Gestation period premature delivery risk auxiliary prediction method and system

Abstract

The invention relates to the technical field of biomedicine, in particular to a gestational period premature delivery risk auxiliary prediction method and system, and the method comprises the steps: obtaining the historical body data and delivery cycle of a pregnant woman, and determining the historical premature delivery degree of the pregnant woman according to the delivery cycle; according to the premature delivery degree difference between different historical pregnant women and the deviation degree change condition difference of the same type of body data during each physical examination, determining the premature delivery correlation degree of different types of body data of the historical pregnant women; clustering the historical pregnant women according to the deviation time difference of the same type of body data among different historical pregnant women, determining the overall correlation of each type of body data according to the number of pregnant women in the cluster and the premature delivery correlation degree of the corresponding type of body data, and screening out the target type of body data; and based on the body data of the target type of the historical pregnant woman during different times of physical examination, constructing a premature delivery prediction model to perform premature delivery prediction. According to the method, the accuracy of premature delivery prediction is effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technology, and in particular to a method and system for auxiliary prediction of premature birth risk during pregnancy. Background Art

[0002] Preterm birth refers to delivery before 37 weeks of gestation. The risk of premature birth is affected by a variety of physiological, pathological and environmental factors, including the health of the pregnant woman, fetal development, lifestyle and external environment. Common risk factors include shortened cervical length, pregnancy-induced hypertension or diabetes, placental abnormalities, infection (such as intrauterine infection or urinary tract infection), multiple pregnancy, history of premature birth, and long-term psychological stress or bad lifestyle.

[0003] Premature birth may not only cause the fetus to face health problems such as low birth weight, respiratory distress, brain damage, etc. after birth, but also increase the risk of delivery for pregnant women. Therefore, monitoring, prediction and intervention of premature birth risk are crucial to the health of mothers and babies. Regular prenatal examinations, dynamic monitoring of key physiological indicators, timely identification of premature birth signs, and necessary medical intervention can effectively reduce the incidence of premature birth and its adverse consequences.

[0004] In the prior art, various data from maternal checkups can be used to construct a decision tree using the random forest method, and the decision tree can be used to predict premature birth during pregnancy. However, due to the presence of redundant or irrelevant features in various data from maternal checkups, when the random forest constructs a decision tree by randomly selecting features, if the feature selection is inappropriate, it may affect the performance of the model, thereby resulting in inaccurate prediction of premature birth during pregnancy. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for auxiliary prediction of the risk of premature birth during pregnancy, so as to solve the problem that the existing prediction of premature birth during pregnancy is not accurate due to improper selection of prenatal examination data.

[0006] In order to solve the above technical problems, in a first aspect, the present invention provides a method for auxiliary prediction of risk of premature birth during pregnancy, comprising the following steps: Acquiring physical data and delivery cycles of pregnant women with different histories, wherein the physical data includes different types of physical data during each physical examination, and the gestational time of pregnant women with different histories during the same physical examination is the same; Determine the degree of premature birth of different historical pregnant women according to the delivery cycle, and determine the deviation of different types of physical data during different physical examinations of each historical pregnant woman according to the degree of premature birth and the difference of the same type of physical data during the same physical examination between different historical pregnant women; Determine the preterm birth correlation degree of different types of physical data for each historical pregnant woman based on the differences in the preterm birth degrees among different historical pregnant women and the differences in the deviation changes of the same type of physical data during different physical examinations; Determine the deviation time of different types of physical data for each historical pregnant woman according to the deviation change situation of the same type of physical data during different physical examinations of each historical pregnant woman, and cluster different historical pregnant women according to the differences in the deviation times of the same type of physical data among different historical pregnant women to obtain several clusters; Determine the overall correlation of different types of physical data according to the preterm birth correlation degree of the corresponding type of physical data of all historical pregnant women in the cluster and the number of all historical pregnant women, and screen out the target type of physical data from different types of physical data according to the overall correlation; Build a preterm birth prediction model based on the target type of physical data during different physical examinations of different historical pregnant women for preterm birth prediction.

[0007] Combined with the first aspect above, in some possible implementation manners, determining the deviation degree of different types of physical data during different physical examinations of each historical pregnant woman includes: Determine the difference value of the same type of physical data during the same physical examination according to the difference in the same type of physical data during the same physical examination between each historical pregnant woman and each other historical pregnant woman; Determine the normal delivery confidence degree of each other historical pregnant woman according to the preterm birth degree, and the greater the preterm birth degree, the smaller the normal delivery confidence degree; Determine the cumulative value of the product value of the normal delivery confidence degree and the physical data difference value between each historical pregnant woman and each other historical pregnant woman, so as to obtain the deviation degree of different types of physical data during different physical examinations of each historical pregnant woman.

[0008] Combined with the first aspect above, in some possible implementation manners, determining the preterm birth correlation degree of different types of physical data for each historical pregnant woman includes: Determine the deviation duration of different types of physical data during different physical examinations of each historical pregnant woman according to the deviation change situation of the same type of physical data during different physical examinations of each historical pregnant woman; Determine the abnormal index of different types of physical data for each historical pregnant woman according to the deviation degree and the deviation duration; Determine the preterm birth correlation index of the same type of physical data between any two historical pregnant women according to the correlation between the difference in the abnormal indexes of the same type of physical data and the difference in the preterm birth degrees among different historical pregnant women; The preterm birth related index that combines the same type of physical data between each historical pregnant woman and every other historical pregnant woman is used to determine the degree of preterm birth related to different types of physical data of each historical pregnant woman.

