Intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment

Abstract

The application provides a cancer patient intelligent old-age care decision-making method based on dynamic psychological evaluation, and belongs to the technical field of intelligent decision-making, and comprises the following steps: obtaining direct evaluation replies of a patient evaluation block and analyzing psychological words, obtaining corresponding contradictory intentions through single reply and contradictory analysis of the evaluation block level; then, global contradictory analysis is performed on the evaluation block core word vector to determine a contradictory resolution probability; then, a first resolution direction is determined in combination with a patient behavior and a psychological dependence image, a second resolution direction is determined according to a contradictory intention connection relationship, a modified reply is generated after the resolution direction is fused and the resolution probability is adjusted; finally, a care sub-scheme is generated according to the modified reply, and the comprehensive old-age care scheme is updated and integrated according to an evaluation period, so that the fitting degree and reliability of the care scheme and the actual state of the patient are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent decision-making technology, and in particular to an intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment. Background Technology

[0002] Cancer patients, as a special group requiring elderly care, have their psychological state directly impacting the suitability and effectiveness of elderly care plans. During the psychological assessment phase before providing elderly care for cancer patients, they are prone to providing subjective responses due to various factors. For example, some patients may deliberately conceal their psychological needs and physical discomfort to avoid burdening their children, while others may exaggerate their negative emotions and psychological distress due to excessive anxiety about their health. Current technologies often rely on the direct results of single psychological assessments for elderly care decisions for cancer patients, failing to identify and correct subjective information in the responses or conduct in-depth analysis of the psychological conflicts revealed during the assessment process. This results in elderly care plans that are poorly aligned with the patient's actual psychological and physical condition, leading to unreliable implementation and an inability to meet the personalized elderly care needs of cancer patients.

[0003] Therefore, this invention proposes an intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment. Summary of the Invention

[0004] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment, in order to solve the aforementioned technical problems.

[0005] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment, comprising: Step 1: Distribute psychological assessment files to cancer patients and capture the initial assessment set of the cancer patients for each assessment block in the psychological assessment files, wherein the initial assessment set contains the direct assessment responses of each assessment question in the assessment block; Step 2: Analyze each direct assessment response according to the assessment indicators of the corresponding assessment question to obtain the direct psychological vocabulary of each assessment indicator. Perform a first contradiction analysis on all direct psychological vocabulary under the same direct assessment response to determine the first contradiction intention. At the same time, perform a second contradiction analysis on the vocabulary vector formed by all direct psychological vocabulary under each direct assessment response in the same assessment block to determine the second contradiction intention. Step 3: Perform a third contradiction analysis on all evaluation blocks according to the core word vector of each evaluation block to determine the third contradiction intention, and perform a fourth contradiction analysis on all evaluation blocks after removing the core word vector of the current block to determine the fourth contradiction intention. Based on the third contradiction intention and the fourth contradiction intention, determine the contradiction resolution probability of the corresponding current block. Step 4: Construct the behavioral dependency profile and psychological dependency profile of the cancer patient based on the historical dependency database of the latest cycle, and determine the first direction for resolving the first contradictory intention. At the same time, determine the second direction for resolving the first contradictory intention based on the first connection relationship between the first contradictory intention and the second contradictory intention. Step 5: Based on the second connection relationship between the first contradictory intention and the third contradictory intention, make a first adjustment to the contradiction resolution probability of the evaluation block that matches the first contradictory intention; Step 6: Adjust the first resolution direction according to the second resolution direction and the first adjustment result, and obtain the corrected response of the corresponding direct assessment response. Obtain the elderly care sub-plan of the corresponding assessment block according to the latest psychological vocabulary of all corrected responses under the same assessment block. Step 7: Update the elderly care sub-plan for each assessment block in real time according to the assessment cycle, and obtain the comprehensive elderly care plan for the latest assessment cycle.

[0006] Preferably, a primary contradiction analysis is performed on all direct psychological terms under the same direct assessment response to determine the primary contradiction intention, including: According to the evaluation indicators of the evaluation questions, all direct psychological terms under the corresponding direct evaluation responses are digitized and the first conflict number with a numerical difference greater than the preset difference is mined. At the same time, according to the penetration length of each evaluation indicator corresponding to the same direct evaluation response in the evaluation question, a penetration mining mechanism is added to mine the second conflict number. Extract the direct mental words of the first conflicting number and the second conflicting number, and regard the overlapping words as the first word and the non-overlapping words as the second word; When the number of the first word is 0 or 1, all the second words are clustered using the words corresponding to the first conflict number as the cluster. When the number of the first word is greater than 1, all the second words are clustered using the first word as the cluster. Based on the clustering results and in combination with all the indicators involved in each clustering block, the first contradictory intent is determined.

[0007] Preferably, a second contradiction analysis is performed on the lexical vectors composed of all direct psychological words under each direct assessment response in the same assessment block to determine the second contradiction intention, including: Retrieve the contradiction assessment model that matches the assessment block from the model database; The vocabulary vectors are input into the contradiction assessment model to perform a second contradiction analysis, and the second contradiction intention is output.

[0008] Preferably, the probability of resolving the contradiction corresponding to the current block is determined based on the third and fourth contradictory intentions, including: The third and fourth contradictory intentions are decomposed into sets of conflict dimensions of assessment indicators. Based on the indicator association topology of psychological assessment of cancer patients, the two sets of conflict dimensions are mapped to a standardized conflict feature space to generate the third conflict feature matrix. Fourth Conflict Feature Matrix In this matrix, the row dimension of the conflict feature matrix represents the evaluation indicators, the column dimension represents the psychological conflict manifestations, and the matrix elements represent the conflict quantification values ​​of the corresponding indicators under the corresponding manifestations. right and Perform element-wise intersection operation to obtain the core conflict feature matrix. At the same time, for and Performing element-wise union operation yields the global conflict feature matrix. The intersection operation is to take... and The minimum value of the elements at corresponding positions, the union operation is to take and The maximum value of the element at the corresponding position; calculate The sum of squares of all elements in the sum of squares and The ratio of the sum of squares of all elements in the formula is considered as the first degree of overlap, C1. calculate The sum of squares of all elements in the sum of squares and The ratio of the sum of squares of all elements in the equation is considered as the second degree of conflict overlap, C2. calculate and The difference of the sum of squares of all elements in the set, and The ratio of the sum of squares of all elements in the equation is considered the conflict dispersion C3. Calculate the probability of conflict resolution for the current block. ,in, It is a natural exponential function.

[0009] Preferably, determining the first direction for resolving the first contradictory intention includes: The historical dependency database is sliced ​​according to the assessment cycle, and the co-occurrence relationship between behavioral feature data and psychological feature data within the same assessment cycle is extracted to construct a behavioral-psychological dual-dimensional association map. Using the psychological words corresponding to the first contradictory intention as retrieval nodes, a neighborhood traversal is performed in the association graph to match behavioral dependency features and psychological dependency features that are strongly associated with the contradictory intention. Based on the causal relationship between the behavioral dependency features and the psychological dependency features, features that only have co-occurrence associations and no causal relationship are eliminated to form a behavioral dependency profile and a psychological dependency profile. Based on the behavioral trend characteristics in the behavioral dependency profile and the emotional stability characteristics in the psychological dependency profile, the causes of the first contradictory intention are analyzed, and the first resolution direction matching the individual characteristics of the cancer patient is determined.

[0010] Preferably, based on the first connection relationship between the first contradictory intention and the second contradictory intention, a second direction for resolving the first contradictory intention is determined, including: The first contradictory intent is mapped as a local conflict feature sequence at the single-response level, and the second contradictory intent is mapped as a global conflict feature sequence within the same evaluation block; Dynamic time warping matching is performed on the local conflict feature sequence and the global conflict feature sequence to obtain the first connection relationship representing the intention transmission path and the degree of first conflict superposition. Based on the constraint direction of the global conflict feature sequence on the local conflict feature sequence, a conflict mitigation path is constructed in the reverse direction of the intention transmission path, and a second resolution direction matching the first connection relationship is adaptively generated according to the first conflict superposition degree.

[0011] Preferably, based on the second connection relationship between the first contradictory intent and the third contradictory intent, a first adjustment is made to the contradiction resolution probability of the evaluation block matching the first contradictory intent, including: The first contradictory intention is mapped as a single-problem local conflict feature, and the third contradictory intention is mapped as a global correlation conflict feature between multiple evaluation blocks; The intent transmission consistency is verified between the local conflict features and the global related conflict features, and a second connection relationship is constructed to characterize the conflict transmission direction, the second conflict superposition intensity, and the cross-block influence range. When the direction of conflict propagation is consistent, the resolution probability is amplified in the positive direction according to the superposition intensity of the second conflict. When the direction of conflict propagation is inconsistent, the resolution probability is negatively reduced according to the cross-block influence range.

[0012] Preferably, the first digestion direction is adjusted a second time based on the second digestion direction and the first adjustment result, and a corrected response corresponding to the direct evaluation response is obtained, including: The first and second resolution directions are mapped to the same standardized psychological conflict feature space, and basic resolution vectors of the same dimension are constructed. With the guided resolution vector ,in, and The dimensions of each feature are equal to the feature dimensions of the standardized psychological conflict feature space. The feature dimensions are determined by the product of the number of psychological assessment indicators and the number of corresponding conflict manifestations. All feature dimensions are normalized. The first adjusted probability of conflict resolution is denoted as Pa. Based on the aforementioned basic resolution vector With the guided resolution vector , determine the meaning Figure 1 Consistency coefficients and conflict offset coefficients, combined with Pa, are used to construct a nonlinear adaptive direction fusion model for the basic resolution vector. Implement the second adjustment; Based on the second adjustment results, conflict feature stripping, semantic vector projection, and logical self-consistency reconstruction were performed on the psychological vocabulary of the direct assessment responses to generate corrected responses.

