A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness

Abstract

The present invention provides a dynamic q-order hesitant fuzzy decision-making method for the diagnosis and treatment of mental illness, including two steps of dynamic weighting and dynamic decision-making. Since the diagnosis and treatment process of complex mental illness usually requires the coordination of various treatment plans, and the treatment plans at each dynamic moment change dynamically, the present invention adopts a dynamic fuzzy weight calculation method and a dynamic fuzzy preference relationship matrix calculation method to provide a comprehensive formulation method for dynamically changing treatment plans. It not only takes into account the dynamic fuzzy correlation between the patient's psychological state and clinical manifestation characteristics in the symptoms of the dynamic fuzzy diagnosis and treatment process, but also ensures the dynamic changes of the patient's psychological characteristics, clinical manifestation characteristics and treatment plans, as well as the historical feedback capability of the dynamic fuzzy diagnosis and treatment evaluation and the combination of the two. Therefore, the present invention has strong dynamic adaptability and flexible processing of fuzzy information in the fuzzy diagnosis and treatment process of mental illness.

Description

Technical Field

[0001] The invention belongs to the technical field of information processing and relates to a dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness. Background Art

[0002] Multi-attribute decision-making refers to the decision-making process of selecting the best option or ranking options when there are multiple attributes in the options. However, complex environments often make it difficult for decision makers to make accurate judgments on a matter. Decision makers are uncertain and hesitant about the attribute evaluation of some options. Therefore, this type of decision-making problem is called a multi-attribute decision-making problem under fuzzy environment.

[0003] When making fuzzy decisions, decision makers are often hesitant in their evaluation of an option, making it difficult to reach a consensus. For example, when a decision maker evaluates something as "not bad but average", the decision maker's evaluation includes two fuzzy evaluations, "not bad" and "average", which is also a drawback of fuzzy sets.

[0004] As an extension of fuzzy sets, the basic components of hesitant fuzzy sets are hesitant fuzzy elements, each of which is a set of several possible memberships; in other words, a fuzzy set can be regarded as a hesitant fuzzy set containing only one membership. Therefore, hesitant fuzzy sets more comprehensively characterize the hesitant information of decision makers than the extensions of other fuzzy sets, which also extends multi-attribute decision-making under fuzzy environments to multi-attribute decision-making under hesitant fuzzy environments. In 2019, Liu et al. extended hesitant Pythagorean fuzzy sets and hesitant Fermat fuzzy sets, and proposed the concept of q-rungorthopair hesitant fuzzy sets (q-ROHFS). The q-rungorthopair hesitant fuzzy sets extend the restrictions of dual hesitant fuzzy sets, hesitant Pythagorean fuzzy sets and hesitant Fermat fuzzy sets, and can describe a larger hesitant fuzzy space, so it is more flexible.

[0005] Based on the q-order pair hesitant fuzzy set, the decision maker has multiple discrete membership and non-membership degrees for judging a thing, which combines the idea of ​​hesitation into fuzzy decision making. Therefore, in a more hesitant and uncertain environment, hesitant fuzzy decision making has greater advantages than fuzzy decision making. In a complex fuzzy background, there may be potential correlations between multiple attributes, and the attribute correlation makes the fuzzy decision problem face many challenges; however, how to consider the correlation and potential dependence between multiple attributes in the q-order pair hesitant fuzzy background is a difficult point: first, most of the existing fuzzy multi-attribute decision-making methods are carried out under the condition that multiple attributes are independent of each other, which is not the case in reality. Under the non-independence condition of the hesitant fuzzy environment, how to ensure the efficiency and rationality of the attribute evaluation of the q-order pair hesitant fuzzy information is a field worthy of research; secondly, in a complex dynamic hesitant fuzzy environment, how to ensure the dynamic attribute correlation and dependence while ensuring efficient computational efficiency is also a problem worth exploring.

[0006] Dynamic multi-attribute decision-making methods need to take into account the characteristics of the dynamic changes of attributes or decision makers over time. In real life, as time goes by, people's views on a thing often change. Affected by time and environmental factors, attributes or weights may also change; however, there are still certain defects in the current research on dynamic multi-attribute decision-making in q-order hesitant fuzzy environments: first, in complex dynamic fuzzy decision-making, it is necessary to consider the dynamics of three aspects of fuzzy multi-attribute decision-making, one is the dynamic change of multiple alternative sets, the second is the dynamic change of multiple attribute sets, and the third is the dynamic fuzzy decision with historical feedback mechanism. However, in the process of disease diagnosis and treatment using the dynamic multi-attribute decision-making method under the existing fuzzy background, it is often impossible to meet all three dynamics at the same time, which will lead to the poor dynamic fuzzy information processing ability and dynamic fuzzy adaptability of complex multi-attribute decision-making methods in dealing with long-term disease diagnosis and treatment; second, in the process of long-term disease diagnosis and treatment, the potential correlation between multiple attributes is difficult to calculate in the diagnosis and treatment process of dynamic fuzzy decision-making using fuzzy information processing, and the existing methods often ignore the potential dynamic correlation between dynamic multi-attributes. Summary of the invention

[0007] The purpose of the present invention is to provide a dynamic q-order hesitant fuzzy decision-making method for the diagnosis and treatment of mental illness in response to the problems existing in the prior art, which solves the following problems: 1. When making complex dynamic fuzzy decisions in the process of long-term disease diagnosis and treatment, the dynamic multi-attribute decision-making method under the existing fuzzy background cannot simultaneously meet the three dynamic characteristics of dynamic changes of multiple alternative sets, dynamic changes of multiple attribute sets, and dynamic fuzzy decisions with a historical feedback mechanism, resulting in the existing complex multi-attribute decision-making method having poor dynamic fuzzy information processing capability and dynamic fuzzy adaptability in handling long-term disease diagnosis and treatment; 2. The potential correlation between multiple attributes is difficult to calculate, and the existing decision-making methods often ignore the potential dynamic correlation between dynamic multi-attributes.

[0008] To this end, the present invention adopts the following technical solutions:

[0009] A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness, including dynamic weighting and dynamic decision-making;

[0010] The dynamic weighting includes the following steps:

[0011] Step 1: Establish the q-order hesitant fuzzy decision matrix D at time t (t) ;

[0012] Specifically, the q-order hesitant fuzzy decision matrix at time t is:

[0013]

[0014] Where t∈T represents the moment of the dynamic moment set T, It represents the hesitant fuzzy number evaluation value of the q-order pair for the i alternative under the j attribute at time t; It represents the membership set of the evaluation value of a candidate i under attribute j at time t; Represents the non-membership set of the evaluation value of a certain candidate i under attribute j at time t.