[0009] Combined with the first aspect above, in some possible implementation manners, determining the degree of deviation persistence of different types of physical data during different physical examinations of each historical pregnant woman, including: Determining the degree of growth of the deviation of different types of physical data during each physical examination of each historical pregnant woman according to the difference in the deviation degree of the same type of physical data between each physical examination of each historical pregnant woman and her previous physical examination; Determining the minimum value among the degrees of growth of the deviation of the same type of physical data during each physical examination of each historical pregnant woman and all subsequent physical examinations of her, to obtain the minimum degree of growth of the deviation of different types of physical data during each physical examination of each historical pregnant woman; Judging whether the minimum degree of growth of the deviation is greater than 0, and determining the degree of deviation persistence of different types of physical data during different physical examinations of each historical pregnant woman according to the judgment result, where the degree of deviation persistence is determined by the maximum value between the corresponding minimum degree of growth of the deviation and 0.

[0010] Combined with the first aspect above, in some possible implementation manners, determining the abnormality index of different types of physical data of each historical pregnant woman, including: Determining the cumulative value of the product values of the deviation degree and the degree of deviation persistence of the same type of physical data during different physical examinations of each historical pregnant woman, so as to obtain the abnormality index of different types of physical data of each historical pregnant woman.

[0011] Combined with the first aspect above, in some possible implementation manners, determining the preterm birth related index of the same type of physical data between any two historical pregnant women, including: Determining the difference in the degree of preterm birth of the same type of physical data between any two historical pregnant women to obtain a first difference value; Determining the difference in the abnormality index of the same type of physical data between any two historical pregnant women to obtain a second difference value; If the first difference value and the second difference value have different signs, setting the preterm birth related index of the same type of physical data between any two historical pregnant women to 0, otherwise setting the preterm birth related index of the same type of physical data between any two historical pregnant women to 1.

[0012] Combined with the first aspect above, in some possible implementation manners, determining the deviation time of different types of physical data of each historical pregnant woman, including: Determine the difference in the degree of deviation persistence of the same type of physical data between each physical examination of each historical pregnant woman and her previous physical examination to obtain the mutation degree; Determine the maximum value among the mutation degrees corresponding to the same type of physical data at each physical examination of each historical pregnant woman to obtain the maximum mutation degree; Determine the physical examination times corresponding to the maximum mutation degree as the deviation time of different types of physical data of each historical pregnant woman.

[0013] Combined with the above first aspect, in some possible implementation manners, determining the overall correlation of different types of physical data includes: Determine the average value of the preterm birth correlation degrees of the corresponding type of physical data of all historical pregnant women in each cluster to obtain the preterm birth correlation degree mean value corresponding to each cluster; According to the number of all historical pregnant women in each cluster and the preterm birth correlation degree mean value, determine the sub-correlation corresponding to each cluster, and both the number of all historical pregnant women and the preterm birth correlation degree mean value are positively correlated with the sub-correlation; Determine the cumulative value of the sub-correlations corresponding to all the clusters of different types of physical data, so as to obtain the overall correlation of different types of physical data.

[0014] Combined with the above first aspect, in some possible implementation manners, constructing a preterm birth prediction model for preterm birth prediction includes: Based on all the physical data of the target types at the same physical examination of different historical pregnant women, construct a decision tree by the random forest method, and use the constructed decision tree as the preterm birth prediction model at the corresponding physical examination for preterm birth prediction.

[0015] To solve the above technical problems, in a second aspect, the present invention also provides a gestational preterm birth risk auxiliary prediction system, including a memory and a processor. The memory is used to store executable program codes, and the processor is used to call and run the executable program codes from the memory, so that the device executes the method in the above first aspect or any one of the possible implementation manners of the first aspect.

[0016] To solve the above technical problems, in a third aspect, the present invention also provides a computer program product, which includes: computer program codes. When the computer program codes run on a computer, the computer executes the method in the above first aspect or any one of the possible implementation manners of the first aspect.

[0017] To solve the above technical problems, in a fourth aspect, the present invention further provides a computer-readable storage medium storing computer program code, which, when running on a computer, causes the computer to execute the method in the first aspect or any possible implementation manner of the first aspect above.

[0018] The present invention has the following beneficial effects: By obtaining the physical data and delivery cycles of different historical pregnant women, determining the degree of preterm birth of different historical pregnant women according to the delivery cycle, and then determining the deviation degree of different types of physical data during different physical examinations of each historical pregnant woman based on the degree of preterm birth and the differences in the same type of physical data during the same physical examination among different historical pregnant women. This deviation degree reflects the differences in physical data during physical examinations of preterm pregnant women compared to those of relatively normal delivery pregnant women at the same gestational time; According to the differences in the degree of preterm birth among different historical pregnant women and the differences in the changes in the deviation degrees of the same type of physical data during different physical examinations, determining the preterm-related degree of different types of physical data of each historical pregnant woman, which reflects the correlation between different types of physical data of pregnant women and the occurrence of preterm birth; Considering that the inducements for preterm birth in different pregnant women are different, determining the deviation time of different types of physical data of each historical pregnant woman based on the changes in the deviation degrees of the same type of physical data during different physical examinations of each historical pregnant woman, and clustering different historical pregnant women according to the differences in the deviation times of the same type of physical data among different historical pregnant women to obtain several clusters; Since the more pregnant women in the same cluster, the more likely the preterm pregnant women in this cluster are caused by the same inducement, this type of physical data is more representative, and the preterm-related degrees of all pregnant women in the cluster are relatively large, indicating that the overall correlation between the corresponding type of physical data and preterm birth is higher. Thus, the target type of physical data that is overall related to preterm birth can be screened out from different types of physical data; Finally, based on the target type of physical data during different physical examinations of different historical pregnant women, a preterm birth prediction model is constructed to achieve accurate preterm birth prediction. The present invention screens the physical data of different historical pregnant women to determine the target type of physical data with a higher overall correlation with preterm birth, so as to construct a more accurate preterm birth prediction model, effectively improving the accuracy of preterm birth prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 Flow chart of steps of an auxiliary prediction method for preterm birth risk in the embodiment of the present invention; Figure 2 Structural schematic diagram of an auxiliary prediction system for preterm birth risk in the embodiment of the present invention. Detailed implementation manners