[0013] Compared with the prior art, the beneficial effects of this application are as follows: 1. It achieves multi-dimensional and full-level identification of contradictory intentions in the responses of psychological assessments of cancer patients. Through the analysis of the first to fourth contradictions, it mines psychological contradictions from four levels: micro level of a single response, meso level of an assessment block, global level of the entire assessment, and global level of the current block relative to other blocks. This avoids the omission and misjudgment of contradictions and improves the comprehensiveness and accuracy of contradiction identification. 2. It realizes the quantitative calculation and dynamic adjustment of the probability of conflict resolution. The standardized resolution probability is obtained through the operation of the conflict feature matrix, and the differential adjustment is made in combination with the second connection relationship between the first and third conflict intentions. This changes the status quo of the traditional subjective judgment of the feasibility of resolution and improves the scientificity and dynamism of the resolution probability. 3. It realizes the personalization and collaboration of the direction of conflict resolution. The first direction of resolution, determined based on the patient's behavior and psychological dependence profile, fits the individual characteristics. The second direction of resolution, determined by the first connection relationship, takes into account the global constraints. The adaptive fusion of the two is achieved through a nonlinear adaptive direction fusion model, so that the direction of resolution is both personalized and holistic. 4. It achieves precise correction of assessment responses and dynamic updating of care plans. The corrected responses generated through methods such as conflict feature stripping eliminate non-objective and contradictory information. The elderly care sub-plans and comprehensive plans generated based on the corrected responses can be updated in real time according to the assessment cycle, realizing the dynamic, personalized and scientific nature of care plans. It effectively solves the problems of low fit and poor reliability of care plans with the actual condition of patients in existing technologies.

[0014] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0015] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of an intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment, as described in an embodiment of the present invention. Detailed Implementation

[0017] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0018] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment, such as... Figure 1 As shown, it includes: Step 1: Distribute psychological assessment files to cancer patients and capture the initial assessment set of the cancer patients for each assessment block in the psychological assessment files, wherein the initial assessment set contains the direct assessment responses of each assessment question in the assessment block; Step 2: Analyze each direct assessment response according to the assessment indicators of the corresponding assessment question to obtain the direct psychological vocabulary of each assessment indicator. Perform a first contradiction analysis on all direct psychological vocabulary under the same direct assessment response to determine the first contradiction intention. At the same time, perform a second contradiction analysis on the vocabulary vector formed by all direct psychological vocabulary under each direct assessment response in the same assessment block to determine the second contradiction intention. Step 3: Perform a third contradiction analysis on all evaluation blocks according to the core word vector of each evaluation block to determine the third contradiction intention, and perform a fourth contradiction analysis on all evaluation blocks after removing the core word vector of the current block to determine the fourth contradiction intention. Based on the third contradiction intention and the fourth contradiction intention, determine the contradiction resolution probability of the corresponding current block. Step 4: Construct the behavioral dependency profile and psychological dependency profile of the cancer patient based on the historical dependency database of the latest cycle, and determine the first direction for resolving the first contradictory intention. At the same time, determine the second direction for resolving the first contradictory intention based on the first connection relationship between the first contradictory intention and the second contradictory intention. Step 5: Based on the second connection relationship between the first contradictory intention and the third contradictory intention, make a first adjustment to the contradiction resolution probability of the evaluation block that matches the first contradictory intention; Step 6: Adjust the first resolution direction according to the second resolution direction and the first adjustment result, and obtain the corrected response of the corresponding direct assessment response. Obtain the elderly care sub-plan of the corresponding assessment block according to the latest psychological vocabulary of all corrected responses under the same assessment block. Step 7: Update the elderly care sub-plan for each assessment block in real time according to the assessment cycle, and obtain the comprehensive elderly care plan for the latest assessment cycle.

[0019] In this embodiment, the psychological assessment document refers to a standardized assessment document designed for the psychological state and elderly care needs of cancer patients. It is pre-set and contains multiple assessment blocks. The assessment content covers psychological dimensions related to elderly care for cancer patients, such as emotional state, social willingness, self-care ability, care needs tendency, and health anxiety level. It is distributed in the form of an online questionnaire. For example, the psychological assessment document designed for elderly patients with advanced gastric cancer includes four assessment blocks: emotional state assessment block, care needs assessment block, health anxiety block, and social willingness assessment block. Each assessment block contains 5-8 assessment questions.

[0020] An assessment block refers to an independent assessment unit in a psychological assessment document, divided according to assessment dimensions. Each assessment block corresponds to a core psychological assessment theme and is a concentrated assessment of a certain psychological state or care needs of cancer patients. Different assessment blocks are logically related and together constitute a complete psychological assessment system.

[0021] The initial assessment set refers to the collection of all assessment responses captured by the platform after cancer patients complete the psychological assessment files. It is stored in categories according to assessment blocks. The initial assessment subset corresponding to each assessment block contains the patient responses to all assessment questions in that assessment block.

[0022] Direct assessment responses refer to the direct answers that cancer patients give to individual assessment questions in a psychological assessment document. These responses can take the form of written answers, options, or scores. They are the patient's direct psychological expression of a single assessment question. For example, if the assessment question is "How low have you been feeling lately? Please rate it on a scale of 1 to 5, where 1 is no low mood and 5 is extremely low mood," the patient's answer of "3" is the direct assessment response to that question. Similarly, if the assessment question is "What type of elderly care service do you hope to receive? Please briefly explain," the patient's answer of "I hope to have someone to accompany me, but I don't need daily living assistance" is the direct assessment response to that question.

[0023] Assessment indicators refer to the core dimensions used to break down and analyze individual assessment questions. They are a refinement of the assessment content of the assessment questions. Each assessment question corresponds to at least one assessment indicator, which serves as the basis for extracting direct psychological terms. For example, the assessment indicator for the question "How low have you been feeling lately? Please rate on a scale of 1-5" is: Depressive mood level; the assessment indicator for the question "What type of elderly care services do you hope to receive? Please briefly describe" is: Daily living care needs and emotional companionship needs.

[0024] Direct psychological vocabulary refers to words that directly reflect a patient's psychological state or needs, extracted from direct assessment responses according to assessment indicators. These can be adjectives, nouns, verbs, etc. For example, from the direct assessment response "I hope someone will accompany me, but I don't need daily care," the direct psychological vocabulary "no need" can be extracted according to the assessment indicator "need for daily care," and the direct psychological vocabulary "high need" can be extracted according to the assessment indicator "need for emotional companionship." From the direct assessment response "3 points (degree of low mood)," the direct psychological vocabulary "moderately low mood" can be extracted.

[0025] The primary contradictory intention refers to the core psychological contradictory intention revealed by the patient in their direct assessment response to a single assessment question, as identified through primary contradictory analysis. It reflects the patient's subjective contradictory tendency under the corresponding psychological dimension of the assessment question. For example, after conducting primary contradictory analysis on the response "I can cook for myself, but I completely need someone to take care of my meals," the primary contradictory intention identified is that "the patient has a cognitive contradiction regarding their needs for self-care and care related to diet; they want to demonstrate self-care ability while also hoping to receive comprehensive dietary care."

[0026] Lexical vectors refer to numerical vectors obtained by vectorizing a set of direct mental words using natural language processing techniques. In this embodiment, the Word2Vec model is used to perform lexical vectorization. For example, the two direct mental words "no need" and "high need" are converted into two-dimensional lexical vectors [0.2, 0.8] using the Word2Vec model, where 0.2 corresponds to the vector value of "no need" and 0.8 corresponds to the vector value of "high need"; "self-care" and "complete dependence" are converted into two-dimensional lexical vectors [0.9, 0.1]. It should be noted that... Yes, the Word2Vec model uses the CBOW training architecture and is based on a national corpus of psychological assessments of cancer patients (including common cancer types such as stomach cancer, lung cancer, and breast cancer, covering the elderly care group aged 60-85). Each data point in the corpus is annotated by professional psychologists and includes psychological vocabulary, semantic strength, corresponding cancer type, and psychological state labels. The model is trained based on the above-mentioned annotated samples. The hyperparameters of the model training are set as follows: word vector dimension 100, number of iterations 20, window size 5, minimum word frequency 5, and word vector similarity is calculated using cosine similarity.

[0027] The second contradictory intention refers to the core psychological contradictory intention of the patient under the same assessment theme, as identified through second contradictory analysis. It is the essence of the mesoscopic contradiction among the responses of multiple assessment questions under the same assessment theme, reflecting the patient's overall subjective contradictory tendency under the assessment theme. For example, after performing second contradictory analysis on the 6 word vectors of the "emotional state assessment block", the identified second contradictory intention is: the patient has a contradiction in the overall emotional state, showing both moderate anxiety and the ability to maintain a peaceful state of mind, resulting in inconsistency in emotional expression.

[0028] In this embodiment, the core vocabulary vector is a standardized digital representation of the core psychological conflict at the assessment block level. It is generated from the conflict pairs obtained from the analysis of the first and second conflicts within the assessment block, after calculation of the absolute value of the difference, threshold filtering, and combination operations. Its dimension is consistent with the standardized psychological conflict feature space, specifically including: Complete the first and second contradiction analysis for all direct assessment responses within the assessment block, and extract all opposing psychological vocabulary contradiction pairs that represent the patient's psychological conflict, such as self-care-dependence and low-needs-high-needs. The Word2Vec model is used to convert the two opposing psychological words of each pair of contradictions into numerical vectors of the same dimension as the standardized psychological conflict feature space, and to perform 0~1 normalization processing. For each pair of contradictory words, perform a dimension-wise difference and take the absolute value operation on the two word vectors to obtain the conflict feature vector of that pair, as shown in the formula: Each dimension value in the vector represents the conflict intensity in that dimension; A preset conflict intensity threshold is set, which is adapted to the psychological assessment scenario of cancer patients, such as 0.5. It is obtained based on the normal distribution analysis of the contradiction to the conflict feature vector, and is generally the mean plus the standard deviation. Calculate the average dimension value of the conflict feature vector of each pair of contradictions, and select the contradiction pairs with an average dimension value greater than a preset threshold as the core contradiction pairs of the evaluation block. For all the selected core contradiction pairs, perform a combination operation that takes the maximum value in each dimension on the conflict feature vectors; The result of the calculation is the core vocabulary vector of this assessment block, and the formula is: , where n is the number of core contradiction pairs, ensuring that the vector maximizes the representation of the core conflict features of the evaluation block. If there are no core contradiction pairs in the evaluation block, the core vocabulary vector is labeled with all zeros.

