Anesthesia assessment decision tree construction method and system

Abstract

The invention relates to the technical field of data analysis, in particular to an anesthesia evaluation decision tree construction method and system, and the system comprises a data collection module, a data integration module, a data analysis module and a model construction module. The method comprises the following steps: establishing an evaluation database based on anesthesia evaluation data, processing and analyzing the database to obtain an evaluation item complete set of an anesthesia evaluation process, performing data analysis on the evaluation item complete set to obtain a node feature set of evaluation items, and constructing a main node guide value, a secondary node guide value and a child node guide value according to the node feature set. The main node, the secondary nodes and the child nodes are constructed based on the node guide values, the anesthesia evaluation decision tree is constructed based on the nodes, and through the constructed nodes and the branches formed among the nodes, anesthesia evaluation can be performed on a patient visually and efficiently, the burden of an anesthetist is shared, and the accuracy of anesthesia evaluation is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a method and system for constructing an anesthesia assessment decision tree. Background Art

[0002] In modern medicine, anesthesia has become an indispensable part of the surgical process. The quality of anesthesia plays a decisive role in whether the operation goes smoothly. Among them, anesthesia evaluation is extremely important to ensure the safety and comfort of the patient's operation. Through comprehensive evaluation methods and detailed evaluation processes, anesthesiologists can develop safe and effective anesthesia plans based on the patient's unique situation to ensure the smooth progress of the operation.

[0003] However, in the existing medical system, anesthesiologists are often the most scarce physician resource. In any hospital, a large number of surgeries require anesthesia for patients, including local anesthesia, partial anesthesia or general anesthesia, which causes anesthesiologists to be overloaded with tasks. This results in an inefficient process in anesthesia assessment. At the same time, it can cause anesthesiologists to be overworked, easily lead to omissions, and cause surgical risks. Therefore, a method and system for constructing an anesthesia assessment decision tree is needed to help solve problems in anesthesia assessment. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the background technology and to propose a method and system for constructing an anesthesia assessment decision tree.

[0005] In order to achieve the above object, the present invention adopts the following technical solution: a method for constructing an anesthesia assessment decision tree, comprising the following steps:

[0006] S1. Collect anesthesia assessment data;

[0007] S2. Establish an evaluation database based on the anesthesia evaluation data, and process the evaluation database to obtain a complete set of evaluation items;

[0008] S3, perform data analysis on the entire set of evaluation items to obtain a node feature set;

[0009] S4. Construct nodes according to the node feature set, and construct an anesthesia assessment decision tree based on the constructed nodes.

[0010] As a further solution of the present invention, with respect to step S2, the method for processing data on the evaluation database is:

[0011] Several evaluation items of anesthesia evaluation data in the evaluation database are determined, and a union operation is performed on the evaluation items of the anesthesia evaluation data contained in the evaluation database to obtain an operation set of the anesthesia evaluation data, which is used as a complete set of evaluation items of the evaluation database.

[0012] As a further solution of the present invention, with respect to step S3, data analysis is performed on the entire set of evaluation items, including the following steps:

[0013] S301, randomly select an evaluation item from the entire set of evaluation items, and mark it as a demand evaluation item;

[0014] S302, obtaining all anesthesia assessment data including demand assessment items in the assessment database, and marking them as target assessment data, obtaining text information of demand assessment items in the target assessment data, using them as demand texts of the target assessment data, integrating the demand texts of all target assessment data to obtain summary texts, splitting the text information of the summary text into multiple assessment phrases by text processing technology, and calculating the word frequencies of different assessment phrases by the following formula:

[0015] In the formula, Ci is the frequency of the evaluation phrase, Ni is the number of times the evaluation phrase appears in the summary text, Z is the total number of demand evaluation items, and i is the number of different evaluation phrases;

[0016] S303, marking the evaluation phrases with a word frequency Ci of 1 as recognition phrases, marking the evaluation phrases with a word frequency Ci not equal to 1 as target phrases, combining the recognition phrases and the target phrases to obtain a characteristic word set of the demand evaluation item; the recognition phrases and the target phrases correspond one to one;

[0017] S304, obtaining a feature word set of all evaluation items in the entire set of evaluation items, and sequentially placing the feature word sets of the evaluation items into the anesthesia coefficient library for identification, determining the phrase type of the identification phrase in the feature word set, determining the anesthesia coefficient of the target phrase corresponding to the feature word set based on the phrase type of the identification phrase, and combining the phrase type and the anesthesia coefficient to obtain a parsed set of the evaluation items;

[0018] S305. Calculate a node feature set of the entire set of evaluation items according to corresponding analysis sets of several evaluation items in the entire set of evaluation items.

