Method for constructing a risk prediction model for oral mucositis care

Abstract

The present invention relates to the field of data processing technology, and more specifically to a method for constructing a risk prediction model for oral mucositis care, comprising: obtaining the data of all patients who have visited the hospital; obtaining the degree of influence of each risk factor on the assessed risk level based on each risk factor of patients at all risk levels, and screening out the key risk factors for each risk level; obtaining the importance of each risk factor based on the degree of influence of each risk factor on all assessed risk levels, and screening out all important risk factors; and constructing an oral mucositis risk prediction model based on the key risk factors and important risk factors for each risk level. The present invention constructs an oral mucositis risk prediction model by analyzing historical patient data, thereby accurately predicting the severity of oral mucositis in patients after chemotherapy and providing corresponding prevention and care for the patients.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for constructing a risk prediction model for oral mucositis care. Background Art

[0002] Oral mucositis, an ulcerative lesion of the oral mucosa, is a common oral complication associated with cancer treatment. Clinically, it manifests as redness, erosion, ulceration, and fibrosis of the oral mucosa. Patients experience discomfort such as dysphagia, dry mouth, and taste disturbances. Furthermore, impaired nutritional status can lead to bacterial, viral, or fungal infections. The inability to accurately predict the progression of a patient's condition can directly or indirectly lead to additional medical costs during care, reducing their quality of life.

[0003] Since the beginning of this year, computer technology has rapidly developed, leading to the rapid adoption and application of intelligent analysis algorithms in the medical field. Artificial intelligence technology is assisting clinical nurses in making more accurate predictions and decisions. However, due to the high incidence of oral mucositis, the large sample size of clinical data, and the complex structure of the data, existing machine learning algorithms struggle to accurately make correct decisions. Summary of the Invention

[0004] The present invention provides a method for constructing a risk prediction model for oral mucositis care to solve the existing problem: it is difficult to accurately predict the oral mucositis grade of patients after chemotherapy based on various data of the patients, resulting in the inability to provide effective care plans for patients.

[0005] The method for constructing a risk prediction model for oral mucositis care of the present invention adopts the following technical solutions:

[0006] The following steps are involved:

[0007] Obtaining information on all patients who visit the hospital, including the patient's risk level and risk factors;

[0008] Marking the patients according to the presence of each risk factor in the patients at each risk level, and obtaining a patient marker sequence for each risk factor at each risk level; obtaining the correlation between the patient marker sequences of different risk factors at the same risk level; screening out the related risk factors of the risk factors at each risk level based on the correlation; obtaining the degree of influence of the risk factors on the assessed risk level based on the correlation between different risk factors at the same risk level, combined with the difference between the risk factors at the risk level and the related risk factors of the risk factors; screening out the key risk factors for each risk level based on the degree of influence of the risk factors on the assessed risk level;

[0009] According to the impact of each risk factor on all assessed risk levels and the number of patients with the corresponding risk factor in the patient data, the importance of each risk factor is obtained; according to the importance of each risk factor, all important risk factors are screened out;

[0010] An oral mucositis risk prediction model is trained based on the key risk factors and important risk factors of each risk level; based on the prediction model, the oral mucositis level of the patient after chemotherapy is predicted, and care is provided to the patient.

[0011] Preferably, the patient is marked according to the presence of each risk factor in the patient at each risk level, and a patient marking sequence for each risk factor at each risk level is obtained, including the following specific methods:

[0012] For any risk factor, among all patients at each risk level, the patients with the risk factor are marked as , patients without the risk factors were labeled ; Obtain a patient tag set corresponding to the risk factors at each risk level, sort the elements in the patient tag set corresponding to the risk factors at each risk level according to the time of information entry, and obtain a patient tag sequence corresponding to the risk factors at each risk level; and are preset non-negative marker values, and Greater than .

[0013] Preferably, the tag value 1, mark value is 0.