[0021] To clearly illustrate the technical features of the present solution, the present invention will be elaborated in detail below through specific implementation manners in conjunction with the accompanying drawings.

[0022] Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments described herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0023] It should be understood that the steps recorded in the method embodiments of the present invention can be executed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this regard.

[0024] As used herein, the term "comprising" and its variants are open-ended, that is, "including but not limited to". The term "based on" is "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0025] It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependent relationships.

[0026] In the embodiments of the present invention, although the operations or steps are described in a specific order in the drawings, it should not be understood that these operations or steps are required to be performed in the specific order shown or in a serial order, or that all the operations or steps shown are required to be performed to obtain the desired result. In the embodiments of the present invention, these operations or steps can be executed serially; they can also be executed in parallel; or a part of these operations or steps can be executed.

[0027] Meanwhile, it can be understood that the data involved in the technical solution of the present invention (including but not limited to the data itself, the acquisition or use of data) should comply with the requirements of corresponding laws, regulations and related provisions. Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs, and all parameters or indicators in the formulas involved in the present invention are numerical values after normalization that eliminate the influence of dimensions.

[0028] To solve the problem that the accuracy of predicting preterm birth during pregnancy becomes poor due to improper selection of antenatal examination data in the prior art, an embodiment of the present invention provides a method and system for assisting in predicting the risk of preterm birth during pregnancy. By obtaining the physical data and delivery cycles of different historical pregnant women, and screening the physical data of different historical pregnant women, the physical data of the target type with a higher overall correlation with preterm birth is determined, so as to construct a more accurate preterm birth prediction model, effectively improving the accuracy of preterm birth risk prediction.

[0029] Next, a method and system for assisting in predicting the risk of preterm birth during pregnancy provided by an embodiment of the present invention will be introduced in detail with reference to the accompanying drawings.

[0030] Figure 1 The basic process schematic diagram of a method for assisting in predicting the risk of preterm birth during pregnancy provided by an embodiment of the present invention is shown, as Figure 1 shown, and the method specifically includes the following steps: Step S100: Obtain the physical data and delivery cycles of different historical pregnant women, where the physical data includes different types of physical data at each physical examination, and the pregnancy time of different historical pregnant women at the same physical examination is the same.

[0031] Specifically, in order to facilitate the subsequent prediction of the risk of preterm birth in pregnant women, it is first necessary to obtain relevant data for constructing a preterm birth prediction model. Therefore, for a number of historical pregnant women who have already given birth, physical examinations are carried out at fixed time intervals starting from the 12th week of pregnancy, such as once a month or once every two weeks, and different types of physical data of each historical pregnant woman at each physical examination are recorded. These physical data refer to the examination data that may cause preterm birth or reflect the phenomenon of preterm birth. Among them, the physical data includes but is not limited to blood pressure, blood sugar, body weight, body fat percentage, heart rate, cervical length, frequency and intensity of uterine contractions, uterine tension, hemoglobin concentration, white blood cell count, platelet count, C-reactive protein, urine protein level, fetal weight, biparietal diameter, femoral length, abdominal circumference, head circumference, fetal heart rate, and amniotic fluid index. It should be understood that since the time intervals for physical examinations of different historical pregnant women starting from the 12th week of pregnancy are the same, the gestational ages of different historical pregnant women at the same physical examination are the same. At the same time, record the conception time and delivery time of each historical pregnant woman, determine the number of weeks included in the time period from the conception time to the delivery time, and use the number of weeks included in this time period as the delivery cycle, so as to obtain the delivery cycle of each historical pregnant woman.

[0032] It should be understood that since it is to predict the risk of preterm birth in pregnant women, in order to reduce the amount of data analysis, late-term pregnant women with a delivery cycle exceeding 42 weeks from the conception time can be excluded. Thus, by selecting the physical data and delivery cycles of different historical pregnant women with a delivery cycle not exceeding 42 weeks as the analysis data, a preterm birth prediction model is finally determined and constructed.

[0033] Step S200: Determine the degree of preterm birth of different historical pregnant women according to the delivery cycle, and determine the deviation degree of different types of physical data of each historical pregnant woman at different physical examinations according to the degree of preterm birth and the differences in the same types of physical data of different historical pregnant women at the same physical examination.

[0034] Specifically, when analyzing the physical data of different historical pregnant women, it is necessary to first judge the degree of preterm birth of each historical pregnant woman, so that the physical data of historical pregnant women with a low degree of preterm birth can be used as control data for the physical data of historical pregnant women with a high degree of preterm birth for analysis. Preterm birth is usually defined as giving birth before 37 weeks of pregnancy, so the earlier the delivery time, the higher the degree of preterm birth. Thus, according to the delivery cycles of different historical pregnant women, the degree of preterm birth of different historical pregnant women can be determined.

