal growth window.

[0165] Step 4: Individual Deviation Assessment

[0166] Given an individual's growth and development data (e.g., a boy, age 8.5 years, height 125cm, genetic height 172cm (belonging to the [170,175) group), weight 22kg, bone age 8.0 years), assess the deviation of the individual's height, weight, bone age, etc., and combine the above rules to evaluate the individual's growth and development status.

[0167] The above are merely preferred embodiments of the present invention and are not intended to limit the invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0168] Example 3:

[0169] A data-knowledge multi-level collaborative fusion-driven child growth and development assessment system, comprising:

[0170] The data processing and coupled representation module is used to acquire and preprocess multi-source datasets of children's growth and development. Based on the standard deviation score, it constructs a knowledge base representation part and a data probability representation part for the data elements of each data attribute in the multi-source dataset of children's growth and development. Based on the knowledge base representation part and the data probability representation part, it constructs a quantum-inspired coupled representation of the data elements.

[0171] The multi-level collaborative fusion module is used to group quantum-inspired coupled representations into multiple levels according to gender, age and genetic height, and perform multi-level collaborative fusion inspired by the principle of quantum interference.

[0172] The assessment and certification module is used to construct assessment indicators based on the results of multi-level collaborative integration; to certify the general and specific laws of children's growth and development based on the assessment indicators; and to compare the corresponding multi-level collaborative integration results and assessment indicators for the children to be assessed, and output the individual deviation assessment results.

Claims

1. A data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development, characterized in that, Includes the following steps: S1. Acquire and preprocess a multi-source dataset of children's growth and development; construct a knowledge base representation and a data probability representation for each data attribute in the multi-source dataset of children's growth and development based on the standard deviation score; construct a quantum-inspired coupled representation for the data elements based on the knowledge base representation and the data probability representation. S2. The quantum-inspired coupling representation is grouped into multiple layers according to gender, age and genetic height, and a multi-level collaborative fusion inspired by the principle of quantum interference is performed; S3. Based on the results of multi-level collaborative integration, construct evaluation indicators; based on the evaluation indicators, identify the general and specific laws of children's growth and development; for the children to be evaluated, compare their corresponding multi-level collaborative integration results and evaluation indicators, and output the individual deviation evaluation results.

2. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 1, characterized in that, S1 includes: Multiply the data probability representation of a data element by the quantized value of its corresponding knowledge base representation to obtain the quantum-inspired coupled representation of the data element. The quantum-inspired coupling representation of each attribute vector under each age subset is the superposition of the quantum-inspired coupling representations corresponding to all knowledge base parts therein.

3. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 1, characterized in that, S1 includes: The multi-source dataset on children's growth and development is grouped using a three-level grouping dimension including sex, age, and genetic height; genetic height is divided into N continuous intervals; each age subset is further subdivided into up to N height groups based on the genetic height interval; S2 includes: For the set of quantum-inspired coupled representations of data elements within the same sex, age, and genetic height group, the first level of quantum interference fusion is performed within the genetic height group to obtain the height group fusion result; then, the height group fusion result is performed within the age group to obtain the age group fusion result; wherein, the quantum interference fusion uses the average probability representation part of the set of quantum-inspired coupled representation data elements as the interference baseline, and the deviations between the data probability representation parts of each data element and the average probability representation parts are algebraically superimposed to simulate the quantum interference effect.

4. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 3, characterized in that, In S2, the quantum-inspired coupling representation set of data elements within the same sex, age, and genetic height group performs first-level quantum interference fusion within the genetic height group, including: For a subset of data elements that share the same knowledge base representation, the quantum heuristic coupling after interference is represented as follows: ； in, For data elements The data probability representation part, This represents the average probability. This refers to the quantized value of the corresponding knowledge base representation. The probabilistic representation of a subset of data elements having the same knowledge base representation is the result of interference; the interference result of the attribute values ​​corresponding to this knowledge base representation is: ； in, The arithmetic mean of the attribute values ​​for this subset; after calculating the attribute value interference results for each of the 7 knowledge base representations, the knowledge base representation with the largest probability representation after interference is selected as the fusion result for the height group: 。 5. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 4, characterized in that, In S2, the height group fusion results are subjected to a second-level quantum interference fusion within the age group, resulting in the following age group fusion results: The fusion result of all height groups within the same age subset is regarded as a new set of data elements, and each new data element has a probability representation part. and knowledge base representation part The second-order average probability representation of this set is calculated as the reference baseline for the second-level interference. For parts representing the same knowledge base Given a set of height groups, calculate their quantum-inspired coupling representation after interference: ； Select the largest height group for fusion results As a result of age group integration.

6. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 3, characterized in that, S2 include: When the dataset lacks genetic height information, the height group division is omitted, and single-level quantum interference-inspired fusion is performed directly within the age subset.

7. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 1, characterized in that, S3 include: For each height group fusion result or age group fusion result, the following indicators are defined: Most frequently occurring SDS: The SDS value corresponding to the knowledge base representation part with the highest probability in the fusion result, taking the representative value of the interval; Probability of the most frequently occurring SDS: This knowledge base represents the output probability of a portion of the corresponding data. SDS weighted average: A probabilities-weighted average of each SDS value; Attribute value deviation rate: The ratio of attribute value deviation to the average value of the source data; General rule verification includes: Authentication of the concentrated range of genetic height: Statistically analyze the most frequent SDS probability and SDS weighted average of each genetic height group to identify the height range in which most children are concentrated; SDS Normal Range Authentication: Calculate the weighted average SDS for each age group. If the weighted average SDS falls within the preset normal range, the child's overall growth and development is authenticated as normal. High-determinism indicator pattern verification: Compare the SDS probability-weighted average of different attributes to identify attribute indicators with higher certainty; Specific pattern authentication includes: Specific patterns within height groups: For height groups where the probability of most frequent SDS exceeds a preset threshold, record the most frequent SDS value to identify typical patterns of weight SDS and bone age SDS within a specific height group. Certification of key age window patterns: Observe the trend of the weighted average SDS of each age group with age, identify the age intervals where SDS increases or decreases significantly, and certify them as rapid growth initiation windows or growth deceleration windows, respectively. Window pattern authentication for entering normal state: For weight SDS or bone age SDS indicators, identify the age at which they first enter the preset normal range from a state that is above or below the normal range, and authenticate it as the window period for entering normal growth.

8. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 7, characterized in that, The output of individual deviation assessment results includes: The results of age group integration are based on the gender and age of the child to be evaluated; Based on the height group fusion results corresponding to the genetic height location of the child to be evaluated; Calculate the current height SDS, weight SDS, and bone age SDS values ​​of the child to be evaluated; The SDS value of the child to be evaluated is compared with the most frequent SDS and the weighted average of SDS in the corresponding fusion results. If the absolute difference between the individual SDS and the most frequent SDS is greater than the preset threshold, or the individual SDS exceeds the preset confidence interval of the weighted average of SDS, it is judged as a significant deviation. The output includes an assessment report showing the degree of deviation, and provides corresponding recommendations based on the growth patterns of certification.

9. The data-knowledge multi-level collaborative fusion-driven method for assessing children's growth and development according to claim 1, characterized in that, S1 includes: Obtain a multi-source dataset of children's growth and development, including height, weight, BMI, bone age, and genetic height attributes, and divide it into multiple age subsets; The standard deviation score range is divided into multiple standard intervals, each standard interval corresponding to a knowledge base representation part; each knowledge base representation part uses binary encoding, and different encoding values ​​correspond to different standard intervals. In each age subset, the proportion of data elements in each standard deviation score interval to the total number of data elements of that attribute is calculated, and the square root of the proportion is used to obtain the probability representation of the corresponding knowledge base representation. For a single data element, its probability representation is calculated based on the proportion of attribute values ​​in its standard deviation score subset and the overall probability of the standard deviation score subset.

10. A data-knowledge multi-level collaborative fusion-driven child growth and development assessment system, characterized in that, include: The data processing and coupled representation module is used to acquire and preprocess multi-source datasets of children's growth and development. Based on the standard deviation score, it constructs a knowledge base representation part and a data probability representation part for the data elements of each data attribute in the multi-source dataset of children's growth and development. Based on the knowledge base representation part and the data probability representation part, it constructs a quantum-inspired coupled representation of the data elements. The multi-level collaborative fusion module is used to group quantum-inspired coupled representations into multiple levels according to gender, age and genetic height, and perform multi-level collaborative fusion inspired by the principle of quantum interference. The assessment and certification module is used to construct assessment indicators based on the results of multi-level collaborative fusion; to certify the general and specific patterns of children's growth and development based on the assessment indicators; and to compare the corresponding multi-level collaborative fusion results and assessment indicators for the children to be assessed, and output the individual deviation assessment results. The aforementioned data-knowledge multi-level collaborative fusion-driven child growth and development assessment system is used to implement the data-knowledge multi-level collaborative fusion-driven child growth and development assessment method and its steps as described in claim 1.

Images