[0014] Preferably, the method of obtaining the correlation between patient marker sequences of different risk factors at the same risk level includes:

[0015] Obtaining the difference between the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level and the sum of the sequence marker values; obtaining the correlation between the patient marker sequences of different risk factors at the same risk level based on the difference between the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level and the sum of the sequence marker values;

[0016] The absolute value of the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level is positively correlated with the correlation between patient marker sequences of different risk factors at the same risk level; the difference in the sum of marker values ​​of patient marker sequences of different risk factors at the same risk level is negatively correlated with the correlation between patient marker sequences of different risk factors at the same risk level.

[0017] Preferably, the method of screening out the related risk factors of the risk factors at each risk level based on the correlation includes:

[0018] Preset a threshold for the number of relevant risk factors , among the correlations between patient marker sequences of different risk factors at the same risk level, the one with the largest correlation risk factors as related risk factors of the risk factors under this risk level.

[0019] Preferably, the method of obtaining the degree of influence of risk factors on the assessed risk level based on the correlation between different risk factors at the same risk level and the difference between the risk factors at the risk level and the risk factors related to the risk factors includes the following specific methods:

[0020] For the degree of influence of any risk factor on the assessment of any risk level, obtain the difference between the risk factor and the related risk factors of the risk factor at the same risk level;

[0021] The correlation between different risk factors at the same risk level is positively correlated with the degree of influence of risk factors on the assessed risk level; the difference between risk factors and related risk factors at the same risk level is positively correlated or negatively correlated with the degree of influence of risk factors on the assessed risk level.

[0022] Preferably, the method of screening out key risk factors for each risk level according to the degree of influence of risk factors on the assessed risk level includes the following specific methods:

[0023] Preset the risk factor impact threshold for a single risk , when the impact of any risk factor on the assessment of any risk level is greater than When the risk factor is ≥ 1, the risk factor is recorded as the key risk factor for assessing the risk level.

[0024] Preferably, the importance of each risk factor is obtained based on the degree of influence of each risk factor on all assessed risk levels, in combination with the number of patients with the corresponding risk factor in the patient data, including the following specific methods:

[0025] The degree of influence of each risk factor on the assessment of all risk levels is positively correlated with the importance of each risk factor; the number of patients with each risk factor is positively correlated with the importance of each risk factor; the difference in the degree of influence of each risk factor on the assessment of different risk levels is negatively correlated with the importance of each risk factor.

[0026] Preferably, the method of screening out all important risk factors according to the importance of each risk factor includes:

[0027] Preset an importance threshold When the importance of risk factors is greater than When the risk factor is significant, it is recorded as a significant risk factor.

[0028] Preferably, the training of the oral mucositis risk prediction model includes the following specific methods:

[0029] The set of key risk factors of all risk levels is recorded as the key factor set, and the set of all important risk factors is recorded as the important factor set; the risk factors in the union of the key factor set and the important factor set are recorded as the target risk factors;

[0030] The set consisting of all target risk factors is used as the first random forest algorithm training set, and the set consisting of all non-target risk factors is used as the second random forest algorithm training set; preset an adjustment weight greater than 0.5 and less than 1 ;

[0031] Let the random forest algorithm be The probability of randomly selecting risk factors from the No. 1 random forest algorithm training set is The probability of randomly selecting risk factors from the No. 2 random forest algorithm training set is used to train the oral mucositis risk prediction model.

[0032] The beneficial effects of the technical solution of the present invention are as follows: the present invention obtains the correlation between different risk factors at the same risk level based on each risk factor of patients at all risk levels; based on the correlation between different risk factors at the same risk level, the key risk factors of each risk level are screened out. Since the greater the impact of a risk factor on the assessed risk level, the more similar the risk factor is to other risk factors, the degree of impact of the risk factor on the assessed risk level can be obtained through the similarity between different risk factors, thereby screening out the key risk factors of each risk level, thereby preparing for the subsequent construction of an oral mucositis nursing knowledge base.