[0029] The third contradictory intention refers to the core intention of the patient's overall psychological contradiction throughout the psychological assessment process, identified through third contradictory analysis. It is an essential distillation of the macroscopic contradictions between different assessment topics, reflecting the patient's overall subjective contradictory tendency. For example, after performing third contradictory analysis on the core vocabulary vectors of the four assessment blocks, the identified third contradictory intention is: the patient has a global contradiction in their overall psychological state, with a significant conflict between their high level of health anxiety and low social willingness and high need for emotional companionship, reflecting the patient's contradictory psychology of both craving external attention and rejecting social interaction.

[0030] The fourth contradictory intention refers to the core psychological contradiction between the current assessment block and all other assessment blocks, identified through fourth contradiction analysis. It is an essential distillation of the relatively global contradiction between the current block and other blocks, and one of the core bases for determining the probability of resolving the contradiction in the current block. For example, after the fourth contradiction analysis using the care needs assessment block as the current block, the identified fourth contradictory intention is: the care needs assessment block reflects the patient's high need for emotional companionship and lack of daily living care needs, which contradicts the low social willingness, high health anxiety, and moderate depression reflected in the other three assessment blocks, demonstrating a mismatch between the patient's care needs and their emotional and social state.

[0031] The probability of conflict resolution refers to the probability that the conflict will be effectively resolved after adopting the corresponding resolution strategy for the psychological conflict of the current assessment block. The value ranges from 0 to 1. The closer the value is to 1, the higher the feasibility of conflict resolution. This probability value is calculated based on the third and fourth conflict intentions and serves as the quantitative basis for subsequent adjustment of the resolution probability and determination of the resolution direction.

[0032] The historical dependency database stores psychological assessment data, behavioral data, and elderly care data for cancer patients across all past assessment cycles. Psychological assessment data includes past assessment responses, psychological vocabulary, and ambivalent intentions; behavioral data includes patients' daily living behaviors, social behaviors, and medical-seeking behaviors; and elderly care data includes past care plans, care implementation effects, and patient feedback. The database is categorized and stored according to assessment cycles, and the latest cycle's data can be retrieved for analysis in real time. For example, the historical dependency database might store a cancer patient's psychological assessment scale responses, daily self-care behavior records, past care plans, and implementation feedback for the past six assessment cycles (30 days each). The database handles abnormal data as follows: missing responses / garbled data are completed using interpolation; extreme responses are handled using... Identify and eliminate based on principles.

[0033] Behavioral dependency profiling refers to the profile of a cancer patient's behavioral characteristics obtained by modeling and refining behavioral data from a historical dependency database. It reflects the patient's behavioral habits and dependency tendencies in daily life, social interactions, and medical care, and serves as one of the bases for determining personalized intervention directions. For example, a patient's behavioral dependency profile constructed based on historical data might be: moderate self-care ability, able to independently wash and dress, but unable to cook independently; minimal social interaction, meeting with children only once a week; requires daily administration of anti-cancer drugs, exhibiting drug dependence and requiring constant reminders.

[0034] Psychological dependency profiling refers to a profile of a cancer patient's psychological characteristics obtained by modeling and refining psychological assessment data from a historical dependency database. It reflects the patient's psychological state and dependency tendencies in areas such as emotional state, health anxiety, care needs, and social willingness, serving as another basis for determining the direction of personalized intervention. For example, a patient's psychological dependency profile constructed based on historical data might show: long-term moderate anxiety, high level of health anxiety, strong psychological dependence on children, a desire for emotional companionship, and high psychological resistance to strangers.

[0035] The first direction for resolving the conflict refers to the core direction determined by combining the patient's behavioral and psychological dependency profiles to resolve the micro-conflict. This direction aligns with the patient's individual behavioral and psychological characteristics and serves as the foundation for conflict resolution. For example, if the first conflict is: the patient has a cognitive conflict regarding their needs for self-care and care related to diet, wanting to demonstrate self-care ability while also desiring comprehensive dietary care, then, based on the patient's behavioral dependency profile (inability to cook independently) and psychological dependency profile (strong psychological dependence on children), the first direction for resolving the conflict is: providing the patient with semi-self-service dietary care, where caregivers prepare the ingredients, the patient performs simple cooking operations, and the patient's children accompany them for meals daily.

[0036] The first connection relationship refers to the correlation between the first contradictory intention and the second contradictory intention, including characteristics such as the transmission path of the contradictory intention and the degree of conflict superposition. It reflects the inherent logical connection between the single-problem contradiction at the micro level and the overall contradiction of the assessment block at the meso level, and is the core basis for determining the second direction of resolution. For example, if the first contradictory intention is: the contradiction in the patient's expression of the degree of depression in eating mood, and the second contradictory intention is: the inconsistency in the expression of the patient's overall emotional state, the first connection relationship between the two is: the micro contradiction in eating mood is a specific manifestation of the meso-level contradiction in the overall emotional state, the contradiction transmission path is from the eating mood dimension to the overall emotional dimension, the degree of conflict superposition is 0.7, and its value ranges from 0 to 1, with higher values ​​indicating stronger superposition.

[0037] The second resolution direction refers to the global constraint direction determined by the first contradiction intention and the first connection relationship to resolve this micro-contradiction. This direction aligns with the overall contradiction characteristics of the assessment block, forming a global constraint and supplement to the first resolution direction. For example, the first resolution direction mentioned above is: semi-self-service meal care + children accompanying meals. Combining the first connection relationship (the emotional conflict related to meals is a specific manifestation of the overall emotional conflict, with a conflict superposition degree of 0.7), the determined second resolution direction is: on the basis of semi-self-service meal care, add an emotional guidance link during the meal process, with professional psychological counselors accompanying the patient during meals to simultaneously alleviate the low mood during meals and the overall emotional anxiety.

[0038] The second connection refers to the relationship between the first and third contradictory intentions, including characteristics such as the direction of transmission of contradictory intentions, the intensity of conflict superposition, and the scope of cross-block influence. It reflects the inherent logical connection between single-problem contradictions at the micro level and global contradictions at the macro level, and is the core basis for the first adjustment of the probability of contradiction resolution. For example, if the first contradictory intention is the cognitive contradiction of the patient's dietary care needs, and the third contradictory intention is the global contradiction between the patient's high health anxiety and low social willingness and high need for emotional companionship, the second connection between the two is: the micro contradiction of dietary care needs is transmitted to the global health anxiety contradiction, the intensity of conflict superposition is 0.8, and the scope of cross-block influence covers the three assessment blocks of care needs, emotional state, and health anxiety.

[0039] The first adjustment refers to the probability of resolving the contradiction in the evaluation block that matches the first contradiction intention. The numerical adjustment is combined with the second connection relationship. The adjustment methods include positive amplification and negative contraction, so that the adjusted probability of resolving the contradiction is more in line with the correlation characteristics between micro contradictions and global contradictions.

[0040] The second adjustment refers to the direction optimization of the first resolution direction, which combines the second resolution direction with the probability of contradiction resolution after the first adjustment. It is a global constraint and quantitative adjustment of the basic resolution direction, so that the final resolution direction after adjustment not only fits the individual characteristics of the patient, but also conforms to the contradiction characteristics of the overall assessment block and the global situation, while taking into account the feasibility of contradiction resolution.

[0041] The revised response refers to the assessment response obtained by modifying and reconstructing contradictory psychological terms in the original direct assessment response, based on the second adjusted final resolution direction. This revised response eliminates non-objective and contradictory information from the original direct assessment response, making it more consistent with the patient's actual psychological and physical state. For example, if the original direct assessment response was: "I can cook for myself, but I completely need someone to take care of my meals," the revised response, based on the second adjusted resolution direction, would be: "I can independently complete simple cooking operations, but I need someone to prepare ingredients and accompany me during meals."

[0042] The latest psychological vocabulary refers to the psychological terms that have been re-analyzed and extracted from the revised responses according to the assessment indicators, eliminating contradictions. It is the core basis for generating sub-plans for elderly care. For example, the latest psychological terms extracted from the revised response: "I can independently complete simple cooking operations, but I need others to prepare ingredients and accompany me during meals" are: self-care in eating - partial ability, dietary care needs - ingredient preparation, and emotional companionship.

[0043] A sub-plan for elderly care refers to a specific elderly care plan generated for a single assessment block based on the latest psychological vocabulary of all corrected responses under that block. Each assessment block corresponds to one sub-plan, which is a component of the comprehensive elderly care plan. For example, the sub-plan generated for the care needs assessment block might include: Dietary care: providing semi-self-service, with caregivers preparing fresh ingredients daily and patients completing simple cooking themselves; Emotional companionship: arranging for children to accompany the patient during meals daily, and professional psychological counselors accompanying the patient during meals three times a week to provide emotional support; Daily living care: requiring no dedicated caregivers, only daily reminders to take medication.

[0044] The assessment cycle refers to the fixed time interval for conducting psychological assessments of cancer patients. It serves as the basis for dynamically updating elderly care plans. In this embodiment, the basic assessment cycle is set at 30 days, which can be dynamically adjusted according to the severity of the patient's condition and the stability of their psychological state. The quantitative adjustment standard is as follows: for patients with advanced cancer / psychological stability <0.5, the assessment cycle is shortened to 15 days; for patients with early-stage cancer / psychological stability ≥0.8, the assessment cycle is extended to 60 days. Psychological stability is obtained by comprehensively quantifying emotional stability characteristics, with a value ranging from 0 to 1.

[0045] The comprehensive elderly care plan refers to the overall elderly care plan for cancer patients obtained by integrating and optimizing the elderly care sub-plans corresponding to all assessment blocks. The plan is updated in real time according to the assessment cycle, and incorporates the latest psychological and behavioral characteristics of patients to achieve dynamic and personalized care.