[0015] Step 2: Through the decision matrix D in step 1 (t) Calculate the fuzzy entropy matrix at time t, and then calculate the cross entropy between every two attributes using the cross entropy formula;

[0016] Specifically, the hesitant fuzzy number of any two q-order pairs is ξ 1 = 1 ,v 1 >、ξ 2 = 2 ,v 2 >, calculate the fuzzy number ξ 1 , 2 Two cross entropy formulas for M-type fuzzy cross entropy CE M and N-type fuzzy cross entropy CE​​N As follows:

[0017]

[0018] In the formula, represents the number of membership degrees of ξ 1 ; represents the number of non - membership degrees of ξ 1 ; represents the number of membership degrees of ξ 2 ; represents the number of non - membership degrees of ξ 2 ; The parameter s satisfies 1 < s ≤ 2; The parameter p satisfies p > 0; u 1 and v 1 respectively represent the membership degree set and non - membership degree set of the fuzzy number ξ 1 ; u 2 and v 2 respectively represent the membership degree set and non - membership degree set of the fuzzy number ξ 2 ; θ 1 represents the membership degree element of the membership degree set u 1 of the fuzzy number ξ 1 ; θ 2 represents the membership degree element of the membership degree set u 2 of the fuzzy number ξ 2 ; represents the non - membership degree element of the non - membership degree set v 1 of the q - order pair fuzzy number ξ 1 ; represents the non - membership degree element of the non - membership degree set v 2 of the q - order pair fuzzy number ξ 2 ; π 1 represents the hesitancy degree of the q - order pair fuzzy number ξ 1 ; π 2 represents the hesitancy degree of the q - order pair fuzzy number ξ 2 ;

[0019] Step 3: Calculate the fuzzy measure and its Banzhaf value of each attribute at time t;

[0020] Specifically, the formula for calculating the fuzzy measure of each attribute at time t is as follows:

[0021]

[0022] In the formula, μ represents the fuzzy measure of the attribute set; C = {c 1 , c 2 , …, c n} represents an attribute set for multi - attribute decision - making; E(ξ ij ) represents the q - order pair hesitant fuzzy number ξij The fuzzy entropy of ij ,ξ sj ) represents the q-order hesitant fuzzy cross entropy; m represents the number of alternatives in the alternative set;

[0023] The calculation formula of each attribute Banzhaf value is as follows:

[0024]

[0025] Where B represents the Banzhaf function and S represents a subset of the attribute set C.

[0026] Step 4: Calculate the static weight of the attribute set at time t and the weight of the historical attribute set at time t;

[0027] Specifically, the calculation formula for the static weight of the attribute set at time t is as follows:

[0028]

[0029] In the formula, Represents the static weight of the attribute at time t;

[0030] The calculation formula of the historical attribute set weight at time t is as follows:

[0031]

[0032] In the formula, B h Represents the historical attributes at time t Banzhaf value; represents the historical attribute set at time t; h represents the historical attribute subscript. If it is the initial time, then

[0033] Step 5: Calculate the attribute set C at time t (t) With historical attributes Dynamic attribute set weights

[0034] Specifically, the attribute set C at time t (t) ={c 1 ,c 2 ,…,c n}'s static weights and historical attribute sets The dynamic weight of the attribute set at time t is obtained by The calculation formula is as follows:

[0035]

[0036] In the formula, F IOWARepresents the weight of the historical attribute set at time t and static weight of attribute set at time t The induced ordered weighted average function; Q(x) = (-x 2 +3x) / 2 represents a regular increasing monotonic quantization function; q represents a dynamic time variable; Represents the union of the historical attribute set and the current attribute set at time i of the dynamic time variable q The weight of Represents the weight vector at each moment i, that is, the degree of attention that the dynamic decision makes to moment i.

[0037] Step 6: Calculate the attribute set C at time t (t) The dynamic weight W (t) ;

[0038] Specifically, the attribute set C at time t (t) The dynamic weight W (t) The calculation formula is as follows:

[0039]

[0040] In the formula, Indicates that the dynamic attribute set at time t contains only the current attribute set C at time t (t) Dynamic weight of .

[0041] The dynamic decision-making includes the following steps:

[0042] Step 7: Calculate the relative closeness of the fuzzy evaluation of each attribute at time t and sort them;

[0043] Specifically, the calculation formula for the relative closeness of the fuzzy evaluation under each attribute at time t is as follows:

[0044]

[0045] In the formula, RC j Represents the fuzzy evaluation value at time t for the relative closeness of positive and negative ideal solutions; Represents the fuzzy evaluation value With positive ideal evaluation value The distance between Represents the fuzzy evaluation value With negative ideal evaluation value The distance between

[0046] According to the relative proximity RC j The options under each attribute are sorted, and the sorting formula is as follows:

[0047]

[0048] In the formula, A j Indicates that in attribute c j The alternative set under ; σ(i) represents the sorting order, which satisfies

[0049] Step 8: Calculate the candidate set at time t with respect to attribute c j The static fuzzy preference relation matrix

[0050] Specifically, the candidate set at time t is about attribute c j The static fuzzy preference relation matrix The calculation formula is as follows

[0051]

[0052] In the formula, The static fuzzy preference relationship conversion formula for the sth candidate to the kth candidate is expressed as:

[0053]

[0054] In the formula, s and k are both positive real numbers greater than 0; r max The relative closeness RC calculated by expression (10) j The maximum value of r(a s ) and r(a k ) represent alternative a s and alternative a k fuzzy preference value.

[0055] Step 9: The dynamic weight W at time t obtained in step 6 (t) And step 8 obtains the static fuzzy preference relationship matrix for each attribute The static fuzzy preference relation matrix of all attributes of the integrated candidate set at time t

[0056] Specifically, the integrated calculation formula in step 9 is as follows:

[0057]

[0058] In the formula, represents the attribute set at time t; Represents the integrated static fuzzy preference relationship between alternatives s and k at time t.