[0035] Further, in the embodiment of the present invention, the above-mentioned method for determining the degree of preterm birth of different historical pregnant women according to the delivery cycle has the corresponding calculation formula: ; Among them, represents the The premature birth degree of the indicating the delivery cycle of the indicating the preset threshold number of weeks of premature birth, ; indicating the normalization function; indicating the maximum value function.

[0036] In the above calculation formula, by comparing the delivery cycle of the th historical pregnant woman with the preset threshold number of weeks of premature birth, the premature birth degree of the th historical pregnant woman is determined. When the delivery cycle is less than the preset threshold number of weeks of premature birth and the difference between them is greater, the value of the corresponding premature birth degree is larger.

[0037] Before a premature birth occurs in a pregnant woman with premature birth, some of her associated physical data will show corresponding persistent deviations compared with normal pregnant women. Such deviations often reflect potential physiological or pathological problems during pregnancy and are important signals of the risk of premature birth. Specifically, the degree of deviation of the pregnant woman's physical data is mainly reflected in the differences in the physical data during the physical examination at the same gestational age in the pregnancy period between the pregnant woman with premature birth and the pregnant woman with normal delivery. The greater the difference, the greater the degree of deviation of this type of physical data during this physical examination.

[0038] Furthermore, the implementation steps for determining the degree of deviation of different types of physical data during different physical examinations of each historical pregnant woman include: Step S201: Determine the difference value of the same type of physical data during the physical examination at the same gestational cycle according to the differences in the same type of physical data during the physical examination at the same gestational cycle between each historical pregnant woman and each other historical pregnant woman; Step S202: Determine the normal delivery confidence level of each other historical pregnant woman according to the premature birth degree. The greater the premature birth degree, the smaller the normal delivery confidence level; Step S203: Determine the cumulative value of the product values of the normal delivery confidence level and the physical data difference value between each historical pregnant woman and each other historical pregnant woman, so as to obtain the degree of deviation of different types of physical data during different gestational cycles of each historical pregnant woman.

[0039] For the above steps, in the embodiments of the present invention, the calculation formula corresponding to determining the degree of deviation of different types of physical data during different physical examinations of each historical pregnant woman is: ; wherein, indicating the Degree of deviation of the th type of physical data during the th physical examination of a historical pregnant woman; Indicates the total number of historical pregnant women; Indicates the degree of preterm birth of the th historical pregnant woman; Indicates the th historical pregnant woman during the th physical examination, the th type of physical data value; Indicates the th historical pregnant woman during the th physical examination, the th type of physical data value; Indicates the normalization function.

[0040] In the above calculation formula, by according to the degree of preterm birth of the th historical pregnant woman, the normal delivery confidence level of the th historical pregnant woman is determined, so as to set a greater confidence level for normal delivery pregnant women with a smaller degree of preterm birth. When the th historical pregnant woman during the th physical examination, the th type of physical data has a greater difference from the th physical examination of other historical pregnant women with a greater confidence level in the th type of physical data, then the th historical pregnant woman during the th physical examination, the th type of physical data corresponds to a higher degree of deviation.

[0041] Step S300: Determine the preterm birth correlation degree of different types of physical data of each historical pregnant woman according to the difference in the degree of preterm birth between different historical pregnant women and the difference in the change of the degree of deviation of the same type of physical data during different physical examinations.

[0042] Specifically, when a pregnant woman has a premature birth during pregnancy, the persistent deviation of her body data usually reflects the cumulative effect of potential problems during pregnancy. This deviation is not a transient fluctuation, but is caused by the gradual imbalance of certain physiological or pathological factors, such as the gradual shortening of the cervix, the gradual decline of placental function, the persistent presence of chronic inflammation or infection, etc. Since these problems usually gradually worsen in a certain trend before premature birth, the body data will show a stable and continuous deviation, that is, as the number of physical examinations increases and the time of delivery approaches, the degree of deviation of the body data of premature pregnant women will continue to increase. When there is a persistent deviation in a certain body data of a pregnant woman, it indicates that the pregnant woman is likely to have an abnormality in this body data. When the abnormal difference in the same body data between a certain pregnant woman and other pregnant women has a greater correlation with the difference in the degree of premature birth, it means that this body data of the pregnant woman is more related to the occurrence of premature birth, and thus the degree of premature birth related to this body data of the pregnant woman can be determined.

[0043] Further, the implementation steps for determining the degree of premature birth related to different types of body data of each historical pregnant woman include: Step S301: Determine the degree of deviation persistence of different types of body data during different physical examinations of each historical pregnant woman according to the change of the degree of deviation of the same type of body data during different physical examinations of each historical pregnant woman. Step S302: Determine the abnormal index of different types of body data of each historical pregnant woman according to the degree of deviation and the degree of deviation persistence. Step S303: Determine the premature birth related index of the same type of body data between any two historical pregnant women according to the correlation between the difference in the abnormal index of the same type of body data and the difference in the degree of premature birth between different historical pregnant women. Step S304: Integrate the premature birth related indexes of the same type of body data between each historical pregnant woman and each other historical pregnant woman to determine the degree of premature birth related to different types of body data of each historical pregnant woman.

[0044] Regarding the above step S301, when the degree of deviation of the same type of body data during different physical examinations of each historical pregnant woman continues to increase, and the higher the minimum degree of continuous increase of the body data after a certain physical examination, the higher the degree of deviation persistence of the body data after this physical examination.