[0046] The beneficial effects of the above technical solution are as follows: Through four-level contradiction intention analysis, comprehensive identification of contradictions at the micro, meso, and macro levels in the psychological assessment responses of cancer patients is achieved; personalized contradiction resolution directions are determined by combining patient behavior and psychological dependence profiles; and the resolution probability and direction are dynamically adjusted through the connection relationship of contradiction intentions. Finally, a comprehensive elderly care plan that fits the actual condition of the patient is generated and updated in real time. This effectively solves the problem of low fit of care plans caused by non-objective assessment responses of patients in existing technologies, greatly improves the scientificity, personalization, and dynamism of elderly care decisions for cancer patients, and enhances the reliability of care plan implementation.

[0047] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment. It performs first contradiction analysis on all direct psychological terms under the same direct assessment response to determine the first contradiction intention, including: According to the evaluation indicators of the evaluation questions, all direct psychological terms under the corresponding direct evaluation responses are digitized and the first conflict number with a numerical difference greater than the preset difference is mined. At the same time, according to the penetration length of each evaluation indicator corresponding to the same direct evaluation response in the evaluation question, a penetration mining mechanism is added to mine the second conflict number. Extract the direct mental words of the first conflicting number and the second conflicting number, and regard the overlapping words as the first word and the non-overlapping words as the second word; When the number of the first word is 0 or 1, all the second words are clustered using the words corresponding to the first conflict number as the cluster. When the number of the first word is greater than 1, all the second words are clustered using the first word as the cluster. Based on the clustering results and in combination with all the indicators involved in each clustering block, the first contradictory intent is determined.

[0048] In this embodiment, digital conversion refers to converting direct psychological terms into standardized numerical values ​​according to assessment indicators, realizing the conversion of textual terms into calculable numbers. The range of numerical values ​​is set according to the dimensions of the assessment indicators. In this embodiment, a numerical scale of 1-10 is used for conversion. The numerical values ​​are positively correlated with the semantic strength of the psychological terms, which is the basis for subsequent conflict data mining. For example, the mapping relationship between direct psychological terms and numerical values ​​corresponding to the assessment indicator "self-care ability - diet" is: fully self-care - 10, partially self-care - 6, no self-care ability - 1, so the direct psychological term "partial self-care" is converted into the numerical value 6; the mapping relationship corresponding to the assessment indicator "care needs - diet" is: no need - 1, low need - 4, medium need - 7, high need - 10, so the direct psychological term "high need" is converted into the numerical value 10.

[0049] The first conflict number refers to a set of conflicting values ​​among the numerical values ​​corresponding to different assessment indicators under the same direct assessment response after digital conversion, where the numerical difference is greater than a preset difference. The preset difference is set according to the number and dimensions of the assessment indicators. In this embodiment, the preset difference is set to 4. This threshold is based on the statistical analysis of 5,000 psychological assessment samples of cancer patients. When the numerical difference is ≥4, 92% of the samples have significant psychological contradictions, which meets the significance test of P<0.01. The numerical difference is the difference between the absolute values ​​of two numerical values. The first conflict number reflects the explicit contradiction of direct psychological terms at the numerical level. For example, if two numerical values ​​under the same response are 6 (partially self-reliant) and 10 (high need), the numerical difference between them is 4, which is equal to the preset difference, and is determined to be the first conflict number. If the numerical values ​​are 9 (fully self-reliant) and 2 (low need), the numerical difference is 7, which is greater than the preset difference, and is also determined to be the first conflict number.

[0050] Penetration length refers to the length and importance of a single assessment indicator within the semantic description of its corresponding assessment question. It is quantified using a numerical value ranging from 1 to 5. A higher numerical value indicates a longer penetration length and greater importance for the assessment indicator within the assessment question, serving as the core quantitative basis for the penetration mining mechanism. Length is quantified by the proportion of characters, while importance is jointly determined by clinical psychologists and elderly care experts, each accounting for 50% of the weight. For example, in the assessment question "How are your current self-care and care needs? Please describe them mainly in terms of diet and clothing," the penetration length of the assessment indicators "self-care ability - diet" and "care needs - diet" is 5, while the penetration length of the assessment indicators "self-care ability - clothing" and "care needs - clothing" is 3, reflecting the higher importance of diet-related indicators in this question.

[0051] The penetration mining mechanism refers to a conflict data mining mechanism that targets all digitized values ​​under the same direct evaluation response and combines the penetration length of each evaluation indicator. Based on the analysis of explicit data differences, this mechanism considers the importance of evaluation indicators and mines out implicit conflict data caused by differences in the importance of indicators, thus supplementing the first conflict data mining.

[0052] The second conflict number refers to the conflict number mined through the penetration mining mechanism under the same direct assessment response. This number not only considers the difference in numerical values, but also combines the penetration length of the assessment indicator for weighted calculation, reflecting the implicit contradictions of direct psychological words in terms of numerical value and indicator importance. For example, the numerical values ​​after digital conversion are 6 (partial self-care, penetration length 5), 10 (high need, penetration length 5), 8 (complete self-care - clothing, penetration length 3), 3 (low need - clothing, penetration length 3). The difference in diet-related numbers is 4 (first conflict number), and the difference in clothing-related numbers is 5. Combining the penetration length for weighted calculation of conflict value, the conflict value for diet-related numbers is 4 × 5 = 20, and the conflict value for clothing-related numbers is 5 × 3 = 15. If the preset weighted conflict value is 18, then the diet-related numbers 6 and 10 are the second conflict numbers. Here, 18 is the optimal critical value for contradiction identification, with a false negative rate of ≤5%. The weighted conflict value C = |V1-V2| × L, where |V1-V2| is the numerical difference, and L is the penetration length of the assessment indicator.

[0053] Overlapping words refer to direct psychological words that simultaneously correspond to both the first and second conflicting numbers. That is, the word is the textual carrier of both the first and second conflicting numbers, and is the direct psychological word with the most prominent contradiction in the same direct assessment response. For example, if the first conflicting number is 6 (partially self-reliant) and 10 (high need), and the second conflicting number is also 6 (partially self-reliant) and 10 (high need), then the corresponding direct psychological words "partially self-reliant" and "high need" are overlapping words.

[0054] The second word refers to the direct psychological word that corresponds only to the first conflict number or only to the second conflict number, i.e., non-overlapping words. For example, if the first conflict number is 6 (partial self-care), 10 (high need), and 8 (complete self-care - dressing), and the second conflict number is 6 (partial self-care) and 10 (high need), then "complete self-care - dressing" which corresponds only to the first conflict number is the second word.

[0055] Clusters refer to the words that serve as the core cluster centers in clustering. The selection of clusters is determined by the number of words in the first category. Clusters are the benchmark for clustering, and all second categories are clustered around the clusters.

[0056] Clustering refers to the process of grouping the second set of words into clusters using a clustering algorithm. In this embodiment, a hierarchical clustering algorithm is used, which clusters words based on semantic similarity and contradictory relevance, so that each group of words forms an independent contradictory sub-module, which is the basis for determining the first contradictory intent. For example, taking the first words "partially self-reliant" and "high demand" as clusters, hierarchical clustering is performed on the second words "fully self-reliant - dressing" and "low demand - dressing". Based on semantic similarity, "fully self-reliant - dressing" is clustered into the "partially self-reliant" cluster, and "low demand - dressing" is clustered into the "high demand" cluster, forming two clustering processing blocks.

[0057] A clustering block refers to an independent module obtained after clustering, consisting of a cluster and a second term that is clustered to that cluster. For example, the partially self-care + fully self-care - dressing module and the high-need + low-need - dressing module formed after the above clustering are two clustering blocks, which correspond to the assessment indicators of self-care ability (eating and dressing) and care needs (eating and dressing), respectively.

[0058] The beneficial effects of the above technical solution are as follows: quantitative analysis of psychological vocabulary is achieved through digital conversion; by combining explicit first conflict numbers and implicit second conflict numbers, comprehensive identification of contradictory words in the responses to individual assessment questions is achieved; orderly grouping of contradictory words is achieved through case-specific clustering; and finally, the intention of the first contradiction is accurately determined by combining the clustering results with assessment indicators. This effectively improves the accuracy and refinement of the first contradiction analysis, avoids the omission and misjudgment of contradictions at the micro level, and provides accurate micro-contradictory basis for determining the direction of subsequent contradiction resolution.

[0059] This invention provides a method for intelligent elderly care decision-making for cancer patients based on dynamic psychological assessment. It performs a second contradiction analysis on the lexical vector composed of all direct psychological words under each direct assessment response in the same assessment block to determine the second contradiction intent, including: Retrieve the contradiction assessment model that matches the assessment block from the model database; The vocabulary vectors are input into the contradiction assessment model to perform a second contradiction analysis, and the second contradiction intention is output.

[0060] In this embodiment, the model database refers to a database that stores various conflict assessment models for different psychological assessment topics. The models are classified and stored according to the topic of the assessment block, and each assessment block topic corresponds to a unique conflict assessment model.

[0061] The contradiction assessment model is trained on a neural network model using labeled samples under the corresponding topic of the assessment block. Each sample is labeled with word vectors, contradiction type, contradiction intention and contradiction degree. The training sample size for each topic is no less than 1000.

[0062] Model matching refers to the process of searching and retrieving the corresponding contradiction assessment model in the model database based on the theme of the assessment block being analyzed.

[0063] It should be noted that the principles for obtaining the third and fourth contradictory intentions are similar to those for obtaining the second contradictory intention, and will not be elaborated here.

[0064] The beneficial effects of the above technical solution are: by constructing a one-to-one matching system between the evaluation block theme and the contradiction assessment model, the modeling and intelligentization of the second contradiction analysis are realized, which can quickly and accurately identify the meso-level contradictions at the evaluation block level and extract the intention of the second contradiction.