[0033] Furthermore, all important risk factors are screened out according to the degree of influence of each risk factor on all assessed risk levels; since the key risk factors of each risk level are indicators for a risk level, the important risk factors are obtained by integrating the degree of influence of all risk factors on all assessed risk levels; further, according to the key risk factors and important risk factors of each risk level, a training set of the random forest algorithm is constructed to train the oral mucositis risk prediction model, so that the oral mucositis risk prediction model can accurately evaluate the patient's risk level, and then provide the patient with an effective care plan, so as to achieve the purpose of reducing the incidence of oral mucositis complications and alleviating the patient's suffering. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0035] Figure 1 A flowchart of the steps of the method for constructing a risk prediction model for oral mucositis care according to the present invention;

[0036] Figure 2 Flowchart for obtaining the oral mucositis risk prediction model. DETAILED DESCRIPTION

[0037] To further illustrate the technical means and effectiveness of the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the method for constructing a risk prediction model for oral mucositis care according to the present invention, including its specific implementation, structure, features, and effectiveness. In the following description, references to different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0038] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0039] The specific scheme of the method for constructing a risk prediction model for oral mucositis care provided by the present invention is described in detail below with reference to the accompanying drawings.

[0040] Example 1:

[0041] Step S1: Collect the patient's oral mucositis risk data.

[0042] The oral mucositis risk factors of the current patient are collected by doctors and nurses, and the risk factors are shown in the following table; at the same time, the oral mucositis risk factors and risk levels of all historical patients who have visited the database are retrieved.

[0043]

[0044] Step S2: predicting the oral mucositis grade of the current patient after chemotherapy based on the oral mucositis risk information of the current patient.

[0045] It should be noted that currently, the severity of oral mucositis after chemotherapy is mainly predicted by a scale before the patient undergoes chemotherapy. However, the induction and influence of the occurrence of oral mucositis involve many complex factors, and the accuracy and applicability of the scale in predicting the severity of oral mucositis after chemotherapy are limited. Therefore, this embodiment uses various historical patient data to train an oral mucositis risk prediction model to predict the severity of oral mucositis after chemotherapy in patients, so as to accurately predict the grade of oral mucositis after chemotherapy and formulate corresponding nursing plans for patients.

[0046] Specifically, the oral mucositis risk factors of the current patient are used as input to the oral mucositis risk prediction model to predict the oral mucositis grade of the current patient after chemotherapy.

[0047] Step S3: Customize a corresponding prevention strategy for the patient based on the predicted oral mucositis grade of the current patient after chemotherapy.

[0048] For patients with predicted oral mucositis grade 1, oral hygiene, moisturizing, and basic care should be the main focus;

[0049] For patients with predicted oral mucositis grade 2, oral irritation should be reduced in addition to maintaining oral cleanliness and moisture, and appropriate medications may be used for intervention.

[0050] For patients with predicted oral mucositis grade three, stronger preventive and treatment measures should be taken on the basis of keeping the mouth clean and moist and avoiding all irritating foods and behaviors. Continuous monitoring and drug intervention by doctors are required.

[0051] Example 2:

[0052] See also Figure 1 , which shows a flowchart of the steps of a method for constructing a risk prediction model for oral mucositis care provided by one embodiment of the present invention, wherein the method is the specific steps of step S2:

[0053] Step S201: Obtain risk factor information of the patient.

[0054] It should be noted that this embodiment is a method for constructing a risk prediction model for oral mucositis care. Specifically, it trains an oral mucositis risk prediction model by analyzing the data of all patients who have visited the hospital in history. The severity of oral mucositis in patients after chemotherapy is predicted based on the oral mucositis risk prediction model. To this end, it is necessary to first obtain the patient's data.

[0055] Specifically, the data of all patients visiting the database at all risk levels are retrieved. The patient data includes the patient's risk level and risk factors. The risk level includes: mild risk, moderate risk and severe risk. The risk factors include: whether taking opioids, whether smoking, whether the oral pH value is less than 6.5, etc.

[0056] Step S202: Mark the patient according to the presence of each risk factor in the patient at each risk level, and obtain the patient marking sequence for each risk factor at each risk level; obtain the correlation of the patient marking sequences between different risk factors at the same risk level; based on the correlation, screen out the relevant risk factors of the risk factors at each risk level; based on the correlation between different risk factors at the same risk level, combined with the difference between the risk factors at the risk level and the relevant risk factors of the risk factors, obtain the degree of influence of the risk factors on the assessed risk level; based on the degree of influence of the risk factors on the assessed risk level, screen out the key risk factors for each risk level.