[0059] Step 10: The static fuzzy preference relationship matrix obtained in step 9 Calculate the dynamic fuzzy preference relationship matrix at time t

[0060] Specifically, the dynamic fuzzy preference relationship matrix at time t is The calculation formula is as follows:

[0061]

[0062] In the formula, represents the dynamic fuzzy preference for alternative s and alternative k at time t, which is expressed as:

[0063]

[0064] Where D E represents an enhanced integration operator.

[0065] Step 11: According to the dynamic fuzzy preference relationship matrix at time t in step 10 Constructing ranking indicators;

[0066] Specifically, the calculation formula of the ranking index is as follows:

[0067]

[0068] In the formula, m represents the number of alternatives in the alternative set; I (t) (a i ) indicates that other alternatives at time t tend to favor alternative a i The preference value of (t) (a i ) represents the alternative a at time t i The preference value towards other alternatives; specifically, V (t) (a i ) is larger, indicating a is more preferred. i .

[0069] The beneficial effects of the present invention are:

[0070] 1. The present invention can take into account the dynamic fuzzy correlation between the long-term dynamic fuzzy changes in the clinical manifestations of mental illness and the psychological state of the patient; specifically, since the uncertainty of long-term dynamic diagnosis and treatment needs to take into account the correlation of fuzzy features changing over time, the present invention provides two q-order cross entropy formulas for hesitant fuzziness, and provides a dynamically changing symptom feature weighting method from the perspective of fuzzy measurement combined with Banzhaf value, which not only takes into account the correlation of dynamically changing symptom features, but also can meet the dynamic calculation of dynamically changing symptom feature weights at different times;

[0071] 2. The present invention can adapt to the dynamic changes of the patient's psychological state and clinical manifestation characteristics in the fuzzy dynamic diagnosis and treatment problem; specifically, since the various psychological states and clinical manifestation characteristics of the patient will change over time, but the traditional fuzzy decision-making diagnosis and treatment method needs to ensure that the characteristics remain unchanged, the present invention can consider dynamic fuzzy diagnosis and treatment during the change of the patient's psychological state and clinical manifestation characteristics, and can ensure the influence of the patient's historical psychological state and clinical manifestation characteristics on the current feature weight, ensuring the fuzzy history feedback mechanism;

[0072] 3. The present invention can take into account the fuzzy dynamic adaptability of dynamically changing treatment plans in the fuzzy dynamic diagnosis and treatment process; specifically, since the diagnosis and treatment process of complex mental illnesses usually requires the coordination of various treatment plans, and the treatment plan at each dynamic moment is not fixed, the present invention adopts a dynamic fuzzy weight calculation method and a dynamic fuzzy preference relationship matrix calculation method to provide a comprehensive formulation method for dynamically changing treatment plans, which not only takes into account the dynamic fuzzy correlation between the patient's psychological state and clinical manifestation characteristics in the symptoms of the dynamic fuzzy diagnosis and treatment process, but also ensures the dynamic changes of the patient's psychological characteristics, clinical manifestation characteristics and treatment plans, as well as the historical feedback capability of the dynamic fuzzy diagnosis and treatment evaluation and the combination of the two. Therefore, the present invention has strong dynamic adaptability and flexible processing of fuzzy information in the fuzzy diagnosis and treatment process of mental illness. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a schematic diagram of the process of the present invention;

[0074] Figure 2 is the bipolar patient in Example 1 at t 1 -t 8 Fuzzy preference matrix of treatment options within a time period. DETAILED DESCRIPTION

[0075] The technical solution of the present invention is described below in conjunction with the accompanying drawings and implementation methods.

[0076] Example

[0077] Since the diagnosis and treatment of mental illness is ambiguous, hesitant and uncertain, the use of precise numerical values ​​to describe the diagnosis and treatment of uncertain mental illness will reduce the effect of diagnosis and treatment. Therefore, the present invention is applied to the diagnosis and treatment of bipolar disorder, and the test results are comprehensively analyzed. The present invention abandons the use of precise numerical values ​​or score values ​​to diagnose and treat patients with mental illness, but describes the psychological state and clinical manifestations of patients with mental illness from the perspective of fuzzy evaluation.

[0078] Bipolar disorder is a complex psychological and mental illness. Its diagnosis and treatment is a complex and long-term process. Assume that a long-term treatment plan is formulated for a bipolar patient. Its symptom characteristics are divided into two parts: mania and depression, which are respectively extremely inflated or arrogant self-esteem (C 1 ), sleep effects (including insomnia and hypersomnia) (C 2 ), abnormal communication desire (C 3 ), quick thinking (C 4 ), distraction (C 5 ), psychomotor agitation or retardation (C 6 ), excessive participation in over-activation (C 7 ), depressed (C 8 ), decreased interest (C 9 ), non-diet weight loss (C 10 ), long-term fatigue (C 11 ), excessive self-blame and guilt (C 12 ), hesitant (C 13 ), self-harm (C 14 ); Long-term treatment for bipolar disorder includes medication and psychotherapy, including lithium salts (T 1 ), lamotrigine (T 2 ), lithium salts and sodium valproate (T 3 ), cognitive behavioral therapy (T 4 ), Family Oriented Therapy (T 5 ), interpersonal and social rhythm therapy (T 6 ), group psychoeducation (T 7 ), Systematic Nursing Management (T 8 ).

[0079] The treatment cycle for patients is one cycle every quarter, and the total treatment duration is two years. The treatment cycle time is expressed as P = {t p |p=1,2,…,8}, a total of eight periods, where the treatment set is T={T i |i=1,2,…,8}, the symptom set is C={C j |j=1,2,…,14}, treatment set T (t) Generate alternative set, symptom set C (t) Generates a property set.

[0080] For each period t, there is a treatment decision matrix in, represents the evaluation of the doctor or patient on their symptoms in each period, and each period t has a treatment set T (t) and symptom set C (t)The two-year treatment regimen evaluation of the patients is shown in Tables 1 to 8. The two-year treatment regimen evaluation of the patients is shown in Tables 1 to 8.