[0065] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment, which determines the probability of conflict resolution for the current block based on the third and fourth contradictory intentions, including: The third and fourth contradictory intentions are decomposed into sets of conflict dimensions of assessment indicators. Based on the indicator association topology of psychological assessment of cancer patients, the two sets of conflict dimensions are mapped to a standardized conflict feature space to generate the third conflict feature matrix. Fourth Conflict Feature Matrix In this matrix, the row dimension of the conflict feature matrix represents the evaluation indicators, the column dimension represents the psychological conflict manifestations, and the matrix elements represent the conflict quantification values ​​of the corresponding indicators under the corresponding manifestations. right and Perform element-wise intersection operation to obtain the core conflict feature matrix. At the same time, for and Performing element-wise union operation yields the global conflict feature matrix. The intersection operation is to take... and The minimum value of the elements at corresponding positions, the union operation is to take and The maximum value of the element at the corresponding position; calculate The sum of squares of all elements in the sum of squares and The ratio of the sum of squares of all elements in the formula is considered as the first degree of overlap, C1. calculate The sum of squares of all elements in the sum of squares and The ratio of the sum of squares of all elements in the equation is considered as the second degree of conflict overlap, C2. calculate and The difference of the sum of squares of all elements in the set, and The ratio of the sum of squares of all elements in the equation is considered the conflict dispersion C3. Calculate the probability of conflict resolution for the current block. ,in, It is a natural exponential function.

[0066] In this embodiment, the set of conflict dimensions for assessment indicators refers to the set of all conflicting assessment indicators and their corresponding conflicting features obtained after decomposing the contradictory intentions according to all assessment indicators of the psychological assessment of cancer patients. Each element in the set is a "assessment indicator-conflict feature" binary, which is the basis for converting contradictory intentions into a quantitative matrix. For example, the third contradictory intention is "the global contradiction between the patient's high health anxiety and low social willingness and high need for emotional companionship". It can be decomposed into the set of conflict dimensions for assessment indicators as: {(health anxiety level - too high), (social willingness - too low), (need for emotional companionship - too high), (emotional stability level - low)}.

[0067] The indicator association topology refers to the topological structure obtained after modeling the inherent logical relationship between all assessment indicators in the psychological assessment of cancer patients. This structure uses assessment indicators as nodes and the strength of the association between indicators as the weight of the edges, reflecting the mutual influence between the assessment indicators. It should be noted that the edge weights are calculated using the Pearson correlation coefficient, the node size is positively correlated with the frequency of the indicator in the assessment, and the edge thickness is positively correlated with the strength of the association.

[0068] The standardized conflict feature space refers to the standardized high-dimensional feature space constructed for the contradiction analysis of psychological assessments of cancer patients. The dimensions of this space are jointly determined by the assessment indicators and the manifestations of psychological conflict. It is a quantitative mapping space of the set of conflict dimensions. All conflict features of contradictory intentions can be quantitatively represented in this space, realizing the standardized analysis of contradictory intentions.

[0069] The manifestations of psychological conflict refer to the common types of contradictions in the psychological assessment of cancer patients. They are the column dimensions of the standardized conflict feature space. In this embodiment, the manifestations of psychological conflict include five types: numerical contrast, semantic contradiction, mismatch of needs, emotional inconsistency, and contradiction between behavior and psychology, covering the main manifestations of psychological conflict in cancer patients.

[0070] The conflict feature matrix refers to a two-dimensional matrix that quantifies the conflict characteristics of contradictory intentions in a standardized conflict feature space. The row dimension represents all assessment indicators, and the column dimension represents the psychological conflict manifestations. Matrix elements are the quantitative conflict values ​​of the corresponding assessment indicators under the corresponding psychological conflict manifestations, ranging from 0 to 1. The closer the value is to 1, the higher the degree of conflict for that indicator under that manifestation. In this embodiment... This is the conflict feature matrix corresponding to the third contradictory intention. This is the conflict feature matrix corresponding to the fourth contradictory intention. For example... The matrix consists of 4 rows and 5 columns. The rows correspond to four assessment indicators: level of health anxiety, social willingness, need for emotional companionship, and level of emotional stability. The columns correspond to five forms of psychological conflict. The matrix elements... The conflict quantification value, representing the level of health anxiety in the numerical contrast representation, is 0.9.

[0071] In this embodiment, such as , After element-wise intersection operation After element-wise union operation .

[0072] The beneficial effects of the above technical solution are as follows: by decomposing contradictory intentions into a standardized set of conflict dimensions and mapping them to a feature space, a quantitative matrix representation of contradictory intentions is realized. The core conflict and global conflict features are extracted by combining intersection and union operations. The overlap and dispersion of the third and fourth contradictory intentions are accurately quantified through three feature values. Finally, a standardized probability of contradiction resolution is generated through a natural exponential function, realizing the scientific and accurate quantification of the probability of contradiction resolution. This changes the current situation of subjective judgment on the feasibility of contradiction resolution in existing technologies and provides a reliable quantitative basis for subsequent adjustment of resolution probability and fusion of resolution direction.

[0073] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment, determining a first direction for resolving the first contradictory intention, including: The historical dependency database is sliced ​​according to the assessment cycle, and the co-occurrence relationship between behavioral feature data and psychological feature data within the same assessment cycle is extracted to construct a behavioral-psychological dual-dimensional association map. Using the psychological words corresponding to the first contradictory intention as retrieval nodes, a neighborhood traversal is performed in the association graph to match behavioral dependency features and psychological dependency features that are strongly associated with the contradictory intention. Based on the causal relationship between the behavioral dependency features and the psychological dependency features, features that only have co-occurrence associations and no causal relationship are eliminated to form a behavioral dependency profile and a psychological dependency profile. Based on the behavioral trend characteristics in the behavioral dependency profile and the emotional stability characteristics in the psychological dependency profile, the causes of the first contradictory intention are analyzed, and the first resolution direction matching the individual characteristics of the cancer patient is determined.

[0074] In this embodiment, time-series slicing refers to dividing the data in the historical dependency database according to the assessment period, dividing continuous historical data into multiple time slices, each time slice corresponding to the complete data of one assessment period. For example, if the historical dependency database contains 180 days of patient historical data and the assessment period is 30 days, then the data is divided into 6 time-series slices, each 30 days long, and each slice corresponds to the behavioral and psychological data of one assessment period.

[0075] Behavioral characteristic data refers to various types of data stored in the historical dependency database that reflect the behavioral habits and status of cancer patients, including data on self-care behavior, social behavior, medical treatment behavior, diet and rest, etc. It is the basic data for building a behavioral dependency profile. For example, a patient's behavioral characteristic data is: wake up at 6:00 am and go to sleep at 9:00 pm every day; can independently complete washing and dressing, but cannot cook; meets with children once a week and has no other social interactions; takes medication on time every day and goes to the hospital for a check-up once a week.

[0076] Psychological characteristic data refers to various types of data reflecting the psychological state and tendencies of cancer patients stored in historical databases, including past psychological assessment data, psychological vocabulary, ambivalent intentions, and emotional state scores. For example, a patient's psychological characteristic data might be: a health anxiety score of 8 (out of 1-10); a depressed mood score of 5; a need for emotional support score of 9; and a social willingness score of 2.

[0077] Co-occurrence association refers to the relationship between a behavioral characteristic and a psychological characteristic that occur simultaneously within the same assessment period. It is quantified using a correlation coefficient, with a value ranging from 0 to 1. For example, if the behavioral characteristic of "inability to cook independently" and the psychological characteristic of "high need for emotional companionship" occur simultaneously within the same assessment period, the correlation coefficient between the two is 0.85, indicating a strong co-occurrence association between them.

[0078] The Behavioral-Psychological Dual-Dimensional Correlation Map is a visual map constructed using the behavioral and psychological characteristics of cancer patients as nodes and the co-occurrence correlation coefficient between them as the weight of the edges. The map consists of two dimensions: a behavioral characteristic layer and a psychological characteristic layer. The size of the nodes represents the frequency of feature occurrence, and the thickness of the edges represents the magnitude of the correlation coefficient, thus achieving a visual and quantitative representation of the relationship between behavioral and psychological characteristics. For example, in the Behavioral-Psychological Dual-Dimensional Correlation Map, the edge between the behavioral characteristic node "inability to cook independently" and the psychological characteristic node "high need for emotional companionship" is relatively thick, representing a correlation coefficient of 0.85; the edge between the behavioral characteristic node "no social behavior" and the psychological characteristic node "low willingness to socialize" is the thickest, with a correlation coefficient of 0.95.

[0079] A retrieval node refers to the starting node for neighborhood traversal in the behavior-psychology dual-dimensional association graph. In this embodiment, the core psychological vocabulary corresponding to the first contradictory intention is used as the retrieval node. The retrieval node is the starting point of the neighborhood traversal and determines the scope and direction of the traversal. For example, if the core psychological vocabulary corresponding to the first contradictory intention is: self-care in eating - partial ability, dietary care needs - high, then these two words will be used as retrieval nodes for traversal in the association graph.

[0080] Neighborhood traversal refers to the process of starting from a retrieval node and traversing its neighboring nodes in the behavioral-psychological dual-dimensional association graph according to a preset association threshold. In this embodiment, the preset association threshold is 0.7, and only neighboring nodes with an association coefficient greater than or equal to 0.7 are traversed to achieve accurate matching of strongly associated features. For example, starting from the retrieval node "Dietary self-care - partial ability", the traversal reaches the behavioral feature node "unable to cook independently" with an association coefficient of 0.8 and the psychological feature node "strong psychological dependence on children" with an association coefficient of 0.75. It should be noted that, based on the significance test of the Pearson correlation coefficient (P<0.01), when the association coefficient is ≥0.7, the behavioral and psychological features have a strong co-occurrence association, and the false positive rate is ≤3%.

[0081] Strongly associated behavioral / psychological dependency features refer to behavioral and psychological features that are matched through neighborhood traversal, have a correlation coefficient ≥ 0.7 with the retrieval node (the core psychological vocabulary of the first contradictory intention), and are verified to have a causal relationship by Granger causality test. These features are the core manifestation of the patient's individual characteristics and are the core basis for constructing behavioral and psychological dependency profiles.