[0057] It should be noted that this embodiment specifically achieves the purpose of improving the data quality of the oral mucositis nursing knowledge base by screening all risk factors in the patient data. Therefore, it is necessary to first obtain the risk factors that have a great impact on the assessment of each risk level; considering that there are some correlations between different risk factors, the correlations between different risk factors can be further considered. In an embodiment of the present invention, by marking the patients, a patient marking sequence for each risk factor at all risk levels is obtained, and the patient marking sequences of different risk factors at the same risk level are used to quantify the similarity between different risk factors. The greater the impact of a risk factor on the assessed risk level, the more similar the risk factor is to other risk factors. The correlation between different risk factors can be used to obtain the degree of influence of the risk factor on the assessed risk level, and the key risk factors for each risk level can be screened out.

[0058] Specifically, for any risk factor, among all patients at each risk level, the patients with the risk factor are marked as , patients without the risk factors were labeled ; Obtain a patient tag set corresponding to the risk factors at each risk level, sort the elements in the patient tag set corresponding to the risk factors at each risk level according to the time of information entry, and obtain a patient tag sequence corresponding to the risk factors at each risk level; and are preset non-negative marker values, and Greater than , and The specific value of can be set according to the actual situation. This embodiment does not make a hard requirement. 、 to give a narrative;

[0059] Preferably, in one embodiment of the present invention, the difference between the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level and the sum of the sequence marker values ​​is obtained; the correlation between the patient marker sequences of different risk factors at the same risk level is obtained based on the difference between the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level and the sum of the sequence marker values;

[0060] The absolute value of the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level is positively correlated with the correlation between patient marker sequences of different risk factors at the same risk level; the difference in the sum of marker values ​​of patient marker sequences of different risk factors at the same risk level is negatively correlated with the correlation between patient marker sequences of different risk factors at the same risk level.

[0061] In a specific embodiment of the present invention, the calculation formula is:

[0062]

[0063] Where, Indicates the Patients in the risk category risk factors and The correlation between risk factors; Indicates the Risk level The patient marker sequence of the first risk factor is Pearson correlation coefficient between patient marker sequences of risk factors; Indicates the Risk level The sum of the marker values ​​in the marker sequence of patients with risk factors; Indicates the Risk level The sum of the marker values ​​in the marker sequence of patients with risk factors; Indicates absolute value operation; Represents an exponential function with a natural constant as the base; in this embodiment, Model to present inverse proportional relationship and normalization processing, As the input of the model, the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0064] Furthermore, a threshold value of the number of relevant risk factors is preset. , The specific value of can be set according to the actual situation. This embodiment does not make a hard requirement. to give a narrative;

[0065] Among the correlations between patient marker sequences of different risk factors at the same risk level, the one with the greatest correlation is risk factors as related risk factors of the risk factors under this risk level;

[0066] For the degree of influence of any risk factor on the assessment of any risk level, obtain the difference between the risk factor and the related risk factors of the risk factor at the same risk level;

[0067] The correlation between different risk factors at the same risk level is positively correlated with the degree of influence of risk factors on the assessed risk level; the difference between risk factors and related risk factors at the same risk level is positively correlated or negatively correlated with the degree of influence of risk factors on the assessed risk level.

[0068] In a specific embodiment of the present invention, the calculation formula is:

[0069]

[0070] Where, Indicates the Risk factors for assessing The degree of impact of each risk level; Indicates the preset threshold value of the number of relevant risk factors; represents the number of risk factors; Indicates the Patients in the risk category risk factors and The correlation between risk factors; Indicates the Patients in the risk category risk factors and The correlation between the risk factors; Indicates the Patients in the risk category The mean of the correlations between a risk factor and all relevant risk factors; Indicates absolute value operation; Represents a normalization function, and the normalization object is the criticality of each risk factor under all risk levels.