[0081] Table 1 Bipolar treatment decision matrix for patients at t1

[0082]

[0083] Table 2 Bipolar treatment decision matrix for patients at t2

[0084]

[0085] Table 3 Bipolar treatment decision matrix for patients at t3

[0086]

[0087] Table 4 Bipolar treatment decision matrix for patients at T4

[0088]

[0089]

[0090] Table 5 Bipolar treatment decision matrix for patients at t5

[0091]

[0092] Table 6 Bipolar treatment decision matrix for patients at t6

[0093]

[0094] Table 7 Bipolar treatment decision matrix for patients at t7

[0095]

[0096] Table 8 Bipolar treatment decision matrix for patients at t8

[0097] <![CDATA[C 8 ]]> <![CDATA[C 9 ]]> <![CDATA[C 11 ]]> <![CDATA[C 13 ]]> <![CDATA[T 4 ]]> <{0.15}，{0.45，0.1}> <{0.6，0.5}，{0.3}> <{0.25}，{0.6，0.8}> <{0.9}，{0.2，0.3}> <![CDATA[T 5 ]]> <{0.7}，{0.4}> <{0.4}，{0.35，0.7}> <{0.6}，{0.15，0.2}> <{0.15}，{0.6}> <![CDATA[T 7 ]]> <{0.6，0.3}，{0.5}> <{0.55}，{0.2，0.35}> <{0.7}，{0.3，0.4}> <{0.5}，{0.6，0.2}> <![CDATA[T 8 ]]> <{0.5}，{0.2，0.6}> <{0.7}，{0.2}> <{0.9，0.5}，{0.15，0.3}> <{0.7}，{0.3，0.6}>

[0098] like Figure 1 As shown, in this embodiment, t 1 From the initial period to t 2 Taking the dynamic period as an example, the specific application steps of the present invention are described in detail. It should be noted that all evaluation values ​​are 3rd-order hesitant fuzzy numbers, and the cross entropy formula adopts the following formula:

[0099]

[0100] Among them, the parameter s=2.

[0101] (1)t 1 Initial period

[0102] Step 1: Establish t 1 Bipolar Disorder Treatment Decision Matrix As shown in Table 1, we can see that t 1 The treatment set for the period is Symptom set

[0103] Step 2: Calculate t 1 The fuzzy entropy and symptom cross entropy of the bipolar treatment plan decision matrix during the period, where the symptom cross entropy is calculated using the following formula:

[0104]

[0105] Step 3: First calculate t 1 The fuzzy measure of symptoms during the period is calculated as follows:

[0106]

[0107] Substitute the fuzzy entropy of the bipolar treatment decision matrix obtained in step 2 and the cross entropy of the symptoms into the above formula to calculate t 1 The fuzzy measure of period symptoms is μ(C 1 )=0.2813,μ(C 2 )=0.3849,μ(C 4 )=0.4264,μ(C 11 )=0.4178;

[0108] Next, calculate the Banzhaf value of each symptom, and the calculation formula is as follows:

[0109]

[0110] According to the above formula, the Banzhaf value of each symptom is B(C 1 )=0.1741,B(C 2 )=0.2485,B(C 4 )=0.2802,B(C 11 )=0.2735.

[0111] Step 4: Calculate t 1 The static weight of the period symptom set is calculated as follows:

[0112]

[0113] Calculated from the above formula, t 1 Period Symptoms The static weight of Because the historical symptom set is empty in the initial period, the historical symptom set and

[0114] Step 5: From step 4, we know that t 1 Period Symptoms Static weights and historical symptom sets The weight of Substitute the two into the following formula:

[0115]

[0116] From the above formula, we can know that:

[0117]

[0118] Step 6: Substitute the result of step 5 into the following formula to calculate the attribute set at time t1 Dynamic weight

[0119]

[0120] Calculated from the above formula, t 1 Dynamic weights of period symptom sets

[0121] Step 7: Calculate t based on the q-order pair hesitant fuzzy distance formula 1 Period symptoms The relative closeness of each alternative The calculation formula is as follows:

[0122]

[0123] Next, the treatment set ranking for each symptom is given, with C 1 For example, It can be seen that:

[0124]

[0125] Step 8: Calculate the candidate set at time t with respect to attribute c j The static fuzzy preference relation matrix

[0126] First, the relative proximity obtained in step 7 The conversion formula is as follows:

[0127]

[0128] in, represents the q-order hesitant fuzzy distance formula, σ(i) represents the The i-th sequence of j Indicates deviation.

[0129] Then, by Calculate the static fuzzy preference relationship matrix under each attribute as follows:

[0130]

[0131] Step 9: Given step 6, we get t 1 Symptoms of the period Weight Step 8 Get each symptom The static fuzzy preference relation matrix of the treatment set is Calculate t 1 The static fuzzy preference relationship matrix of the period treatment set is calculated as follows:

[0132]

[0133] From the above formula, we can get: 1 The static fuzzy preference relationship matrix of the period treatment set is:

[0134]

[0135] Step 10: Since it is the initial t 1 period, so the dynamic fuzzy preference relationship matrix

[0136] Step 11: According to the dynamic fuzzy preference relationship matrix at time t in step 10, a ranking index is constructed, and its calculation formula is as follows:

[0137]

[0138] Substituting the dynamic fuzzy preference relationship matrix obtained in step 10 into the above formula, we can get: 1 The treatment preferences for the period were

[0139] In summary, the following conclusions are drawn:

[0140] t 1 The sequence of treatment plans during the period is (T 4 , T 2 , T 5 , T 3 ), which means that t 1 The treatment plan during this period was cognitive behavioral therapy (T4 ) is the main drug, lamotrigine (T 2 ) Medication-Assisted Treatment and Family-Oriented Therapy (T 5 ) as a supplement, lithium salt and sodium valproate (T 3 ) Drug therapy is a secondary treatment option.

[0141] (2)t 2 Dynamic period

[0142] Step 1: According to t 2 The decision matrix for bipolar treatment in the period is shown in Table 2. 2 Symptoms of the period Treatment set Historical symptom set

[0143] Step 2: Calculate t 2 The fuzzy entropy of the decision matrix of the treatment plan for bipolar disease during the period and the fuzzy cross entropy of the symptoms, among which the fuzzy cross entropy of the symptoms was calculated using the M-type fuzzy cross entropy.

[0144] Step 3: First, calculate t 2 The fuzzy measure of the symptoms of the period is obtained from the fuzzy entropy of the bipolar treatment decision matrix obtained in step 2 and the fuzzy cross entropy of the symptoms. 2 The fuzzy measure of period symptoms is μ(C 2 )=0.2421,μ(C 3 )=0.3646,μ(C 5 )=0.4062,μ(C 6 )=0.3368,μ(C 11 )=0.5985;

[0145] Continue to calculate the Banzhaf value of each attribute, the calculation formula is as follows:

[0146]

[0147] The Banzhaf values ​​of each symptom calculated by the above formula are B(C 2 )=0.1038,B(C 3 )=0.1664,B(C 5 )=0.1896,B(C 6 )=0.1515,B(C 11 )=0.3119.