[0045] Further, in the embodiment of the present invention, determining the degree of deviation persistence of different types of body data during different physical examinations of each historical pregnant woman in the above step S301 includes: First, based on the difference in the deviation degree of the same type of physical data between each physical examination of each historical pregnant woman and the previous physical examination, determine the degree of deviation growth of different types of physical data at each physical examination of each historical pregnant woman, that is, determine the difference in the deviation degree of the same type of physical data between each physical examination of each historical pregnant woman and the previous physical examination, and use this difference as the degree of deviation growth of the corresponding type of physical data at each physical examination of each historical pregnant woman.

[0046] Secondly, determine the minimum value among the degree of deviation growth of the same type of physical data at each physical examination of each historical pregnant woman and all subsequent physical examinations, and obtain the minimum degree of deviation growth of different types of physical data at each physical examination of each historical pregnant woman. For the sake of easy understanding, for the th historical pregnant woman at the th physical examination for the th type of physical data, determine the minimum value among all the degree of deviation growth of the th historical pregnant woman's th type of physical data from the th physical examination to the last physical examination, and use this minimum value as the minimum degree of deviation growth of the th historical pregnant woman's th type of physical data at the th physical examination.

[0047] Finally, determine whether the minimum degree of deviation growth of different types of physical data at each physical examination of each historical pregnant woman is greater than 0, and determine the deviation persistence degree of different types of physical data at different physical examinations of each historical pregnant woman according to the judgment result. The deviation persistence degree is determined by the maximum value between the corresponding minimum degree of deviation growth and 0.

[0048] In the embodiment of the present invention, to determine the deviation persistence degree of different types of physical data at different physical examinations of each historical pregnant woman, the corresponding calculation formula is: ; where represents the deviation persistence degree of the th historical pregnant woman's th type of physical data at the th physical examination; represents the minimum degree of deviation growth of the th historical pregnant woman's th type of physical data at the th physical examination, that is, the th historical pregnant woman's From the first physical examination to the last physical examination, the minimum value of the growth degree of all deviation degrees of the types of physical data; Indicates the maximum value function.

[0049] Regarding the above step S302, when the degree of continuous deviation of the physical data of a certain type with a higher deviation degree in a pregnant woman is greater, it indicates that the physical data of this type is more abnormal, the corresponding abnormal index is greater, and the degree of inducing premature birth is greater.

[0050] Furthermore, the above step S302 determines the abnormal index of different types of physical data of each historical pregnant woman, including: determining the cumulative value of the product of the deviation degree and the deviation persistence degree of the same type of physical data at different physical examinations of each historical pregnant woman, so as to obtain the abnormal index of different types of physical data of each historical pregnant woman. The corresponding calculation formula is: ; Among them, Indicates the th historical pregnant woman's abnormal index of the th type of physical data; Indicates the number of examinations of the th historical pregnant woman; Indicates the th physical examination of the th historical pregnant woman's degree of deviation persistence of the th type of physical data; Indicates the th physical examination of the

[0051] In the above calculation formula, the greater the deviation persistence degree and the deviation degree of the th historical pregnant woman's th type of physical data at different physical examinations, it indicates that the th historical pregnant woman's th type of physical data is more likely to be abnormal, and its corresponding abnormal index is greater.

[0052] Regarding the above step S303, the greater the degree of deviation of a certain type of physical data related to preterm birth from the normal level for a pregnant woman and the longer the duration of continuous deviation, the earlier the preterm birth usually occurs. This is because severe and long-term deviations usually reflect more serious physiological or pathological problems. For example, a significant shortening of the cervical length may indicate cervical incompetence, and a continuous decline in placental function may lead to insufficient fetal oxygen supply or intrauterine growth restriction. When these problems accumulate to a certain extent, they will trigger uterine contractions, premature rupture of membranes or other preterm birth inducing factors, forcing the delivery to occur earlier. Therefore, the amplitude and duration of the deviation of the pregnant woman's physical data will be directly related to the exacerbation of the preterm birth risk and the time of occurrence, that is, for all pregnant women, the greater the degree of correlation between the abnormal index of a certain physical data and the degree of preterm birth of the pregnant woman, the greater the degree of correlation of this physical data with preterm birth.

[0053] Further, the above step S303 determines the preterm birth related index of the same type of physical data between any two historical pregnant women, including: determining the difference in the degree of preterm birth of the same type of physical data between any two historical pregnant women to obtain a first difference value; determining the difference in the abnormal index of the same type of physical data between any two historical pregnant women to obtain a second difference value; if the first difference value and the second difference value have different signs, then set the preterm birth related index of the same type of physical data between any two historical pregnant women to 0, otherwise set the preterm birth related index of the same type of physical data between any two historical pregnant women to 1.

[0054] In the embodiment of the present invention, the above step S303 determines the preterm birth related index of the same type of physical data between any two historical pregnant women, and the corresponding calculation formula is: ; where represents the preterm birth related index of the th historical pregnant woman and the th historical pregnant woman for the th type of physical data; represents the abnormal index of the th historical pregnant woman for the th type of physical data; represents the abnormal index of the th historical pregnant woman for the th type of physical data; represents the degree of preterm birth of the th historical pregnant woman; represents the degree of preterm birth of the th historical pregnant woman.

[0055] In the above calculation formula, if the th historical pregnant woman and the th historical pregnant woman for the If the preterm birth related index of a certain type of body data is 1, it indicates that the th historical pregnant woman and the th historical pregnant woman have a type of body data that is correlated with the degree of preterm birth; if the th historical pregnant woman and the th historical pregnant woman have a preterm birth related index of 0 for the type of body data, it indicates that the th historical pregnant woman and the th historical pregnant woman have a type of body data that is not correlated with the degree of preterm birth.