[0071] It should be noted that and The larger the value, the Patients in the risk category The stronger the correlation between a risk factor and all other risk factors, the greater the impact of the risk factor on the risk level assessment. and The larger the value of Risk factors for assessing The higher the degree of influence of a risk factor on the assessed risk level, the higher the similarity between the risk factor and all related factors. The smaller the value, the greater the impact of the risk factor on the assessed risk level.

[0072] Furthermore, a threshold for the impact of risk factors under a single risk is preset. , The specific value of can be set according to the actual situation. This embodiment does not make a hard requirement. Describe, when any risk factor has a greater impact on the assessment of any risk level When the risk factor is ≥ 1, the risk factor is recorded as the key risk factor for assessing the risk level.

[0073] At this point, the key risk factors for assessing each risk level are obtained.

[0074] Step S203: Obtain the importance of each risk factor based on the degree of influence of each risk factor on all assessed risk levels and the number of patients with the corresponding risk factor in the patient data; and screen out all important risk factors based on the importance of each risk factor.

[0075] It should be noted that the key risk factors for evaluating each risk level obtained in step S202 are indicators for a risk level. Since there are multiple analysis levels for the patient evaluation results, and the more patients with corresponding risk factors in the patient data, the more likely the risk factor is to cause oral mucositis, that is, the more important the risk factor is, it is also necessary to obtain the importance of each risk factor based on the degree of influence of each risk factor on all evaluated risk levels and the number of patients with corresponding risk factors in the patient data, and further screen out all important risk factors.

[0076] Specifically, the importance of each risk factor is obtained based on the degree of influence of each risk factor on all assessed risk levels and the number of patients with the corresponding risk factor in the patient data;

[0077] The degree of influence of each risk factor on the assessment of all risk levels is positively correlated with the importance of each risk factor; the number of patients with each risk factor is positively correlated with the importance of each risk factor; the difference in the degree of influence of each risk factor on the assessment of different risk levels is negatively correlated with the importance of each risk factor;

[0078] In a specific embodiment of the present invention, the calculation formula is:

[0079]

[0080] Where, Indicates the The importance of each risk factor; Indicates the Risk factors for assessing The degree of impact of each risk level; A number indicating the level of risk; Indicates the The average value of the impact of each risk factor on the assessment of all risk levels; Indicates absolute value operation; Represents an exponential function with a natural constant as the base; in this embodiment, Model to present inverse proportional relationship and normalization processing, As the input of the model, the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0081] It should be noted that The larger the value of The greater the impact of each risk factor on the assessment of all risk levels; The larger the value, the higher the The greater the number of patients with each risk factor, the and The larger the value of The more important the risk factor is; The larger the value, the The more stable the impact of each risk factor on the assessment of each risk level, the The more important the risk factor.

[0082] Furthermore, a threshold of importance is preset , The specific value of can be set according to the actual situation. This embodiment does not make a hard requirement. Describe; when the importance of risk factors is greater than When the risk factor is significant, it is recorded as a significant risk factor.

[0083] At this point, all important risk factors are obtained.

[0084] Step S204: training an oral mucositis risk prediction model based on the key risk factors and important risk factors of each risk level; predicting the oral mucositis level of the patient after chemotherapy based on the prediction model, and providing care for the patient.

[0085] It should be noted that after obtaining the key risk factors and important risk factors of each risk level respectively through step S202 and step S203, the target risk factors can be obtained according to the key risk factors and important risk factors of each risk level; the target risk factors are risk factors that play a decisive role in assessing the risk level, so the oral mucositis risk prediction model can be trained according to the target risk factors to accurately predict the oral mucositis level of patients who are going to undergo chemotherapy after chemotherapy, and further provide corresponding care for the patients before chemotherapy based on the predicted level, thereby reducing the incidence of oral mucositis complications and alleviating the suffering of patients.