[0148] Step 4: Calculate the static weight of the attribute set at time t using the following formula:

[0149]

[0150] Get 2 Period Symptoms The static weight of

[0151] Due to t 2 The process of determining the dynamic weight of period symptoms needs to consider the historical symptom set for dynamic calculation, according to the calculation formula of the weight of the historical attribute set at time t:

[0152]

[0153] Available, t 2 The historical symptom set of the period is The weight of the historical symptom set is

[0154] Step 5: Calculate the attribute set at time t2 With historical attributes Dynamic attribute set weight The calculation formula is as follows:

[0155]

[0156] By calculating the above formula, we can get: 2 period and The dynamic weight of

[0157]

[0158] Step 6: According to the dynamic fuzzy decision, the attributes at time t Dynamic weight The calculation formula is:

[0159]

[0160] By calculating the above formula, we can get: 2 Symptoms of the period The dynamic weight of

[0161] Step 7: Based on the q-order hesitant fuzzy distance formula D γ , t 2 The evaluation value of each attribute at the moment The calculation formula for the relative proximity is:

[0162]

[0163] By calculating the above formula, we can get: 2 Period symptoms The relative closeness of each alternative as follows:

[0164]

[0165] And sort the treatment sets under each symptom according to formula (11), with C 2 For example,

[0166] Step 8: Through the fuzzy preference relationship formula:

[0167]

[0168] in, represents the q-order hesitant fuzzy distance formula, σ(i) represents the The i-th sequence of j Indicates deviation.

[0169] Use with Convert and calculate t 2 Period Treatment Set About Symptom Sets Each symptom The static fuzzy preference relation matrix as follows:

[0170]

[0171] Step 9: Given step 6, we get t 2 Symptoms of the period Dynamic weight and each symptom obtained in step 8 The static fuzzy preference relation matrix of the treatment set is Treatment set according to time t2 The static fuzzy preference relationship matrix of:

[0172]

[0173] From the above calculation, we know that t 2 The static fuzzy preference relationship matrix of the period treatment set is:

[0174]

[0175] Step 10: The static fuzzy preference relationship matrix from step 9 According to the enhanced integration operator formula:

[0176]

[0177] and calculate t 2 At each treatment sequence The dynamic fuzzy preference is calculated as follows:

[0178]

[0179] From the above two formulas, we know that t 2 Treatment episodes of the period Dynamic fuzzy preference relationship matrix As shown below:

[0180]

[0181] Step 11: The dynamic fuzzy preference relationship matrix obtained in step 10 is used to construct a ranking index, and the calculation formula of the ranking index is as follows:

[0182]

[0183] From the above calculation, we know that t 2 The treatment preferences for the period were

[0184] In summary, the following conclusions are drawn:

[0185] t 2 The sequence of treatment plans during the period is (T 2 , T 3 , T 8 , T 1 , T 6 ), which means that t 2 Lamotrigine (T 2 ) Drug treatment is mainly lithium salt and sodium valproate (T 3 ) and lithium salt (T 1 ) drug therapy combined with systematic nursing management (T 8 ) as adjunctive therapy, interpersonal and social rhythm therapy (T 6 ) is a secondary treatment option.

[0186] Then, with t 2 The calculation steps of the treatment plan for the period are the same. 3 to 8 The same treatment options were selected during the same period.

[0187] First, the dynamic weights of symptoms for assessing patients' bipolar disorder during the two-year period were calculated as shown in Table 9.

[0188] Table 9 Dynamic symptom weights of bipolar patients during two years of treatment

[0189] <![CDATA[t 1 ]]> <![CDATA[t 2 ]]> <![CDATA[t 3 ]]> <![CDATA[t 4 ]]> <![CDATA[t 5 ]]> <![CDATA[t 6 ]]> <![CDATA[t 7 ]]> <![CDATA[t 8 ]]> <![CDATA[C 1 ]]> 0.1783 - - - - 0.2809 - - <![CDATA[C 2 ]]> 0.2545 0.2007 - - 0.1995 - 0.2309 - <![CDATA[C 3 ]]> - 0.1365 - - - - - - <![CDATA[C 4 ]]> 0.2870 - - - - - - - <![CDATA[C 5 ]]> - 0.1555 0.1858 - - - - - <![CDATA[C 6 ]]> - 0.1242 - 0.1391 - - 0.1064 - <![CDATA[C 7 ]]> - - - 0.0923 - - - - <![CDATA[C 8 ]]> - - 0.1521 0.2081 0.1425 - 0.1682 0.1893 <![CDATA[C 9 ]]> - - 0.1578 0.3075 0.1224 - - 0.1829 <![CDATA[C 10 ]]> - - 0.1493 0.1223 0.0895 - 0.1134 - <![CDATA[C 11 ]]> 0.2801 0.3831 - - 0.2600 - 0.3040 0.4638 <![CDATA[C 12 ]]> - - 0.1886 0.1306 0.0941 0.3096 - - <![CDATA[C 13 ]]> - - - - - 0.1423 0.0771 0.1640 <![CDATA[C 14 ]]> - - 0.1666 - 0.0920 0.2671 - -

[0190] As can be seen from Table 9, since the patient suffered from bipolar disease for two years, and the diagnostic process may take into account the patient's condition over a period of time in the past, there are dynamic differences in the emphasis of the patient's bipolar symptom assessment at different times.

[0191] For example, in t 1 Period, the impact on patients' sleep (C 2 ), Quick Thinking (C 4 ) and chronic fatigue (C 11 ) as the main focus; 2 The main focus of the period is on sleep effects (C 2 ) and chronic fatigue (C 11 );t 3 The main concern during this period is the patient's distraction (C 5 ), excessive self-blame and guilt (C 12 ) situation; at t 4 The patient's interest decreases during this period (C 9 ) Highest attention; 5 to 8 During this period, patients experience mild mania and severe depression, with a focus on long-term fatigue (C 11 ) and excessive self-blame and guilt (C 12 ).

[0192] Next, a treatment plan for the patient's bipolar disorder over a two-year period is given.