[0082] Behavioral trend characteristics refer to the trends in a patient's behavioral characteristics over a certain period of time, extracted from the behavioral dependence profile. These include trends in the improvement / deterioration of behavioral abilities and trends in the stability / change of behavioral habits. For example, a patient's behavioral trend characteristics might be: a slow decline in self-care ability over the past three assessment periods, from partial self-care (able to do simple cooking) to only being able to process ingredients but unable to cook; and an increasing trend in dependence on medication reminders.

[0083] In this embodiment, emotional stability characteristics refer to the degree of emotional stability of a patient over a certain period of time, extracted from the psychological dependence profile. These characteristics include quantitative features such as the amplitude of emotional fluctuations, the duration of low mood / anxiety, and emotional regulation ability, with values ​​ranging from 0 to 1. Values ​​closer to 1 indicate greater emotional stability and serve as the core psychological basis for analyzing the triggers of the primary conflict intention. For example, a cancer patient's emotional stability characteristics might be "emotional fluctuation amplitude 0.7, duration of anxiety 21 days / assessment cycle, and emotional regulation ability 0.3," reflecting poor emotional stability and a weak ability to self-regulate emotions.

[0084] The primary cause of the contradictory intention refers to the root reason why a patient experiences psychological conflict in responding to a single assessment question. This cause is derived from an analysis combining behavioral trend characteristics and emotional stability characteristics, and is categorized into three levels: behavioral, psychological, and psychosomatic. It is the core basis for determining the primary direction of resolution. For example, combining the patient's behavioral trend characteristic of "slowly declining self-care ability" with the emotional stability characteristics of "high level of health anxiety and poor emotional stability," the analysis reveals that the primary cause of the contradictory intention "conflict between self-care ability and care needs" is: At the psychosomatic level: The patient experiences health anxiety due to declining self-care ability, wanting to alleviate anxiety by demonstrating self-care ability, yet simultaneously desiring care due to insufficient actual ability.

[0085] The beneficial effects of the above technical solution are as follows: the periodic analysis of historical data is realized through time-series slicing; the constructed behavioral-psychological dual-dimensional correlation map accurately mines the intrinsic relationship between patients' behavior and psychological characteristics; the neighborhood traversal with the core words of the first conflict intention as nodes realizes the accurate matching of strong correlation features; and the causal analysis based on behavioral trends and emotional stability features allows the determination of the first resolution direction to be anchored to the individual behavior and psychological characteristics of the patient, effectively solving the problem of low adaptability between traditional resolution directions and individual patients, improving the personalization and pertinence of the conflict resolution direction, and laying a foundation for the subsequent integration and adjustment of resolution directions to fit the patient's actual situation.

[0086] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment. According to a first connection relationship between a first contradictory intention and a second contradictory intention, a second direction for resolving the first contradictory intention is determined, including: The first contradictory intent is mapped as a local conflict feature sequence at the single-response level, and the second contradictory intent is mapped as a global conflict feature sequence within the same evaluation block; Dynamic time warping matching is performed on the local conflict feature sequence and the global conflict feature sequence to obtain the first connection relationship representing the intention transmission path and the degree of first conflict superposition. Based on the constraint direction of the global conflict feature sequence on the local conflict feature sequence, a conflict mitigation path is constructed in the reverse direction of the intention transmission path, and a second resolution direction matching the first connection relationship is adaptively generated according to the first conflict superposition degree.

[0087] In this embodiment, the local conflict feature sequence refers to a one-dimensional sequence formed by arranging the conflict features at the single response level in a logical order after decomposing the first contradictory intention according to the evaluation index. Each element in the sequence is a triplet of evaluation index-conflict degree-conflict manifestation, reflecting the micro-conflict features and logical relationships in the response to a single evaluation question. For example, if the first contradictory intention is: the contradiction between self-care in eating and the cognition of care needs, the resulting local conflict feature sequence is: [(self-care ability-eating, 0.8, numerical contrast), (care needs-eating, 0.85, semantic contradiction)].

[0088] The global conflict feature sequence refers to a one-dimensional sequence formed by arranging the conflict features at the assessment block level, obtained after decomposing the second contradictory intention according to the assessment indicators, in a logical order. The sequence covers the conflict features of all assessment questions within the assessment block. For example, if the second contradictory intention is: the patient's overall care needs are inconsistent, the resulting global conflict feature sequence is: [(self-care ability - diet, 0.8, numerical contrast), (care needs - diet, 0.85, semantic contradiction), (self-care ability - clothing, 0.6, numerical contrast), (care needs - clothing, 0.7, semantic contradiction)].

[0089] Dynamic time warping matching (VTMM) is an algorithm used to match the similarity of sequences of different lengths and with different feature point distributions. It aligns feature points and plans paths between locally conflicting and globally conflicting feature sequences, calculating their similarity and feature transmission relationships. It is a core algorithm for uncovering the intrinsic connections between two sequences. For example, by using VTMM to match the aforementioned locally and globally conflicting feature sequences, it can achieve precise alignment of dietary-related conflicting feature points, revealing that the local sequence is a core subsequence of the global sequence.

[0090] The intent transmission path refers to the path of conflict feature transmission between the local conflict feature sequence and the global conflict feature sequence obtained through dynamic time warping matching. It reflects the direction and logical order of transmission from micro-level conflict intent to meso-level conflict intent and is one of the core features of the first connection relationship. For example, the intent transmission path of the local and global sequences mentioned above is: self-care ability - eating conflict → care needs - eating conflict → self-care ability - clothing conflict → care needs - clothing conflict, reflecting that the micro-conflict related to eating is the starting point of the global conflict of the assessment block.

[0091] The first degree of conflict overlap refers to the proportion of local conflict features in the global conflict features calculated through dynamic time warping matching. The value ranges from 0 to 1; the closer the value is to 1, the greater the impact of the local conflict on the global conflict. This is another core feature of the first connection relationship. For example, if the calculated degree of overlap of diet-related local conflicts in the global conflict of the care needs assessment block is 0.8, it reflects that this local conflict is a major component of the global conflict.

[0092] The first connection relationship refers to the inherent connection between the first contradictory intention and the second contradictory intention, which are jointly constituted by the intention transmission path and the degree of superposition of the first conflict. This relationship is quantitatively represented in the form of path + superposition degree, reflecting the connection strength and transmission logic between micro-contradictions and meso-contradictions.

[0093] The constraint direction refers to the direction in which the global conflict feature sequence constrains the local conflict feature sequence. In other words, it reflects the global rules that must be followed in resolving local conflicts, indicating the direction of the global conflict resolution needs to guide the resolution of local conflicts. For example, if the global conflict resolution need of the care needs assessment block is: to unify the expression of care needs and adapt to the patient's actual self-care ability, then the constraint direction for local conflicts related to diet is: to resolve the cognitive contradiction between dietary self-care and care needs, so that the expression of dietary care needs matches the actual dietary self-care ability.

[0094] A conflict mitigation path refers to a conflict resolution path constructed from the global to the local level, based on the direction of constraints and in reverse along the intention transmission path. This path represents the concrete implementation logic of global conflict resolution needs in local conflict resolution. For example, a conflict mitigation path constructed in reverse along the intention transmission path could be: first, unify the global goal expressed by care needs → resolve the conflict between care needs and clothing → resolve the conflict between self-care ability and clothing → resolve the conflict between care needs and eating → resolve the conflict between self-care ability and eating.

[0095] Adaptive generation refers to dynamically adjusting the intensity and implementation details of care measures in the second resolution direction based on the numerical value of the first conflict superposition degree. The higher the superposition degree, the more specific and intense the care measures, so that the matching degree between the second resolution direction and the first connection relationship is adaptively adjusted. For example, if the first conflict superposition degree is 0.8 (high), the details and implementation frequency of dietary care measures are increased in the second resolution direction; if the superposition degree is 0.4 (low), the dietary care measures are simplified, focusing on resolving other conflicts globally.

[0096] The beneficial effects of the above technical solution are as follows: by mapping contradictory intentions into standardized conflict feature sequences, a quantitative comparison of micro and meso-level contradictions is achieved. The dynamic time warping matching algorithm accurately mines the first connection relationship between the two. The conflict mitigation path constructed in reverse along the intention transmission path allows the resolution of local contradictions to follow the global resolution requirements. Based on the adaptive generation mechanism of the first conflict superposition degree, the second resolution direction is highly matched with the first connection relationship, realizing the coordinated unity of the resolution directions of micro and meso-level contradictions and improving the overall and rational nature of contradiction resolution.

[0097] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment. According to a second connection relationship between a first contradictory intention and a third contradictory intention, a first adjustment is made to the probability of conflict resolution of the assessment block matching the first contradictory intention, including: The first contradictory intention is mapped as a single-problem local conflict feature, and the third contradictory intention is mapped as a global correlation conflict feature between multiple evaluation blocks; The intent transmission consistency is verified between the local conflict features and the global related conflict features, and a second connection relationship is constructed to characterize the conflict transmission direction, the second conflict superposition intensity, and the cross-block influence range. When the direction of conflict propagation is consistent, the resolution probability is amplified in the positive direction according to the superposition intensity of the second conflict. When the direction of conflict propagation is inconsistent, the resolution probability is negatively reduced according to the cross-block influence range.

[0098] In this embodiment, the amplification factor calibration formula is: k1=1+S2, where S2 is the second collision superposition intensity, and the adjusted probability is: Pa×k1; The shrinkage coefficient calibration formula is: k2=1 / (N+1), where N is the number of evaluation blocks in the cross-block influence range, and the adjusted probability is: Pa×k2; It should be noted that the adjusted probability shall not exceed 1. If it is greater than 1, the adjusted probability shall be 1.

[0099] In this embodiment, the single-problem local conflict feature refers to the micro-conflict feature corresponding to a single assessment question obtained after refining the first contradictory intention. It is represented in the form of a feature vector, where the vector dimension is the assessment index and the element is the degree of conflict of the corresponding index, reflecting the core features of the single-problem contradiction. For example, if the first contradictory intention is the contradiction between self-care in eating and the perception of care needs, the single-problem local conflict feature obtained by mapping is: [0.8 (self-care ability - eating), 0.85 (care needs - eating)].