[0056] Regarding the above step S304, when a certain historical pregnant woman and other historical pregnant women show a correlation in a certain type of body data, it indicates that the preterm birth related degree of this type of body data of this historical pregnant woman is greater. Therefore, by fusing the preterm birth related indexes of the same type of body data between each historical pregnant woman and each other historical pregnant woman, the preterm birth related degree of different types of body data of each historical pregnant woman can be determined.

[0057] Furthermore, in the embodiment of the present invention, the above step S304 determines the preterm birth related degree of different types of body data of each historical pregnant woman, and the corresponding calculation formula is: ; where represents the preterm birth related degree of the th historical pregnant woman for the type of body data; represents the total number of historical pregnant women; represents the th historical pregnant woman and the th historical pregnant woman for the type of body data of the preterm birth related index.

[0058] Step S400: According to the change situation of the deviation degree of the same type of body data of different physical examinations of each historical pregnant woman, determine the deviation time of different types of body data of each historical pregnant woman, and cluster different historical pregnant women according to the difference in the deviation time of the same type of body data between different historical pregnant women, to obtain several clusters.

[0059] Specifically, since the inducements for premature birth vary among different pregnant women, the physical data with the same degree of correlation to premature birth will also differ. For example, a relatively high degree of correlation to premature birth in the physical data of the c-th type may only manifest in some patients with premature birth caused by the same inducement. When a specific inducement for premature birth leads to premature birth, the premature pregnant women related to this type of inducement usually have similar abnormal stages in certain relevant physical data. This is because the same inducement often follows a similar pathological mechanism. For example, pregnant women with cervical incompetence may show a significant shortening of the cervical length in the second trimester of pregnancy, while pregnant women with placental dysfunction may exhibit fetal growth restriction or abnormal amniotic fluid volume in the third trimester. This consistency in the stages of the pathological mechanism makes the time periods when relevant physical data become abnormal highly similar.

[0060] Based on the above analysis, according to the change in the deviation degree of the same type of physical data in different physical examinations of each historical pregnant woman, the deviation persistence degree of different types of physical data in different physical examinations of each historical pregnant woman can be determined. The process of obtaining this deviation persistence degree has been introduced in detail in step S301 above and will not be elaborated here. When a certain type of physical data of a pregnant woman shows continuous deviation, it indicates that the physical data of this type of the pregnant woman has become abnormal. When the deviation persistence degree of a certain type of physical data in a physical examination of a pregnant woman suddenly increases, it indicates that this physical examination reflects an abnormality in this type of physical data. Thus, the deviation time of the pregnant woman corresponding to this type of physical data can be determined. Pregnant women with more similar deviation times in a certain type of physical data are more likely to have similar inducements for premature birth. If their degrees of correlation to premature birth are relatively high, it indicates that the overall correlation between this type of physical data and premature birth is higher.

[0061] Furthermore, the steps to determine the deviation time of different types of physical data of each historical pregnant woman include: Step S401: Determine the difference in the deviation persistence degree of the same type of physical data between each physical examination of each historical pregnant woman and the previous physical examination to obtain the mutation degree; Step S402: Determine the maximum value among the mutation degrees corresponding to the same type of physical data in each physical examination of each historical pregnant woman to obtain the maximum mutation degree; Step S403: Determine the physical examination number corresponding to the maximum mutation degree as the deviation time of different types of physical data of each historical pregnant woman.

[0062] Regarding the above steps, for any historical pregnant woman, at any physical examination, the deviation persistence degree of any type of physical data and the deviation persistence degree of this same arbitrary The difference in the persistence degree of deviation of a certain type of body data is denoted as the arbitrary mutation degree of a certain type of body data at any physical examination, and determine the arbitrary maximum value of the mutation degree of a certain type of body data at all physical examinations to obtain the maximum mutation degree. Record the physical examination times corresponding to the maximum mutation degree as the deviation time of a certain type of body data for any historical pregnant woman at the th type of body data.

[0063] According to the differences in the deviation times of the same type of body data among different historical pregnant women, cluster different historical pregnant women to obtain several clusters. The more historical pregnant women in a cluster, the more likely it is that the preterm pregnant women in this cluster are caused by the same incentive. The fewer historical pregnant women in a cluster, the more random the deviation of the historical pregnant women in this cluster in the corresponding type of body data, and it is not representative.

[0064] Furthermore, in the embodiments of the present invention, determine the absolute value of the difference in the deviation times of the same type of body data between any two different historical pregnant women, and use this absolute value of the difference as the distance metric between different historical pregnant women. Based on this distance metric, use the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm to cluster all historical pregnant women, thereby obtaining several clusters.

[0065] Step S500: Determine the overall correlation of different types of body data according to the preterm-related degree of the corresponding type of body data of all historical pregnant women in the cluster and the number of all historical pregnant women, and screen out the target type of body data from different types of body data according to the overall correlation.

[0066] Specifically, for several clusters corresponding to any type of body data, since the more pregnant women in the same cluster, the more likely it is that the preterm pregnant women in this cluster are caused by the same incentive, this type of body data is more representative, and the preterm-related degree of all pregnant women in the cluster is relatively large, indicating that the overall correlation between the corresponding type of body data and preterm birth is higher. At this time, this type of body data can provide more basis for preterm birth prediction.