[0193] According to the calculation, the fuzzy preference matrix of the patient's dynamic treatment plan for two years is as follows: Figure 2 As shown, Figure 2 (a) in the equation is t 1 Period treatment plan preference relationship matrix, Figure 2 (b) in the equation is t 2 Period treatment plan preference relationship matrix, Figure 2 (c) in the equation is t 3 Period treatment plan preference relationship matrix, Figure 2 (d)t 4 Period treatment plan preference relationship matrix, Figure 2 (e) in the equation is t 5 Period treatment plan preference relationship matrix, Figure 2 (f) in the equation is t 6 Period treatment plan preference relationship matrix, Figure 2 (g) in the equation is t 7 Period treatment plan preference relationship matrix, Figure 2 (h) in is t 8 Period treatment plan preference relationship matrix; according to Figure 2 You can clearly understand the focus of each period on the treatment plan for bipolar disorder patients:1 For example, cognitive behavioral therapy (T 4 ) compared with lamotrigine (T 2 ) and lithium and sodium valproate (T 3 ) and family-oriented therapy (T 5 ) have a greater preference, but cognitive behavioral therapy (T 4 ) therapy, but rather focuses more on cognitive behavioral therapy (T 4 ).

[0194] The ranking index results for each period are shown in Table 10.

[0195] Table 10 Ranking of dynamic treatment options for bipolar patients during a two-year period

[0196]

[0197]

[0198] From Table 10, we can see that:

[0199] t 1 Cognitive behavioral therapy (CBT) 4 ) is the main drug, lamotrigine (T 2 ) Medication-Assisted Treatment and Family-Oriented Therapy (T 5 ) as a supplement, lithium salt and sodium valproate (T 3 ) Drug therapy is a secondary treatment option;

[0200] t 2 Lamotrigine (T 2 ) Drug treatment is mainly lithium salt and sodium valproate (T 3 ) and lithium salt (T 1 ) drug therapy combined with systematic nursing management (T 8 ) as adjunctive therapy, interpersonal and social rhythm therapy (T 6 ) is a secondary treatment option;

[0201] t 3 Systematic nursing management (T 8 ) and group psychoeducation (T 7 ) is the main treatment option, family-oriented therapy (T 5 ) and cognitive behavioral therapy (T 4 ) as an auxiliary treatment program, interpersonal and social rhythm therapy (T 6 ) is a secondary treatment option;

[0202] t 4 Cognitive behavioral therapy (CBT) 4 ) as the main treatment plan, with group psychoeducation (T7 ) and lithium and sodium valproate (T 3 ) is assisted by drug therapy and systematic nursing management (T 8 ) is a secondary treatment option;

[0203] t 5 Family-Oriented Therapy (T 5 ) and interpersonal and social rhythm therapy (T 6 ) as the main treatment program and supplemented by group psychoeducation (T 7 ), lithium salt and sodium valproate (T 3 ) Drug treatment and cognitive behavioral therapy (T 4 ) supplemented by systematic nursing management (T 8 ) is a secondary treatment option;

[0204] t 6 Lithium and sodium valproate (T 3 ) Drug therapy supplemented by systematic nursing management (T 8 ) is the main treatment option, family-oriented therapy (T 5 ) and lamotrigine (T 2 )Lithium salt (T 1 ) combined with cognitive behavioral therapy (T 4 ) is a secondary treatment option;

[0205] t 7 During this period, lithium salts and sodium valproate (T 3 ) of drug treatment, group psychoeducation (T 7 ) as an aid, combined with cognitive behavioral therapy (T 4 ) and system care management (T 8 ), interpersonal and social rhythm therapy (T 6 ) is a secondary treatment option;

[0206] t 8 Group Psychological Education (T 7 ) as the main treatment option, supplemented by family-oriented therapy (T 5 ) and cognitive behavioral therapy (T 4 ), with system nursing management (T 8 ) is a secondary treatment option.

[0207] Finally, this embodiment is compared and analyzed with other three prior arts from three aspects: dynamic decision-making, static decision-making and differently weighted decision-making.

[0208] 1. In terms of dynamic decision-making, this embodiment is compared and analyzed with the prior art 1 "a dynamic group decision-making method based on AQM (referred to as AQM method)".

[0209] Prior art 1 converts the dynamic fuzzy preference relationship matrix into a 0-1 matrix according to the idea of ​​AQM, and then makes a decision sorting. However, in a complex fuzzy background, it is not possible to only consider the binary preference for the alternative plan, even if this binary preference is converted from the fuzzy preference. Table 11 shows the comparison of the dynamic treatment plan using the AQM method and the fuzzy AQM method used in this embodiment. Prior art 1 converts the fuzzy preference matrix into a 0-1 preference relationship matrix, which causes the fuzzy preference information to be lost in the final plan sorting, which makes it impossible to fit the patient's symptoms well when the treatment plan is formulated in some periods.

[0210] Table 11 Comparison of the AQM method and the dynamic treatment scheme of the present invention

[0211]

[0212]

[0213] 2. In terms of static decision-making, the present invention is compared and analyzed with the static fuzzy treatment plan of each period (i.e., the decision-making method based on static weighting of Banzhaf index).

[0214] The comparison results are shown in Table 12. There are some differences between the static treatment plan and the dynamic treatment plan in each period. The reason is that the static treatment plan does not have dynamic weight feedback and dynamic fuzzy preference relationship feedback.

[0215] Table 12 Comparative analysis of static treatment regimen and dynamic treatment regimen for patients in each period

[0216] Static treatment plan Dynamic treatment plan <![CDATA[t 1 ]]> <![CDATA[T 4 >T 2 >T 5 >T 3 ]]> <![CDATA[T 4 >T 2 >T 5 >T 3 ]]> <![CDATA[t 2 ]]> <![CDATA[T3 3 >T 2 >T 8 ≥T 1 >T 6 ]]> <![CDATA[T 2 >T 3 ≥T 8 ≥T 1 >T 6 ]]> <![CDATA[t 3 ]]> <![CDATA[T 7 >T 5 ≥T 4 ≥T 8 >T 6 ]]> <![CDATA[T 8 >T 7 ≥T 5 ≥T 4 >T 6 ]]> <![CDATA[t 4 ]]> <![CDATA[T 4 >T 3 >T 7 >T 8 ]]> <![CDATA[T 4 >T 7 ≥T 3 >T 8 ]]> <![CDATA[t 5 ]]> <![CDATA[T 5 >T 6 >T 7 >T 3 >T 8 >T 4 ]]> <![CDATA[T 5 >T 6 >T 7 >T 3 ≥T 4 >T 8 ]]> <![CDATA[t 6 ]]> <![CDATA[T 8 >T 3 >T 2 >T 5 >T 1 >T 4 ]]> <![CDATA[T 3 >T 8 >T 5 >T 2 >T 1 >T 4 ]]> <![CDATA[t 7 ]]> <![CDATA[T 4 ≥T 7 >T 3 >T 8 >T 6 ]]> <![CDATA[T 3 >T 7 >T 4 ≥T 8 >T 6 ]]> <![CDATA[t 8 ]]> <![CDATA[T 7 >T 4 >T 5 >T 8 ]]> <![CDATA[T 7 >T 5 ≥T 4 >T 8 ]]>