[0100] The global relational conflict feature refers to the macro-level conflict features covering all assessment blocks obtained after extracting the third contradictory intention. It is represented in the form of a feature matrix, where the row dimension represents the assessment block, the column dimension represents the assessment indicator, and the element represents the degree of conflict (dimensionless value from 0 to 1) between the corresponding assessment block and assessment indicator, reflecting the global relational conflict features between multiple assessment blocks. For example, if the third contradictory intention is the global conflict between a patient's high health anxiety and low social willingness and high need for emotional companionship, the resulting global relational conflict feature is a 4-row, 8-column matrix. The rows correspond to 4 assessment blocks (emotional state, care needs, health anxiety, social willingness), the columns correspond to 8 core assessment indicators (degree of depressive mood, self-care ability, care need tendency, etc.), and the element values ​​are the degree of conflict in each dimension.

[0101] Intent transmission consistency verification refers to verifying the consistency of the transmission logic between local conflict characteristics of a single problem and global related conflict characteristics. The verification content includes whether the direction, intensity and scope of conflict transmission match. The consistency coefficient is used to quantify the verification result, with a value range of 0-1. The closer the value is to 1, the higher the consistency.

[0102] The direction of conflict transmission refers to the direction in which local conflict features, obtained through the consistency check of intent transmission, are transmitted to globally related conflict features. It is divided into two cases: consistent and inconsistent. Consistency indicates that the transmission direction of local conflict is the same as the formation direction of global conflict; inconsistency indicates that the transmission directions are opposite. For example, if the transmission direction of local dietary conflict is: care needs → health anxiety, and the formation direction of global conflict is: conflicting care needs → increased health anxiety → decreased social willingness, then the two conflict transmission directions are consistent.

[0103] The second conflict superposition intensity refers to the superposition intensity of local conflict features within the global interconnected conflict features. Its value ranges from 0 to 1; the closer the value is to 1, the greater the contribution and the stronger the influence of the local conflict on the global conflict. For example, the second conflict superposition intensity of a diet-related local conflict within the global interconnected conflict is 0.75, reflecting that this local conflict has a strong driving effect on the formation of global contradictions.

[0104] Cross-block influence range refers to the extent to which a local conflict feature extends beyond its own assessment block and affects other assessment blocks. It is quantified by the number of affected assessment blocks and takes the value of an integer greater than or equal to 1. The larger the value, the wider the cross-block influence. For example, a local conflict related to diet not only affects the care needs assessment block but also the emotional state assessment block and the health anxiety assessment block, with a cross-block influence range of 3.

[0105] The second connection relationship refers to the inherent connection between the first and third contradictory intentions, which are composed of the direction of conflict transmission, the intensity of the second conflict superposition, and the cross-block influence range. This relationship is quantitatively represented in the form of direction + intensity + range, reflecting the connection characteristics and influence degree of micro-conflicts and macro-global conflicts. For example, the second connection relationship between the first and third contradictory intentions mentioned above is: the direction of conflict transmission is consistent (dietary conflict → health anxiety conflict), the intensity of the second conflict superposition is 0.8, and the cross-block influence range is 3.

[0106] Positive amplification of the resolution probability refers to multiplying the original resolution probability by an amplification factor determined by the superposition intensity of the second conflict when the conflict propagation direction is consistent. This results in an adjusted resolution probability higher than the original probability. The amplification factor is positively correlated with the superposition intensity of the second conflict and ranges from 1 to 2. For example, if the original resolution probability is 0.6 and the superposition intensity of the second conflict is 0.75, corresponding to an amplification factor of 1.5, the resolution probability after positive amplification is 0.6 × 1.5 = 0.9.

[0107] The negative contraction resolution probability refers to the probability of conflict resolution being lower than the original probability when the direction of conflict transmission is inconsistent. This is achieved by multiplying the original resolution probability by a contraction coefficient determined by the cross-block influence range. The contraction coefficient is negatively correlated with the cross-block influence range and ranges from 0 to 1. For example, if the original resolution probability is 0.6 and the cross-block influence range is 3, corresponding to a contraction coefficient of 0.5, the resolution probability after negative contraction would be 0.6 × 0.5 = 0.3.

[0108] The beneficial effects of the above technical solution are as follows: by mapping the intentions of micro and macro contradictions into standardized conflict characteristics, the quantitative comparison and analysis of the two are realized. The consistency verification of intention transmission accurately uncovers the second connection relationship between the two. Differentiated probability adjustment strategies are adopted according to different conflict transmission directions. Positive amplification and negative contraction make the adjusted contradiction resolution probability more in line with the correlation characteristics of local and global contradictions, improving the dynamism and scientific nature of the resolution probability, and providing a more accurate quantitative basis for the subsequent integration and adjustment of resolution direction.

[0109] This invention provides a smart elderly care decision-making method for cancer patients based on dynamic psychological assessment. The method involves adjusting the first resolution direction according to a second resolution direction and a first adjustment result, and obtaining a corrected response to the corresponding direct assessment response. The method includes: The first and second resolution directions are mapped to the same standardized psychological conflict feature space, and basic resolution vectors of the same dimension are constructed. With the guided resolution vector ,in, and The dimensions of each feature are equal to the feature dimensions of the standardized psychological conflict feature space. The feature dimensions are determined by the product of the number of psychological assessment indicators and the number of corresponding conflict manifestations. All feature dimensions are normalized. The first adjusted probability of conflict resolution is denoted as Pa. Based on the aforementioned basic resolution vector With the guided resolution vector , determine the meaning Figure 1 Consistency coefficients and conflict offset coefficients, combined with Pa, are used to construct a nonlinear adaptive direction fusion model for the basic resolution vector. Implement the second adjustment; Based on the second adjustment results, conflict feature stripping, semantic vector projection, and logical self-consistency reconstruction were performed on the psychological vocabulary of the direct assessment responses to generate corrected responses.

[0110] In this embodiment, it is intended Figure 1 Coherence coefficient Conflict offset coefficient .

[0111] In this embodiment, the nonlinear adaptive direction fusion model is as follows: The gradient descent method is used to solve for the final solution vector. The learning rate is set to 0.01, and iterations continue until the vector converges, meaning the difference between two adjacent iterations is ≤0.001. , For vectors of the same dimension, through Achieving global control over the feasibility of conflict, through To suppress the excessive dominance of the fundamental resolution direction in high-offset scenarios, by... Strengthening the constraint role of the second resolution direction in low-consistency scenarios ensures that the final resolution direction simultaneously anchors to the patient's individual behavioral psychological profile, assesses intra-block conflict consistency, and evaluates the feasibility of cross-block conflict resolution. It should be noted that... , , .

[0112] In this embodiment, the basic resolution vector Specific calculation method: Based on behavioral and psychological dependency profiles of cancer patients, n core features related to the primary conflict intention are extracted, namely, patient behavioral habit fit. Psychological and emotional compatibility Correlation of Contradictory Causes Past resolution fit rate Individual tolerance Furthermore, the remaining characteristics are sequentially ordered according to the psychological assessment indicators. After quantifying each core characteristic, the following formula is used to calculate... : ,in, , The maximum value in set X; Guided resolution vector The specific calculation method is as follows: Based on the global conflict features within the evaluation block, n corresponding core features related to the second conflict intention are extracted, which are the intra-block conflict fit degrees. Alignment with overall intent Conflict mitigation correlation Intrablock digestion fit rate Constraint fit Furthermore, the remaining characteristics are sequentially ordered according to the psychological assessment indicators. After quantifying each core characteristic, the following formula is used to calculate... : ,in, , The maximum value in set Y.

[0113] In this embodiment, the standardized psychological conflict feature space refers to a standardized high-dimensional feature space constructed for the resolution of psychological conflicts in cancer patients. The feature dimension of this space is determined by the product of the number of psychological assessment indicators and the number of psychological conflict manifestations. Each point in the space represents a specific resolution direction, realizing the quantification and standardized representation of the resolution direction. For example, if there are 8 psychological assessment indicators and 5 psychological conflict manifestations, the feature dimension of the standardized psychological conflict feature space is 40, and all resolution directions are mapped to a 40-dimensional vector.

[0114] For example, the first resolution direction is: semi-self-service dietary care + children accompanying meals, which maps to a 40-dimensional basic resolution vector. In the study, the reduction intensity of the dimensions related to self-care in eating and emotional companionship was 0.8-0.9, while the other dimensions were 0.1-0.3.

[0115] For example, the second resolution direction is: semi-self-service dietary care + emotional support + children accompanying meals, which maps to a 40-dimensional directional resolution vector. In the study, the resolution intensity of the dimensions related to self-care in eating, emotional companionship, and emotional guidance was 0.75-0.95, while the resolution intensity of the other dimensions was 0.1-0.3.

[0116] Conflict feature stripping refers to the process of removing features based on the final resolution vector. The aim is to extract contradictory characteristic words from the original psychological vocabulary of direct assessment responses, eliminate non-objective and contradictory semantic information, and retain true and effective psychological characteristic information. For example, from the original psychological vocabulary of "self-care" and "complete dependence," the contradictory characteristic word "complete dependence" is extracted, "self-care" is retained and modified to "partial self-care," targeting... Output interpretation rules: Final resolution vector The dimensions are consistent with the standardized psychological conflict characteristic space (40 dimensions, 8 assessment indicators × 5 conflict manifestations). Dimension value ≥ 0.7: This is identified as a core area for resolution, and care measures should be designed around this dimension. Dimension values ​​∈ [0.3, 0.7): These are determined to be secondary resolution directions, appropriately reflected in care measures, and supplementing the core directions; Dimension value < 0.3: Determined to have no need for resolution, and does not need to be considered in care measures; Sort by dimension value from high to low, extract the top 3 core resolution directions, and use them as the core basis for subsequent correction and response generation and care plan design.