[0067] Furthermore, the implementation steps of determining the overall correlation of different types of body data include: Step S501: Determine the average value of the preterm-related degree of the corresponding type of body data of all historical pregnant women in each cluster to obtain the preterm-related degree mean value corresponding to each cluster; Step S502: Determine the sub - correlation corresponding to each cluster class according to the number of all historical pregnant women in each cluster class and the mean value of the preterm - birth correlation degree. Both the number of all historical pregnant women and the mean value of the preterm - birth correlation degree are positively correlated with the sub - correlation; Step S503: Determine the cumulative value of the sub - correlations corresponding to all the cluster classes of different types of body data, so as to obtain the overall correlation of different types of body data.

[0068] For the above steps, in the embodiment of the present invention, the calculation formula for determining the overall correlation of different types of body data is as follows: ; where, represents the overall correlation of the th type of body data; represents the total number of cluster classes of the th type of body data; represents the number of historical pregnant women in the th cluster class of the th type of body data; represents the mean value of the preterm - birth correlation degree of the th type of body data of all historical pregnant women in the th cluster class of the th type of body data; represents the normalization function.

[0069] In the above calculation formula, by normalizing the number of historical pregnant women in each cluster class of the th type of body data as the weight of the mean value of the preterm - birth correlation degree of the th type of body data of all historical pregnant women in each cluster class, and by accumulating the sub - correlations obtained by weighting the mean value of the preterm - birth correlation degree for each cluster class, the overall correlation of the th type of body data is obtained. The greater the overall correlation, the greater the possibility that the th type of body data is related to preterm birth, and the more it should be used as reference data for preterm - birth prediction.

[0070] A correlation threshold is preset. The specific value of the correlation threshold can be reasonably selected according to needs. In the embodiment of the present invention, the value of the correlation threshold is set to 0.7. Compare the overall correlation of different types of body data with this correlation threshold. If the overall correlation is greater than this correlation threshold, the corresponding type of body data is used as the target - type body data. Thus, according to the overall correlation, the screening of all target - type body data among different types of body data can be realized.

[0071] Step S600: Based on the physical data of target types during different physical examinations of different historical pregnant women, construct a preterm birth prediction model to predict preterm birth.

[0072] Specifically, based on the physical data of all target types during the same physical examination of all historical pregnant women, and combined with the delivery cycles of all historical pregnant women, use random forest to construct decision trees for each physical examination. The constructed decision trees are the preterm birth prediction models for each physical examination. Since the specific implementation process of constructing decision trees belongs to the prior art, it will not be elaborated here.

[0073] When it is necessary to predict preterm birth for a certain target pregnant woman, obtain the physical data of all target types of this target pregnant woman, and input these physical data into the preterm birth prediction model, i.e., the decision tree, corresponding to the number of physical examinations. The decision tree performs preterm birth prediction, and finally obtains the corresponding preterm birth prediction result. Since the specific implementation process of obtaining the prediction result by using the decision tree also belongs to the prior art, it will not be elaborated here. It should be understood that since the gestational time of the target pregnant woman can be determined, and each physical examination corresponds to a gestational time, the preterm birth prediction model corresponding to the number of physical examinations of the target pregnant woman can be determined.

[0074] Based on the same inventive concept, an embodiment of the present invention also provides a system for assisting in predicting the risk of preterm birth during pregnancy, as Figure 2 shown. The prediction system includes: a memory 201, a processor 202, and a computer program 203 stored in the memory 201 and running on the processor 202. When the processor 202 executes the computer program 203, the system can execute any one of the methods for assisting in predicting the risk of preterm birth during pregnancy described above.

[0075] An embodiment of the present invention can divide the functions of the system according to the above method examples. For example, each function module can be corresponded, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware. It should be noted that the division of modules in this embodiment is illustrative, only a logical function division, and there can be other division methods in actual implementation.

[0076] Based on the same inventive concept, an embodiment of the present invention also provides a computer program product, which includes: computer program code. When the computer program code runs on a computer, the computer executes any one of the methods for assisting in predicting the risk of preterm birth during pregnancy described above.

[0077] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium, which stores computer program code. When the computer program code runs on a computer, it causes the computer to execute any one of the foregoing assisted prediction methods for preterm birth risk during pregnancy.

[0078] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for auxiliary prediction of risk of premature birth during pregnancy, characterized in that: The following steps are involved: Acquiring physical data and delivery cycles of pregnant women with different histories, wherein the physical data includes different types of physical data during each physical examination, and the gestational time of pregnant women with different histories during the same physical examination is the same; Determine the degree of premature birth of different historical pregnant women according to the delivery cycle, and determine the deviation of different types of physical data during different physical examinations of each historical pregnant woman according to the degree of premature birth and the difference of the same type of physical data during the same physical examination between different historical pregnant women; Determine the degree of relevance of different types of physical data of each historical pregnant woman to premature birth according to the differences in the degree of premature birth between different historical pregnant women and the differences in the changes in the deviations of the same type of physical data during different physical examinations; According to the change of the deviation degree of the same type of physical data during different physical examinations of each historical pregnant woman, the deviation time of different types of physical data of each historical pregnant woman is determined, and according to the difference of the deviation time of the same type of physical data between different historical pregnant women, different historical pregnant women are clustered to obtain a number of clusters; Determine the overall correlation of different types of body data according to the degree of relevance to premature birth of the corresponding types of body data of all historical pregnant women in the cluster and the number of all historical pregnant women, and screen out target types of body data from the different types of body data according to the overall correlation; Based on the target type of physical data of different historical pregnant women at different physical examinations, a premature birth prediction model is constructed to predict premature birth.

