Example 1 Clinical Example
 1 Data source: From January 2009 to January 2012, cases recorded in the Dermatology Outpatient and Inpatient Department of Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine, a total of 202 cases, including 127 males and 75 females, age Between the ages of 17-79, the average age is 42.70±14.65 years, and the course of disease is 4-360 months, with an average of 113.60±94.07 months. All cases were diagnosed with psoriasis vulgaris at the time of treatment.
 2.1 Diagnostic criteria: All diagnosed as psoriasis vulgaris (according to "Clinical Disease Diagnosis Basis for Cure and Improvement Criteria", edited by Sun Chuanxing, Beijing People's Military Medical Publishing House).
 2.2 Exclusion criteria:
 ①Have a history of fever of unknown origin within 2 months;
 ②Throughout systemic treatment within 3 months, including hormones and immunosuppressive agents;
 ③A large area of strong corticosteroids have been used within 3 months;
 ④Other types of psoriasis;
 ⑤ Pregnant or lactating women;
 ⑥ Patients with severe primary diseases such as cardio-cerebrovascular, liver and kidney damage and mental illness.
 2.3 Treatment: TCM treatment based on syndrome differentiation-clearing away heat and cooling blood (the main prescription for blood-heat syndrome and the addition and subtraction of medicines depending on the symptoms).
 2.4 Evaluation of curative effect: According to the "Guiding Principles for Clinical Research of New Chinese Medicines (Trial)", the evaluation criteria for the curative effect of psoriasis vulgaris (National Administration of Traditional Chinese Medicine. Standards for Diagnosis and Curative Effect of Chinese Medicine Diseases[M]. Nanjing. Nanjing University Press. 1994:154-155), the curative effect is divided into 4 levels: "cure", "significantly effective", "improved" and "ineffective".
 2 methods:
 2.1 Establish a four-element connection
 Taking psoriasis symptomatic addition and subtraction medication selection as an example, establish the quaternary connection number of curative effect of different symptoms: μ=a+bi+cj+dk, a corresponds to the percentage of recovery after symptomatic medication, and b corresponds to obvious effect after symptomatic medication The percentage of c corresponds to the percentage of improvement after symptomatic medication, and d corresponds to the percentage of ineffectiveness after symptomatic medication. Among them, the value interval of i, j, and k is defined according to the clinical significance. We usually set improvement as positive development. The definition of i takes a value in the interval [0,1], which means that "marked effect" has a certain range of "cure" and Uncertainty in the development of "improvement"; j takes a value in the interval of [-1,0], indicating the uncertainty of the development of "improvement" to "significant" and "ineffective"; k = -1 means ineffective (or worse ).
 2.2 Calculate the first and third order partial connection numbers
 According to the need to calculate the effective connection value, at this time i∈[0,1], i is 0.5, j∈[0,-1], j is -0.5.
 First-order positive connection number
 Third-order positive connection number
 Similarly, calculate the first-order and third-order partial negative connection numbers, according to Calculate the corresponding first-order and third-order full partial connection numbers.
 3 results:
 Table 1 Distribution of accompanying symptoms and tongue coating pulse conditions in patients with psoriasis in each treatment outcome (cases)
 Table 2 The number of quaternary links of curative effect
 Table 3 First-order and third-order total partial connection numbers
 2.4 Discussion:
 In Table 1 and Table 2, the distribution of symptoms and tongue coating pulse conditions of patients with psoriasis in each treatment outcome (example) and the quaternary connection number are listed, and the first-order and third-order fully partial connection numbers are calculated respectively. In Table 3. Taking the reduction of sweating as an example, we can see that whether it is calculating the first or third order, the total partial connection number is all negative. Theoretically, the development trend of symptomatic treatment for the reduction of sweating is a negative development, and we Observed results of clinical efficacy are mostly positive development, indicating that the theoretical results are inconsistent with the actual clinical observations. According to the theory of connection mathematics, only when a, b, c, d of the quaternary connection number μ=a+bi+cj+dk satisfy a>(b+c+d), the total partial connection number is positive. But the clinical situation obviously cannot always meet this condition.
 Therefore, we try to improve the partial connection number method of curative effect analysis based on the practical significance of medicine. For example, for the quaternary connection number, only the first-order fully partial connection number that directly reflects the current treatment measures and the more clinically meaningful first-order total partial connection number is calculated and ranked in i ∈ [0.5, 1], j ∈ [- 0.5, 0] Inner pair i, j are taken to make the first-order total partial connection number less than 0 (in accordance with clinical practice). The adjustment results are as follows:
 Table 4 Adjust i, j to define the interval based on the first-order total partial connection number ranking
 From the result observation, the value interval of i and j has been in line with the clinical situation under the sample size. According to the theory of set pair analysis, in the past, the third-order partial connection number method was often used to calculate the quaternion connection number to eliminate the uncertain factors i, j to obtain a real number. However, the real number is a measure reflecting the trend of contradictory movement at the micro level of the internal relations of "cure", "significant effect", "improvement" and "ineffectiveness", and cannot fully satisfy the measurement of the movement trend of each curative effect level. And only the real numbers are kept, which is also different from the many uncertain situations included in the clinic. Regarding the value of i and j, when the value and ranking reflected by the total partial connection number do not match the clinical value, sampling error should be considered first. Assuming that 100 drugs are evaluated for the symptomatic treatment of psoriasis, the results of the third-order total partial connection number of more than 80 drugs are in line with clinical reality, and the others are somewhat different. We believe that the third-order total partial connection number should be accepted and the sample size should be expanded. Second calculation. Conversely, if there are 100 drugs, of which there are more than 80 drugs, the results of the third-order total partial connection number are different from the clinical results, then use the method described in this study to only calculate the change trend that directly reflects the drug efficacy under the current treatment measures, and The more clinically meaningful first-order total partial connection number, and then according to its ranking in i∈[0.5,1], j∈[-0.5,0], the values of i, j are further adjusted to make the first-order total partial connection The numbers are in line with clinical reality. In the subsequent evaluation of the symptomatic treatment of psoriasis, each drug selects the corrected i, j value, and further expands the sample size and repeats the calculation to obtain the exact value range of i, j.