[0217] In t 1 Period and t 2 During the period, the treatment scheme of this embodiment is basically the same as that of the static fuzzy treatment method; 3 period, since this embodiment takes into account the historical impact, the system care management (T 8 ) need to be given priority, but in static treatment programs, since only the symptom description of the current period is considered, the main focus is on group psychoeducation (T 7 );t 4 With t 5 The static and dynamic treatment plans are similar; t 6 The treatment plan provided in this embodiment is mainly based on system nursing management (T 8 ) combined with sodium valproate (T 3 ) Drug treatment; 7 The static treatment program during the period is different from the treatment program provided in this embodiment. The static treatment program mainly uses cognitive behavioral therapy (T 4 ) and group psychoeducation (T7 ), and in this embodiment, sodium valproate (T 3 ) Drug treatment and group psychoeducation (T 7 ) is the main treatment option.

[0218] 3. In terms of different weighted decisions, the present invention is compared and analyzed with the prior art 2 "objective weighting method combining basic probability distribution vector and belief entropy".

[0219] Prior art 2 provides a dynamic feedback decision framework from the three aspects of time, information volume and information consistency of each period, but cannot handle dynamic decision-making with dynamic changes of multiple attributes and multiple alternatives, and cannot effectively handle hesitant fuzzy information. The results of the comparative analysis between the present invention and prior art 2 are shown in Table 13.

[0220] Table 13 Comparative analysis of the proposed dynamic decision-making method and the existing technology 2

[0221] Treatment plan of prior art 2 The treatment plan of this embodiment <![CDATA[t 1 ]]> <![CDATA[T 4 >T 2 >T 3 >T 5 ]]> <![CDATA[T 4 >T 2 >T 5 >T 3 ]]> <![CDATA[t 2 ]]> <![CDATA[T 2 >T 3 >T 1 >T 8 >T 6 ]]> <![CDATA[T 2 >T 3 ≥T 8 ≥T 1 >T 6 ]]> <![CDATA[t 3 ]]> <![CDATA[T 7 >T 4 =T 6 =T 8 >T 5 ]]> <![CDATA[T 8 >T 7 ≥T 5 ≥T 4 >T 6 ]]> <![CDATA[t 4 ]]> <![CDATA[T 7 >T 3 >T 8 >T 4 ]]> <![CDATA[T 4 >T 7 ≥T 3 >T 8 ]]> <![CDATA[t 5 ]]> <![CDATA[T 7 >T 6 >T 5 >T 3 >T 4 >T 8 ]]> <![CDATA[T 5 >T 8 >T 7 >T 3 ≥T 4 >T 8 ]]> <![CDATA[t 6 ]]> <![CDATA[T 8 >T 4 >T 3 >T 2 >T 5 >T 1 ]]> <![CDATA[T 3 >T 8 >T 5 >T 2 >T 1 >T 4 ]]> <![CDATA[t 7 ]]> <![CDATA[T 7 >T 3 >T 4 >T 8 >T 8 ]]> <![CDATA[T 3 >T 7 >T 4 ≥T 8 >T 6 ]]> <![CDATA[t 8 ]]> <![CDATA[T 7 >T 4 =T 5 =T 8 ]]> <![CDATA[T 7 >T 5 ≥T 4 >T 8 ]]>

[0222] It can be seen from Table 13 that the prior art 2 and the present embodiment have different 1 to 4 The treatment plans formulated during the period were basically the same. 5 ,t 6 and t 7 The period difference is large. 8 The treatment plans for the two periods are basically the same; the main reason for the difference between the two is that the patient's symptoms have been changing with time, but the dynamic correlation of the changing symptoms has not been considered by the prior art 2.

[0223] The advantages of this embodiment over the prior arts 1 and 2 are shown in Table 14.

[0224] Table 14 Comparative analysis of advantages of prior art 1, prior art 2 and the present invention

[0225]

[0226]

[0227] In summary, the present embodiment can be more flexible in dealing with uncertain and hesitant dynamic decision-making problems. Since the present embodiment takes into account the correlation of dynamic attributes at each moment, it is more flexible and effective in dealing with decision-making problems when some dynamic attributes are not independent. Secondly, the present embodiment adapts to the changes of multiple alternatives and multiple attributes in the dynamic decision-making process, so it has stronger adaptability and dynamic processing capabilities than existing methods. Finally, the present embodiment retains the impact of historical decisions on the current decision when dealing with dynamic decision-making problems, that is, the historical feedback mechanism, including the historical feedback mechanism of dynamically changing attribute weights and the dynamically changing multiple alternative historical feedback mechanism, making the dynamic decision-making method more reasonable and effective.