[0019] As a further solution of the present invention, the phrase type includes a quantization phrase and an adjustment phrase, the anesthesia coefficient includes a quantization coefficient and an adjustment coefficient, the quantization phrase corresponds to the quantization coefficient, and the adjustment phrase corresponds to the adjustment coefficient.

[0020] As a further solution of the present invention, if the phrase type is a quantized phrase, a quantized coefficient of a target phrase corresponding to the identified phrase is obtained, and a calculation is performed based on the numerical value of the target phrase and the quantized coefficient, and the result of the calculation is used as an anesthetic dose estimation of the demand assessment item;

[0021] If the phrase type is an adjustment phrase, the adjustment coefficient of the target phrase corresponding to the adjustment phrase is obtained, and it is determined whether there is an anesthetic dose estimation; if there is an anesthetic dose estimation, the anesthetic dose estimation is multiplied by the adjustment coefficient, and the result of the multiplication is used as the dose adjustment value of the demand assessment item; if there is no anesthetic dose estimation, wait until there is an anesthetic dose estimation.

[0022] As a further solution of the present invention, for step S305, the method for calculating the node feature set of the entire set of evaluation items is:

[0023] The dose adjustment ratio TZ of this demand assessment item is calculated by the following formula:

[0024]

[0025] In the formula, MG is the estimated anesthetic dose, and MT is the dose adjustment value;

[0026] The node feature set is obtained by combining the dose adjustment ratios TZ of different evaluation items contained in the full set of evaluation items.

[0027] As a further solution of the present invention, for step S4, the method for optimizing the node feature set is:

[0028] Sort several dose adjustment ratios in the node feature set in ascending order to obtain a node feature sequence, obtain the median of the dose adjustment ratio in the node feature sequence, mark it as a main node guide value, compare the main node guide value with other sequence elements of the node feature sequence, mark all sequence elements greater than the main node guide value as left branch elements, and mark all sequence elements less than the main node guide value as right branch elements; establish two corresponding main branch sequences according to the ascending order of the left branch elements and the right branch elements;

[0029] The same method is applied to the two main branch sequences to obtain two secondary node guide values ​​and two secondary branch sequences corresponding to each secondary node guide value;

[0030] By analogy, the subsequently obtained nodes and the corresponding two branch sequences are used as the child node guide value and the subsequence, respectively, until there is only one dose adjustment ratio in the subsequence;

[0031] The evaluation items corresponding to the decision tree constructed by the main node guide value, the secondary node guide value and the child node guide value are marked as the main node, the secondary node and the child node respectively, and the anesthesia evaluation decision tree is constructed based on the main node, the secondary node and the child node.

[0032] In addition, the present invention also discloses an anesthesia assessment decision tree construction system, including a data acquisition module for collecting anesthesia assessment data;

[0033] A data integration module is used to establish an evaluation database based on anesthesia evaluation data, and to process the evaluation database to find a complete set of evaluation items;

[0034] The data analysis module is used to perform data analysis on the entire set of evaluation items to obtain a node feature set;

[0035] The model building module is used to construct nodes for the node feature set and build an anesthesia assessment decision tree based on the constructed nodes.

[0036] Compared with the existing technology, the advantages of the present invention are: establishing an evaluation database through the collected anesthesia evaluation data, obtaining a full set of evaluation items of the anesthesia evaluation process by processing the database, obtaining a node feature set of the evaluation items by performing data analysis on the full set of evaluation items, constructing a main node guide value, a secondary node guide value and a sub-node guide value according to the node feature set, constructing a main node, a secondary node and a sub-node based on the node guide value, constructing an anesthesia evaluation decision tree based on the node, and facilitating an intuitive and efficient anesthesia evaluation of the patient through the constructed nodes and the branches formed between several nodes, thereby sharing the burden of the anesthesiologist and improving the accuracy of the anesthesia evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0038] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0039] Embodiment 1, refer to Figure 1 , a method for constructing an anesthesia assessment decision tree, comprising the following steps:

[0040] S1. Collect anesthesia assessment data;

[0041] S2. Establish an evaluation database based on the anesthesia evaluation data, and process the evaluation database to obtain a complete set of evaluation items;

[0042] For step S2, the method for processing data on the evaluation database is:

[0043] Determine several evaluation items of anesthesia evaluation data in the evaluation database, perform a union operation on the evaluation items of anesthesia evaluation data contained in the evaluation database, obtain an operation set of anesthesia evaluation data, and use it as the full set of evaluation items of the evaluation database; it should be noted that one evaluation item only contains one evaluation type, for example, weight: 50kg; another example, gender: male; another example, medical history: heart disease; another example, medical history: heart disease, liver disease;

[0044] S3, perform data analysis on the entire set of evaluation items to obtain a node feature set;

[0045] For step S3, data analysis is performed on the entire set of evaluation items, including the following steps:

[0046] S301, randomly extracting evaluation items from the entire set of evaluation items, and marking them as demand evaluation items; in this embodiment, the method of randomly extracting evaluation items is implemented by a sampling without replacement algorithm;

[0047] S302, obtaining all anesthesia assessment data including demand assessment items in the assessment database, and marking them as target assessment data, obtaining text information of demand assessment items in the target assessment data, using them as demand texts of the target assessment data, integrating the demand texts of all target assessment data to obtain summary texts, splitting the text information of the summary text into multiple assessment phrases by text processing technology, and calculating the word frequencies of different assessment phrases by the following formula:

[0048] In the formula, Ci is the frequency of the evaluation phrase, Ni is the number of times the evaluation phrase appears in the summary text, Z is the total number of demand evaluation items, and i is the number of different evaluation phrases;

[0049] S303, marking the evaluation phrases with a word frequency Ci of 1 as recognition phrases, marking the evaluation phrases with a word frequency Ci not equal to 1 as target phrases, combining the recognition phrases and the target phrases to obtain a characteristic word set of the demand evaluation item; the recognition phrases and the target phrases correspond one to one;

[0050] S304, obtaining a feature word set of all evaluation items in the entire set of evaluation items, and sequentially placing the feature word sets of the evaluation items into the anesthesia coefficient library for identification, determining the phrase type of the identification phrase in the feature word set, determining the anesthesia coefficient of the target phrase corresponding to the feature word set based on the phrase type of the identification phrase, and combining the phrase type and the anesthesia coefficient to obtain a parsed set of evaluation items; the phrase type includes a quantification phrase and an adjustment phrase, the anesthesia coefficient includes a quantification coefficient and an adjustment coefficient, the quantification phrase corresponds to the quantification coefficient, and the adjustment phrase corresponds to the adjustment coefficient;

[0051] It should be noted that the anesthesia coefficient library is obtained based on the statistics of the existing anesthesia dosage system; the target phrases include but are not limited to Chinese phrases, English phrases, digital phrases, and collocation phrases containing Chinese, English and numbers;

[0052] If the phrase type is a quantitative phrase, the quantitative coefficient of the target phrase corresponding to the identified phrase is obtained, and a calculation is performed according to the numerical value of the target phrase and the quantitative coefficient, and the calculated result is used as the anesthetic dose estimation of the demand assessment item;

[0053] If the phrase type is an adjustment phrase, the adjustment coefficient of the target phrase corresponding to the adjustment phrase is obtained, and it is determined whether there is an anesthetic dose estimation; if there is an anesthetic dose estimation, the anesthetic dose estimation is multiplied by the adjustment coefficient, and the result of the multiplication is used as the dose adjustment value of the demand assessment item; if there is no anesthetic dose estimation, wait until there is an anesthetic dose estimation;

[0054] The anesthetic dose estimation and anesthetic dose adjustment values ​​are both realized based on the existing anesthetic dose modulation theory, which will not be elaborated here;

[0055] S305, calculating a node feature set of the entire set of evaluation items according to corresponding analysis sets of several evaluation items in the entire set of evaluation items;

[0056] For step S305, the method for calculating the node feature set of the entire set of evaluation items is:

[0057] The dose adjustment ratio TZ of this demand assessment item is calculated by the following formula:

[0058]

[0059] In the formula, MG is the estimated anesthetic dose, and MT is the dose adjustment value;

[0060] The node feature set is obtained by combining the dose adjustment ratios TZ of different evaluation items contained in the full set of evaluation items;

[0061] S4, constructing nodes according to the node feature set, and constructing an anesthesia assessment decision tree based on the constructed nodes;

[0062] For step S4, the method for optimizing the node feature set is:

[0063] Sort several dose adjustment ratios in the node feature set in ascending order to obtain a node feature sequence, obtain the median of the dose adjustment ratio in the node feature sequence, mark it as a main node guide value, compare the main node guide value with other sequence elements of the node feature sequence, mark all sequence elements greater than the main node guide value as left branch elements, and mark all sequence elements less than the main node guide value as right branch elements; establish two corresponding main branch sequences according to the ascending order of the left branch elements and the right branch elements;

[0064] The same method is applied to the two main branch sequences to obtain two secondary node guide values ​​and two secondary branch sequences corresponding to each secondary node guide value;

[0065] By analogy, the subsequently obtained nodes and the corresponding two branch sequences are used as the child node guide value and the subsequence, respectively, until there is only one dose adjustment ratio in the subsequence;

[0066] The evaluation items corresponding to the decision tree constructed by the main node guide value, the secondary node guide value and the child node guide value are marked as the main node, the secondary node and the child node respectively, and the anesthesia evaluation decision tree is constructed based on the main node, the secondary node and the child node.