[0086] Optionally, in a specific embodiment of the present invention, a set of key risk factors of all risk levels is recorded as a key factor set, and a set of all important risk factors is recorded as an important factor set; the risk factors in the union of the key factor set and the important factor set are recorded as target risk factors;

[0087] Furthermore, the set consisting of all target risk factors is used as the first random forest algorithm training set, and the set consisting of all non-target risk factors is used as the second random forest algorithm training set; a weight adjustment value greater than 0.5 and less than 1 is preset. , The specific value of can be set according to the actual situation. This embodiment does not make a hard requirement. to give a narrative;

[0088] Let the random forest algorithm be The probability of randomly selecting risk factors from the No. 1 random forest algorithm training set is The probability of randomly selecting risk factors from the No. 2 random forest algorithm training set is used to train a risk prediction model for oral mucositis. Since the random forest algorithm is a well-known existing technology, it will not be described in detail in this embodiment.

[0089] At this point, the oral mucositis risk prediction model in step S2 is obtained. Subsequently, the oral mucositis risk prediction model can be used to predict the oral mucositis grade of the patient after chemotherapy, so as to provide the same prevention and care for the patient.

[0090] Figure 2 Flowchart for obtaining the oral mucositis risk prediction model.

[0091] Example 3:

[0092] Example 3 shows the specific implementation steps of step S3:

[0093] It should be noted that after obtaining the oral mucositis risk prediction model, the oral mucositis risk prediction model can be used to accurately predict the oral mucositis grade of patients who are undergoing chemotherapy after chemotherapy, so as to further provide corresponding care for the patients in advance based on the predicted oral mucositis grade, thereby reducing the incidence of oral mucositis complications and alleviating the patients' pain.

[0094] Optionally, in one embodiment of the present invention, the patient's risk factors are input into an oral mucositis risk prediction model to obtain the patient's oral mucositis grade after chemotherapy.

[0095] For patients with predicted oral mucositis grade 1, dietary care is required;

[0096] For patients with predicted oral mucositis grade 2, dietary care and physical care are required;

[0097] For patients with predicted oral mucositis grade three, dietary care, physical care and drug care are required;

[0098] The dietary care includes but is not limited to: providing patients with soft, small, easy-to-swallow foods, while avoiding acidic and spicy foods, and finally providing patients with a meal plan of eating small and frequent meals;

[0099] The physical treatment includes but is not limited to: cryotherapy and low-intensity laser therapy;

[0100] The medical care includes, but is not limited to, having the patient take nonsteroidal anti-inflammatory drugs.

[0101] It should be noted that physical care can produce side effects such as swelling and edema, while drug care has side effects that affect the patient's digestive system. Therefore, in order to alleviate the patient's pain, it is necessary to provide the patient with corresponding care in advance based on the patient's predicted oral mucositis level.

[0102] At this point, this embodiment is completed.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for constructing a risk prediction model for oral mucositis care, characterized in that: The method comprises the following steps: Obtaining information on all patients who visit the hospital, including the patient's risk level and risk factors; Marking patients based on the presence of each risk factor in patients at each risk level to obtain a patient marker sequence for each risk factor at each risk level; obtaining correlations between patient marker sequences for different risk factors at the same risk level; screening out related risk factors for risk factors at each risk level based on the correlations; obtaining the degree of influence of risk factors on the assessed risk level based on the correlations between different risk factors at the same risk level and the differences between risk factors at the risk level and their related risk factors; screening out key risk factors for each risk level based on the degree of influence of risk factors on the assessed risk level; According to the impact of each risk factor on all assessed risk levels and the number of patients with the corresponding risk factor in the patient data, the importance of each risk factor is obtained; according to the importance of each risk factor, all important risk factors are screened out; An oral mucositis risk prediction model is trained based on the key risk factors and important risk factors of each risk level; based on the prediction model, the oral mucositis level of the patient after chemotherapy is predicted, and care is provided to the patient.

2. The method for constructing a risk prediction model for oral mucositis care according to claim 1, characterized in that: The method of marking the patient according to the presence of each risk factor in the patient at each risk level and obtaining the patient marking sequence for each risk factor at each risk level includes the following specific methods: For any risk factor, among all patients at each risk level, the patients with the risk factor are marked as , patients without the risk factors were labeled ; Obtain a patient tag set corresponding to the risk factors at each risk level, sort the elements in the patient tag set corresponding to the risk factors at each risk level according to the time of information entry, and obtain a patient tag sequence corresponding to the risk factors at each risk level; and are preset non-negative marker values, and Greater than .