Claims

1. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness, characterized in that: Including dynamic weighting and dynamic decision-making; The dynamic weighting includes the following steps: Step 1: Establish the q-order hesitant fuzzy decision matrix D at time t (t) ; The decision matrix D (t) It includes the membership set and non-membership set of the evaluation value of a candidate i under attribute j at time t; among which: Time t is the treatment cycle; the alternative set is generated by the treatment set, and alternative i refers to the i-th treatment plan; the symptom set is generated by the attribute set, and attribute j refers to the j-th disease symptom; Step 2: Through the decision matrix D in step 1 (t) Calculate the fuzzy entropy matrix at time t, and then calculate the cross entropy between every two attributes using the cross entropy formula; Step 3: Calculate the fuzzy measure and Banzhaf value of each attribute at time t; Step 4: Calculate the static weight of the attribute set at time t and the weight of the historical attribute set at time t; Step 5: Calculate the attribute set C at time t (t) With historical attribute set Dynamic attribute set weights Step 6: Calculate the attribute set C at time t (t) The dynamic weight W (t) ; The dynamic decision-making includes the following steps: Step 7: Calculate the relative closeness of the fuzzy evaluation of each attribute at time t and sort them; Step 8: Calculate the candidate set at time t with respect to attribute C j The static fuzzy preference relation matrix Step 9: The dynamic weight W at time t obtained in step 6 (t) And step 8 obtains the static fuzzy preference relationship matrix for each attribute Integrate the static fuzzy preference relationship matrix under each attribute to obtain the integrated static fuzzy preference relationship matrix at time t Step 10: The integrated static fuzzy preference relationship matrix obtained in step 9 Calculate the dynamic fuzzy preference relationship matrix at time t Step 11: According to the dynamic fuzzy preference relationship matrix at time t in step 10 Constructing ranking indicators; In step 2, the hesitant fuzzy number of any two q-order pairs is ξ1=<u1,v1> ,ξ2=<u2,v2> , calculate the two cross entropy formulas of the fuzzy numbers ξ1 and ξ2 M-type fuzzy cross entropy CE M and N-type fuzzy cross entropy CE N as follows: Wherein, represents the number of membership degrees of ξ1; represents the number of non - membership degrees of ξ1; represents the number of membership degrees of ξ2; represents the number of non - membership degrees of ξ2; The parameter s satisfies 1 < s ≤ 2; The parameter p satisfies p > 0; u1 and v1 respectively represent the membership degree set and non - membership degree set of the fuzzy number ξ1; u2 and v2 respectively represent the membership degree set and non - membership degree set of the fuzzy number ξ2; θ1 represents the membership degree element of the membership degree set u1 of the fuzzy number ξ1; θ2 represents the membership degree element of the membership degree set u2 of the fuzzy number ξ2; represents the non - membership degree element of the non - membership degree set v1 of the q - order paired fuzzy number ξ1; represents the non - membership degree element of the non - membership degree set v2 of the q - order paired fuzzy number ξ2; π1 represents the hesitancy degree of the q - order paired fuzzy number ξ1; π2 represents the hesitancy degree of the q - order paired fuzzy number ξ2; The formula for calculating the fuzzy measure of each attribute at time t in step 3 is as follows: In the formula, μ represents the fuzzy measure of the attribute set; c j represents attribute decision; E(ξ ij ) represents the q-order hesitant fuzzy number ξ ij Fuzzy entropy of CE(ξ ij ,ξ sj ) represents the q-order hesitant fuzzy cross entropy; m represents the number of alternatives in the alternative set; The calculation formula of each attribute Banzhaf value is as follows: In the formula, B represents the Banzhaf function, S represents the subset of the attribute set C; The calculation formula of the static weight of the attribute set at time t in step 4 is as follows: In the formula, Represents the static weight of the attribute at time t; The calculation formula of the historical attribute set weight at time t is as follows: In the formula, B h Represents the historical attributes at time t Banzhaf value; represents the historical attribute set at time t; h represents the historical attribute subscript. If it is the initial time, then According to the attribute set C at time t in step 5 (t) ={c1,c2,…,c n }'s static weights and historical attribute sets The dynamic weight of the attribute set at time t is obtained by The calculation formula is as follows: In the formula, F IOWA Represents the weight of the historical attribute set at time t and static weight of attribute set at time t The induced ordered weighted average function; Q(x) = (-x 2 +3x) / 2 represents a regular increasing monotonic quantization function; q represents a dynamic time variable; Represents the union of the historical attribute set and the current attribute set at time i of the dynamic time variable q The weight of Represents the weight vector at each moment i, that is, the degree of attention that the dynamic decision makes to moment i.

2. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness according to claim 1, characterized in that: The q-order hesitant fuzzy decision matrix D at time t in step 1 (t) for: Where t∈T represents the moment of the dynamic moment set T, It represents the hesitant fuzzy number evaluation value of the q-order pair for the i alternative under the j attribute at time t; It represents the membership set of the evaluation value of a candidate i under attribute j at time t; It represents the non-membership set of the evaluation value of a candidate i under attribute j at time t.

3. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness according to claim 1, characterized in that: In step 1, the attribute set C at time t (t) The dynamic weight W (t) The calculation formula is as follows: In the formula, Indicates that the dynamic attribute set at time t contains only the current attribute set C at time t (t) Dynamic weight.

4. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness according to claim 1, characterized in that: The calculation formula of the relative closeness of the fuzzy evaluation under each attribute at time t in step 7 is as follows: In the formula, RC j Represents the fuzzy evaluation value at time t for the relative closeness of positive and negative ideal solutions; Represents the fuzzy evaluation value With positive ideal evaluation value The distance between Represents the fuzzy evaluation value With negative ideal evaluation value The distance between According to the relative proximity RC j The options under each attribute are sorted, and the sorting formula is as follows: In the formula, A j Indicates that in attribute c j The alternative set under ; σ(i) represents the sorting order, which satisfies 5. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness according to claim 1, characterized in that: In step 8, the candidate set at time t is about attribute c j Integrated static fuzzy preference relationship matrix The calculation formula is as follows: In the formula, The static fuzzy preference relationship conversion formula for the sth candidate to the kth candidate is expressed as: In the formula, s and k are both positive real numbers greater than 0; r max The relative closeness RC calculated by expression (10) j The maximum value of r(a s ) and r(a k ) represent alternative a s and alternative a k fuzzy preference value.

6. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness according to claim 1, characterized in that: The calculation formula for the integrated static fuzzy preference in step 9 is as follows: In the formula, represents the attribute set at time t; Represents the integrated static fuzzy preference relationship between alternatives s and k at time t.

7. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness according to claim 1, characterized in that: The dynamic fuzzy preference relationship matrix at time t in step 10 The calculation formula is as follows: In the formula, represents the dynamic fuzzy preference for alternative s and alternative k at time t, which is expressed as: Where D E represents an enhanced integration operator.

8. A dynamic q-order hesitant fuzzy decision-making method for diagnosis and treatment of mental illness according to claim 1, characterized in that: The calculation formula of the ranking index in step 11 is as follows: In the formula, m represents the number of alternatives in the alternative set; I (t) (a i ) indicates that other alternatives at time t tend to favor alternative a i Preference value; O (t) (a i ) represents the alternative a at time t i The preference value towards other alternatives; specifically, V (t) (a i ) is larger, indicating a is more preferred. i .

Images