[0117] Semantic vector projection employs a cosine similarity projection algorithm to project the semantic vectors of mental lexical terms, after removing conflict features, onto a standardized mental conflict feature space, ensuring that the semantics of the mental lexical terms align with the final resolution vectors. The direction of resolution must be consistent, i.e., the cosine similarity is ≥0.8. Otherwise, the vocabulary needs to be readjusted to achieve semantic matching with the resolution direction.

[0118] The logically consistent reconstruction is achieved by using a large language model trained on a corpus of psychological assessments of cancer patients, which is adapted to the language expression habits of elderly cancer patients. This large language model is trained on samples based on the results of semantic recombination and logical optimization of psychological vocabularies after conflict feature stripping and semantic vector projection.

[0119] The revised response refers to the assessment response obtained by removing contradictions and non-objective information through conflict feature stripping, semantic vector projection and logical self-consistency reconstruction. It is a precise correction of the original direct assessment response and provides reliable psychological data basis for the generation of elderly care sub-plans.

[0120] The beneficial effects of the above technical solution are: by mapping the digestion direction to a standardized digestion vector, the quantitative fusion of digestion directions in different dimensions is achieved, which means... Figure 1 The consistency coefficient and conflict offset coefficient accurately quantify the correlation characteristics of the two resolution directions. The constructed nonlinear adaptive direction fusion model realizes the adaptive fusion of the basic resolution direction and the guided resolution direction, so that the final resolution direction simultaneously anchors the individual characteristics of the patient, the global constraints within the assessment block and the feasibility of cross-block resolution. The corrected response generated by conflict feature stripping, semantic vector projection and logical self-consistency reconstruction effectively eliminates the contradictions and non-objective information in the original assessment response, improves the authenticity and reliability of the assessment data, provides accurate psychological data support for the generation of elderly care sub-plans, and greatly improves the adaptability and scientificity of the care plan.

[0121] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A smart elderly care decision-making method for cancer patients based on dynamic psychological assessment, characterized in that, include: Step 1: Distribute psychological assessment files to cancer patients and capture the initial assessment set of the cancer patients for each assessment block in the psychological assessment files, wherein the initial assessment set contains the direct assessment responses of each assessment question in the assessment block; Step 2: Analyze each direct assessment response according to the assessment indicators of the corresponding assessment question to obtain the direct psychological vocabulary of each assessment indicator. Perform a first contradiction analysis on all direct psychological vocabulary under the same direct assessment response to determine the first contradiction intention. At the same time, perform a second contradiction analysis on the vocabulary vector formed by all direct psychological vocabulary under each direct assessment response in the same assessment block to determine the second contradiction intention. Step 3: Perform a third contradiction analysis on all evaluation blocks according to the core word vector of each evaluation block to determine the third contradiction intention, and perform a fourth contradiction analysis on all evaluation blocks after removing the core word vector of the current block to determine the fourth contradiction intention. Based on the third contradiction intention and the fourth contradiction intention, determine the contradiction resolution probability of the corresponding current block. Step 4: Construct the behavioral dependency profile and psychological dependency profile of the cancer patient based on the historical dependency database of the latest cycle, and determine the first direction for resolving the first contradictory intention. At the same time, determine the second direction for resolving the first contradictory intention based on the first connection relationship between the first contradictory intention and the second contradictory intention. Step 5: Based on the second connection relationship between the first contradictory intention and the third contradictory intention, make a first adjustment to the contradiction resolution probability of the evaluation block that matches the first contradictory intention; Step 6: Adjust the first resolution direction according to the second resolution direction and the first adjustment result, and obtain the corrected response of the corresponding direct assessment response. Obtain the elderly care sub-plan of the corresponding assessment block according to the latest psychological vocabulary of all corrected responses under the same assessment block. Step 7: Update the elderly care sub-plan for each assessment block in real time according to the assessment cycle, and obtain the comprehensive elderly care plan for the latest assessment cycle.

2. The intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment according to claim 1, characterized in that, Perform a primary contradiction analysis on all direct psychological terms under the same direct assessment response to determine the primary contradiction intention, including: According to the evaluation indicators of the evaluation questions, all direct psychological terms under the corresponding direct evaluation responses are digitized and the first conflict number with a numerical difference greater than the preset difference is mined. At the same time, according to the penetration length of each evaluation indicator corresponding to the same direct evaluation response in the evaluation question, a penetration mining mechanism is added to mine the second conflict number. Extract the direct mental words of the first conflicting number and the second conflicting number, and regard the overlapping words as the first word and the non-overlapping words as the second word; When the number of the first word is 0 or 1, all the second words are clustered using the words corresponding to the first conflict number as the cluster. When the number of the first word is greater than 1, all the second words are clustered using the first word as the cluster. Based on the clustering results and in combination with all the indicators involved in each clustering block, the first contradictory intent is determined.

3. The intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment according to claim 1, characterized in that, A second contradiction analysis is performed on the lexical vectors formed by all direct psychological words under each direct assessment response in the same assessment block to determine the second contradiction intention, including: Retrieve the contradiction assessment model that matches the assessment block from the model database; The vocabulary vectors are input into the contradiction assessment model to perform a second contradiction analysis, and the second contradiction intention is output.

4. The intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment according to claim 1, characterized in that, Based on the third and fourth contradictory intentions, the probability of contradiction resolution for the current block is determined, including: The third and fourth contradictory intentions are decomposed into sets of conflict dimensions of assessment indicators. Based on the indicator association topology of psychological assessment of cancer patients, the two sets of conflict dimensions are mapped to a standardized conflict feature space to generate the third conflict feature matrix. Fourth Conflict Feature Matrix In this matrix, the row dimension of the conflict feature matrix represents the evaluation indicators, the column dimension represents the psychological conflict manifestations, and the matrix elements represent the conflict quantification values ​​of the corresponding indicators under the corresponding manifestations. right and Perform element-wise intersection operation to obtain the core conflict feature matrix. At the same time, for and Performing element-wise union operation yields the global conflict feature matrix. The intersection operation is to take... and The minimum value of the elements at corresponding positions, the union operation is to take and The maximum value of the element at the corresponding position; calculate The sum of squares of all elements in the sum of squares and The ratio of the sum of squares of all elements in the formula is considered as the first degree of overlap, C1. calculate The sum of squares of all elements in the sum of squares and The ratio of the sum of squares of all elements in the equation is considered as the second degree of conflict overlap, C2. calculate and The difference of the sum of squares of all elements in the set, and The ratio of the sum of squares of all elements in the equation is considered the conflict dispersion C3. Calculate the probability of conflict resolution for the current block. ,in, It is a natural exponential function.

5. The intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment according to claim 1, characterized in that, Determining the first direction for resolving the first contradictory intention includes: The historical dependency database is sliced ​​according to the assessment cycle, and the co-occurrence relationship between behavioral feature data and psychological feature data within the same assessment cycle is extracted to construct a behavioral-psychological dual-dimensional association map. Using the psychological words corresponding to the first contradictory intention as retrieval nodes, a neighborhood traversal is performed in the association graph to match behavioral dependency features and psychological dependency features that are strongly associated with the contradictory intention. Based on the causal relationship between the behavioral dependency features and the psychological dependency features, features that only have co-occurrence associations and no causal relationship are eliminated to form a behavioral dependency profile and a psychological dependency profile. Based on the behavioral trend characteristics in the behavioral dependency profile and the emotional stability characteristics in the psychological dependency profile, the causes of the first contradictory intention are analyzed, and the first resolution direction matching the individual characteristics of the cancer patient is determined.

6. The intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment according to claim 1, characterized in that, Based on the first connection relationship between the first contradictory intention and the second contradictory intention, a second direction for resolving the first contradictory intention is determined, including: The first contradictory intent is mapped as a local conflict feature sequence at the single-response level, and the second contradictory intent is mapped as a global conflict feature sequence within the same evaluation block; Dynamic time warping matching is performed on the local conflict feature sequence and the global conflict feature sequence to obtain the first connection relationship representing the intention transmission path and the degree of first conflict superposition. Based on the constraint direction of the global conflict feature sequence on the local conflict feature sequence, a conflict mitigation path is constructed in the reverse direction of the intention transmission path, and a second resolution direction matching the first connection relationship is adaptively generated according to the first conflict superposition degree.

7. The intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment according to claim 1, characterized in that, Based on the second connection relationship between the first contradictory intent and the third contradictory intent, a first adjustment is made to the contradiction resolution probability of the evaluation block matching the first contradictory intent, including: The first contradictory intention is mapped as a single-problem local conflict feature, and the third contradictory intention is mapped as a global correlation conflict feature between multiple evaluation blocks; The intent transmission consistency is verified between the local conflict features and the global related conflict features, and a second connection relationship is constructed to characterize the conflict transmission direction, the second conflict superposition intensity, and the cross-block influence range. When the direction of conflict propagation is consistent, the resolution probability is amplified in the positive direction according to the superposition intensity of the second conflict. When the direction of conflict propagation is inconsistent, the resolution probability is negatively reduced according to the cross-block influence range.

8. The intelligent elderly care decision-making method for cancer patients based on dynamic psychological assessment according to claim 1, characterized in that, Based on the second resolution direction and the first adjustment result, the first resolution direction is adjusted a second time, and a corrected response to the corresponding direct evaluation response is obtained, including: The first and second resolution directions are mapped to the same standardized psychological conflict feature space, and basic resolution vectors of the same dimension are constructed. With the guided resolution vector ,in, and The dimensions of each feature are equal to the feature dimensions of the standardized psychological conflict feature space. The feature dimensions are determined by the product of the number of psychological assessment indicators and the number of corresponding conflict manifestations. All feature dimensions are normalized. The first adjusted probability of conflict resolution is denoted as Pa. Based on the aforementioned basic resolution vector With the guided resolution vector The intention consistency coefficient and conflict offset coefficient are determined, and combined with Pa, a nonlinear adaptive direction fusion model is constructed to resolve the basic resolution vector. Implement the second adjustment; Based on the second adjustment results, conflict feature stripping, semantic vector projection, and logical self-consistency reconstruction were performed on the psychological vocabulary of the direct assessment responses to generate corrected responses.

Images