2. The method for auxiliary prediction of premature birth risk during pregnancy according to claim 1, characterized in that: Determine the deviation of different types of physical data at different physical examinations for each historical pregnant woman, including: Determine the difference value of the same type of physical data at the same physical examination according to the difference of the same type of physical data at the same physical examination between each historical pregnant woman and each other historical pregnant woman; Determine the normal delivery confidence of each other historical pregnant woman according to the degree of premature delivery, wherein the greater the degree of premature delivery, the smaller the normal delivery confidence; The cumulative value of the product of the normal delivery confidence and the physical data difference value between each historical pregnant woman and each other historical pregnant woman is determined, so as to obtain the deviation of different types of physical data of each historical pregnant woman at different physical examinations.

3. The method for auxiliary prediction of premature birth risk during pregnancy according to claim 1, characterized in that: The degree to which different types of physical data were associated with preterm birth was determined for each historical pregnancy, including: Determine the persistence of deviation of different types of physical data during different physical examinations for each historical pregnant woman according to the change of deviation of the same type of physical data during different physical examinations for each historical pregnant woman; Determining abnormal indexes of different types of physical data of each historical pregnant woman according to the degree of deviation and the degree of persistence of the deviation; Determine the preterm birth related index of the same type of physical data between any two historical pregnant women according to the correlation between the difference in abnormal index of the same type of physical data between different historical pregnant women and the difference in the degree of preterm birth; The preterm birth correlation index of the same type of physical data between each historical pregnant woman and each other historical pregnant woman is integrated to determine the preterm birth correlation degree of different types of physical data of each historical pregnant woman.

4. The method for auxiliary prediction of risk of premature birth during pregnancy according to claim 3, characterized in that: Determine the degree of persistence of deviations in different types of physical data at different physical examinations for each historical pregnant woman, including: Determine the degree of increase of the deviation of different types of physical data during each physical examination of each historical pregnant woman according to the difference in the deviation of the same type of physical data between each physical examination of each historical pregnant woman and her previous physical examination; Determine the minimum value of the deviation growth degree of the same type of physical data during each physical examination of each historical pregnant woman and all subsequent physical examinations, and obtain the minimum deviation growth degree of different types of physical data during each physical examination of each historical pregnant woman; Determine whether the minimum deviation growth rate is greater than 0, and determine the deviation persistence of different types of physical data during different physical examinations of each historical pregnant woman based on the judgment result, wherein the deviation persistence is determined by the maximum value between the minimum deviation growth rate and 0.

5. The method for auxiliary prediction of risk of premature birth during pregnancy according to claim 3, characterized in that: Determine the abnormal index of different types of physical data for each historical pregnant woman, including: The cumulative value of the product value of the deviation degree and the deviation persistence degree of the same type of physical data at different physical examinations of each historical pregnant woman is determined, so as to obtain the abnormal index of different types of physical data of each historical pregnant woman.

6. The method for auxiliary prediction of risk of premature birth during pregnancy according to claim 3, characterized in that: Determine the preterm birth-related index between any two historical pregnant women with the same type of physical data, including: Determine the difference in the degree of premature birth of the same type of physical data between any two historical pregnant women to obtain a first difference value; Determine the difference in abnormality index of the same type of physical data between any two historical pregnant women to obtain a second difference value; If the first difference value and the second difference value have different signs, the premature birth correlation index of the same type of physical data between any two historical pregnant women is set to 0; otherwise, the premature birth correlation index of the same type of physical data between any two historical pregnant women is set to 1.

7. The method for auxiliary prediction of risk of premature birth during pregnancy according to claim 3, characterized in that: Determine the deviation time of different types of physical data for each historical pregnant woman, including: Determine the difference in the degree of deviation of the same type of physical data between each physical examination of each historical pregnant woman and the previous physical examination, and obtain the mutation degree; Determine the maximum value of the mutation degree corresponding to the same type of physical data during each physical examination of each historical pregnant woman to obtain the maximum mutation degree; The number of physical examinations corresponding to the maximum mutation degree is determined as the deviation time of different types of physical data of each historical pregnant woman.

8. The method for auxiliary prediction of risk of premature birth during pregnancy according to claim 1, characterized in that: Determine the overall relevance of different types of physical data, including: Determine the average value of the degree of relevance of premature birth of the corresponding type of physical data of all historical pregnant women in each of the clusters, and obtain the mean value of the degree of relevance of premature birth corresponding to each of the clusters; Determine the sub-correlation corresponding to each cluster according to the number of all historical pregnant women and the mean value of the premature birth correlation degree in each cluster, wherein the number of all historical pregnant women and the mean value of the premature birth correlation degree are both positively correlated with the sub-correlation; The accumulated values ​​of the sub-correlations corresponding to all the clusters of different types of body data are determined, thereby obtaining the overall correlations of different types of body data.

9. The method for auxiliary prediction of risk of premature birth during pregnancy according to claim 1, characterized in that: Construct a preterm birth prediction model to predict preterm birth, including: Based on the physical data of all target types during the same physical examination of different historical pregnant women, a decision tree is constructed by the random forest method, and the constructed decision tree is used as the premature birth prediction model during the corresponding physical examination to predict premature birth.

10. A system for assisting in predicting the risk of premature birth during pregnancy, characterized in that: It comprises a memory, a processor and an executable computer program stored in the memory and runnable on the processor, wherein when the processor executes the computer program, it executes a method for auxiliary prediction of the risk of premature birth during pregnancy as described in any one of claims 1 to 9.

Images