3. The method for constructing a risk prediction model for oral mucositis care according to claim 2, characterized in that: Tagged Value 1, mark value is 0.

4. The method for constructing a risk prediction model for oral mucositis care according to claim 2, wherein: The specific method for obtaining the correlation between patient marker sequences of different risk factors at the same risk level includes: Obtaining the difference between the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level and the sum of the sequence marker values; obtaining the correlation between the patient marker sequences of different risk factors at the same risk level based on the difference between the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level and the sum of the sequence marker values; The absolute value of the Pearson correlation coefficient between patient marker sequences of different risk factors at the same risk level is positively correlated with the correlation between patient marker sequences of different risk factors at the same risk level; the difference in the sum of marker values ​​of patient marker sequences of different risk factors at the same risk level is negatively correlated with the correlation between patient marker sequences of different risk factors at the same risk level.

5. The method for constructing a risk prediction model for oral mucositis care according to claim 1, wherein: The method of screening out the related risk factors of the risk factors at each risk level based on the correlation includes: Preset a threshold for the number of relevant risk factors , among the correlations between patient marker sequences of different risk factors at the same risk level, the one with the largest correlation risk factors as related risk factors of the risk factors under this risk level.

6. The method for constructing a risk prediction model for oral mucositis care according to claim 1, characterized in that: The method of obtaining the degree of influence of risk factors on the assessed risk level based on the correlation between different risk factors at the same risk level and the differences between risk factors at the risk level and their related risk factors is as follows: For the degree of influence of any risk factor on the assessment of any risk level, obtain the difference between the risk factor and the related risk factors of the risk factor at the same risk level; The correlation between different risk factors at the same risk level is positively correlated with the degree of influence of risk factors on the assessed risk level; the difference between risk factors and related risk factors at the same risk level is positively correlated or negatively correlated with the degree of influence of risk factors on the assessed risk level.

7. The method for constructing a risk prediction model for oral mucositis care according to claim 1, characterized in that: The method of screening out the key risk factors for each risk level based on the degree of influence of the risk factors on the risk level assessment includes: Preset the risk factor impact threshold for a single risk , when the impact of any risk factor on the assessment of any risk level is greater than When the risk factor is ≥ 1, the risk factor is recorded as the key risk factor for assessing the risk level.

8. The method for constructing a risk prediction model for oral mucositis care according to claim 1, characterized in that: The importance of each risk factor is obtained based on the degree of influence of each risk factor on all assessed risk levels, combined with the number of patients with the corresponding risk factor in the patient data, including the following specific methods: The degree of influence of each risk factor on the assessment of all risk levels is positively correlated with the importance of each risk factor; the number of patients with each risk factor is positively correlated with the importance of each risk factor; the difference in the degree of influence of each risk factor on the assessment of different risk levels is negatively correlated with the importance of each risk factor.

9. The method for constructing a risk prediction model for oral mucositis care according to claim 1, wherein: The specific methods for screening out all important risk factors according to the importance of each risk factor include: Preset an importance threshold When the importance of risk factors is greater than When the risk factor is significant, it is recorded as a significant risk factor.

10. The method for constructing a risk prediction model for oral mucositis care according to claim 1, characterized in that: The specific method of training the oral mucositis risk prediction model includes: The set of key risk factors of all risk levels is recorded as the key factor set, and the set of all important risk factors is recorded as the important factor set; the risk factors in the union of the key factor set and the important factor set are recorded as the target risk factors; The set consisting of all target risk factors is used as the first random forest algorithm training set, and the set consisting of all non-target risk factors is used as the second random forest algorithm training set; preset an adjustment weight greater than 0.5 and less than 1 ; Let the random forest algorithm be The probability of randomly selecting risk factors from the No. 1 random forest algorithm training set is The probability of randomly selecting risk factors from the No. 2 random forest algorithm training set is used to train the oral mucositis risk prediction model.

Images