[0067] In addition, the present invention also discloses an anesthesia assessment decision tree construction system, including a data acquisition module for collecting anesthesia assessment data;

[0068] A data integration module is used to establish an evaluation database based on anesthesia evaluation data, and to process the evaluation database to find a complete set of evaluation items;

[0069] The data analysis module is used to perform data analysis on the entire set of evaluation items to obtain a node feature set;

[0070] The model building module is used to construct nodes for the node feature set and build an anesthesia assessment decision tree based on the constructed nodes.

[0071] Example 2: Before surgery, a basic dose needs to be set according to the patient's weight. The weights in 10 anesthesia assessment data are set as follows: weight: Wk, the unit of weight is kg, k is the number of the anesthesia assessment data, k∈[1, 10]; by splitting the text information, the evaluation phrases are obtained as follows:

[0072] "Weight", "Wk";

[0073] Among them, the number of times "weight" appears is 10, and its word frequency is 1. "weight" is marked as a recognition phrase, and "Wk" is marked as a target phrase. The recognition phrase "weight" is put into the anesthesia coefficient library for recognition, and the corresponding coefficient type is obtained as a quantization phrase. The quantization coefficient of the target phrase "Wk" is obtained, and the value of the target phrase "Wk" is multiplied by the quantization coefficient to obtain the anesthetic dose estimation.

[0074] In this embodiment, the weights of three patients B1, B2 and B3 were obtained before surgery, and the corresponding anesthetic dose estimates G1, G2 and G3 were obtained. At the same time, other evaluation items were obtained as follows:

[0075] B1, medical history: heart disease;

[0076] B2, medical history: none;

[0077] B3, medical history: liver disease;

[0078] In Example 3, the multiple evaluation phrases obtained by splitting are: "medical history", "heart disease", "liver disease", and the word frequency of "medical history" is 1. It is marked as an identification phrase and identified, and the corresponding phrase type is an adjustment phrase. "Heart disease" and "liver disease" are both target phrases, and the adjustment coefficients C1 and C2 corresponding to the target phrases "heart disease" and "liver disease" are obtained; then the anesthetic dose adjustment value of patient B1 is G1×C1; the anesthetic dose adjustment value of patient B2 is G1; the anesthetic dose adjustment value of patient B3 is G1×C2; it should be noted that if the patient's medical history is "none", that is, it does not exist, its adjustment coefficient is set to 1; wherein the parsed sets of the evaluation items of patients B1, B2, and B3 are {adjustment phrase, adjustment coefficient corresponding to heart disease}, {adjustment phrase, 1}, {adjustment phrase, adjustment coefficient corresponding to liver disease}, respectively;

[0079] In the second and third embodiments, the method for identifying the recognition phrase is implemented by database retrieval, which is a prior art and will not be described in detail here.

[0080] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for constructing an anesthesia assessment decision tree, characterized in that: The steps include: S1. Collect anesthesia assessment data; S2. Establish an evaluation database based on the anesthesia evaluation data, and process the evaluation database to obtain a complete set of evaluation items; S3, perform data analysis on the entire set of evaluation items to obtain a node feature set; S4. Construct nodes according to the node feature set, and construct an anesthesia assessment decision tree based on the constructed nodes.

2. The method for constructing an anesthesia assessment decision tree according to claim 1, characterized in that: For step S2, the method for processing data on the evaluation database is: Several evaluation items of anesthesia evaluation data in the evaluation database are determined, and a union operation is performed on the evaluation items of the anesthesia evaluation data contained in the evaluation database to obtain an operation set of the anesthesia evaluation data, which is used as a complete set of evaluation items of the evaluation database.

3. The method for constructing an anesthesia assessment decision tree according to claim 2, characterized in that: For step S3, data analysis is performed on the entire set of evaluation items, including the following steps: S301, randomly select an evaluation item from the entire set of evaluation items, and mark it as a demand evaluation item; S302, obtaining all anesthesia assessment data including demand assessment items in the assessment database, and marking them as target assessment data, obtaining text information of the demand assessment items in the target assessment data, using them as demand texts of the target assessment data, integrating the demand texts of all target assessment data to obtain summary texts, splitting the text information of the summary text into multiple assessment phrases by text processing technology, and calculating the word frequencies of different assessment phrases by the following formula: In the formula, Ci is the frequency of the evaluation phrase, Ni is the number of times the evaluation phrase appears in the summary text, Z is the total number of demand evaluation items, and i is the number of different evaluation phrases; S303, marking the evaluation phrases with a word frequency Ci of 1 as recognition phrases, marking the evaluation phrases with a word frequency Ci not equal to 1 as target phrases, combining the recognition phrases and the target phrases to obtain a characteristic word set of the demand evaluation item; the recognition phrases and the target phrases correspond one to one; S304, obtaining a feature word set of all evaluation items in the entire set of evaluation items, and sequentially placing the feature word sets of the evaluation items into the anesthesia coefficient library for identification, determining the phrase type of the identification phrase in the feature word set, determining the anesthesia coefficient of the target phrase corresponding to the feature word set based on the phrase type of the identification phrase, and combining the phrase type and the anesthesia coefficient to obtain a parsed set of the evaluation items; S305. Calculate a node feature set of the entire set of evaluation items according to corresponding analysis sets of several evaluation items in the entire set of evaluation items.

4. The method for constructing an anesthesia assessment decision tree according to claim 3, characterized in that: The phrase types include quantification phrases and adjustment phrases, the anesthesia coefficients include quantification coefficients and adjustment coefficients, the quantification phrases correspond to the quantification coefficients, and the adjustment phrases correspond to the adjustment coefficients.

5. The method for constructing an anesthesia assessment decision tree according to claim 4, characterized in that: If the phrase type is a quantitative phrase, the quantitative coefficient of the target phrase corresponding to the identified phrase is obtained, and a calculation is performed according to the numerical value of the target phrase and the quantitative coefficient, and the calculated result is used as the anesthetic dose estimation of the demand assessment item; If the phrase type is an adjustment phrase, the adjustment coefficient of the target phrase corresponding to the adjustment phrase is obtained, and it is determined whether there is an anesthetic dose estimation; if there is an anesthetic dose estimation, the anesthetic dose estimation is multiplied by the adjustment coefficient, and the result of the multiplication is used as the dose adjustment value of the demand assessment item; if there is no anesthetic dose estimation, wait until there is an anesthetic dose estimation.

6. The method for constructing an anesthesia assessment decision tree according to claim 5, characterized in that: For step S305, the method for calculating the node feature set of the entire set of evaluation items is: The dose adjustment ratio TZ of this demand assessment item is calculated by the following formula: In the formula, MG is the estimated anesthetic dose, and MT is the dose adjustment value; The node feature set is obtained by combining the dose adjustment ratios TZ of different evaluation items contained in the full set of evaluation items.

7. The method for constructing an anesthesia assessment decision tree according to claim 6, characterized in that: For step S4, the method for optimizing the node feature set is: Sort several dose adjustment ratios in the node feature set in ascending order to obtain a node feature sequence, obtain the median of the dose adjustment ratio in the node feature sequence, mark it as a main node guide value, compare the main node guide value with other sequence elements of the node feature sequence, mark all sequence elements greater than the main node guide value as left branch elements, and mark all sequence elements less than the main node guide value as right branch elements; establish two corresponding main branch sequences according to the ascending order of the left branch elements and the right branch elements; The same method is applied to the two main branch sequences to obtain two secondary node guide values ​​and two secondary branch sequences corresponding to each secondary node guide value; By analogy, the subsequently obtained nodes and the corresponding two branch sequences are used as the child node guide value and the subsequence, respectively, until there is only one dose adjustment ratio in the subsequence; The evaluation items corresponding to the decision tree constructed by the main node guide value, the secondary node guide value and the child node guide value are marked as the main node, the secondary node and the child node respectively, and the anesthesia evaluation decision tree is constructed based on the main node, the secondary node and the child node.

8. The anesthesia assessment decision tree construction system according to claim 1, characterized in that: It includes a data acquisition module for collecting anesthesia assessment data; A data integration module is used to establish an evaluation database based on anesthesia evaluation data, and to process the evaluation database to find a complete set of evaluation items; The data analysis module is used to perform data analysis on the entire set of evaluation items to obtain a node feature set; The model building module is used to construct nodes for the node feature set and build an anesthesia assessment decision tree based on the constructed nodes.

Images