Optimization system for predicting cumulative pregnancy rate of assisted reproductive technology

Abstract

The invention relates to an optimization system for predicting the cumulative pregnancy rate of an assisted reproductive technology. Reproductive health information of the patient is accumulated through multi-cycle data collection; based on the multi-cycle data, constructing an initial cumulative pregnancy rate prediction model; after each new treatment cycle is finished, comparing a cycle success or failure result with the initial prediction, and inputting new data into the model; updating the predicted value of the accumulated pregnancy rate through a dynamic adjustment mechanism; generating a staged cumulative pregnancy rate curve according to the dynamic prediction result of each period; intelligent decision support is provided in combination with the change trend of the curve; the key risk point of the next period is predicted by analyzing the inflection point and the ascending or descending trend of the curve, and personalized treatment suggestions are provided for doctors and patients, including medicine dosage adjustment or treatment scheme change. And by performing multi-dimensional analysis on the prediction result of the accumulated pregnancy rate, key factors influencing success of pregnancy, including patient age, embryo quality and effectiveness of a treatment scheme, are identified.

Description

Technical Field

[0001] The present invention relates to the field of assisted reproductive technology, and in particular to an optimization system for predicting the cumulative pregnancy rate of assisted reproductive technology. Background Art

[0002] Existing methods for predicting the cumulative pregnancy rate of assisted reproductive technology, such as the pregnancy success rate prediction method described in Chinese invention patent publication number CN116884620A, mainly obtain the biological characteristics of the test object, including physical characteristics and hormone characteristics, and then input these biological characteristics into the pregnancy success rate prediction model, and output the pregnancy success rate of the test object through the model. The model is trained by the correspondence between the sample biological characteristics and the pregnancy category (successful or failed pregnancy). However, this method has some significant deficiencies and drawbacks in practical applications, which affects its wide application and prediction effect in assisted reproductive technology.

[0003] First, this method mainly relies on physical characteristics and hormone characteristics as key input data for predicting pregnancy success rate, but these characteristics only cover part of the patient's physiological state and ignore other important multidimensional factors, such as the patient's psychological state, living habits, external environmental factors, and previous treatment responses. These factors have an important impact on the pregnancy success rate during assisted reproduction. Due to the lack of comprehensive consideration of these multidimensional factors, the existing methods cannot fully capture the complexity and diversity of factors affecting the pregnancy success rate, resulting in the accuracy and personalization of the prediction results being limited. For patient groups with obvious individual differences, a single physiological characteristic analysis often cannot provide accurate predictions. In addition, this method lacks the ability to dynamically capture and adjust real-time data. The assisted reproductive treatment process is a dynamic and continuous process, and the patient's physiological state and treatment response will continue to change during each treatment cycle. The existing pregnancy success rate prediction method fails to make full use of real-time data for dynamic adjustment of the model, and only relies on periodically collected physical characteristics and hormone characteristics for prediction, which makes it difficult for the model to reflect the actual changes in the patient's pregnancy success rate throughout the treatment process. The lack of the ability to respond to real-time data makes it impossible for existing models to be optimized in real time during the treatment process, and it is impossible to provide doctors with timely and personalized treatment recommendations, which affects the effectiveness and success rate of assisted reproductive treatment. Another significant shortcoming of existing methods is the lack of analysis and optimization of the cumulative pregnancy rate curve. The change in pregnancy rate is a cumulative process, which not only depends on the physiological characteristics of the current cycle, but is also affected by the results of previous cycles and the overall trend. The existing model fails to dynamically analyze and predict the cumulative pregnancy rate curve as a whole, and only focuses on the pregnancy results of a single cycle.

[0004] This prediction method that lacks a cumulative perspective cannot accurately reflect the patient's cumulative pregnancy success rate in multiple treatment cycles, which can easily lead to one-sidedness and limitations in the prediction results. The analysis of the cumulative pregnancy rate curve can not only help doctors understand the patient's long-term treatment trend, but also identify the risk points that may occur during the treatment process, and guide doctors to make more scientific decisions in each cycle. Finally, the existing pregnancy success rate prediction method lacks intelligent and personalized support in risk warning and treatment adjustment. Although the existing method can output the predicted value of the pregnancy success rate, it is unable to intelligently analyze and warn of changes in the predicted value and potential risks. During assisted reproductive treatment, the patient's physiological state and treatment effect may change at any time. Failure to identify and warn these changes in time may cause the treatment strategy to lag and affect the final treatment success rate. Summary of the invention

[0005] The purpose of the present invention is to provide an optimized system for predicting the cumulative pregnancy rate of assisted reproductive technology, thereby solving some of the drawbacks and shortcomings pointed out in the background technology.

[0006] The present invention solves the above-mentioned technical problems by adopting the following technical solutions: an optimization system for predicting cumulative pregnancy rate of assisted reproductive technology, S1, multi-cycle data accumulation and initial prediction modeling;

[0007] S1.1. First, accumulate the patient's reproductive health information through multi-cycle data collection, including follicle development, embryo quality, and endometrial thickness during each assisted reproductive treatment;

[0008] S1.2. Construct an initial cumulative pregnancy rate prediction model based on multi-cycle data;

[0009] S2. Dynamic prediction and adjustment of cumulative pregnancy rate;

[0010] S2.1. After each new treatment cycle, the success or failure of the cycle is compared with the initial prediction and the new data is input into the model.

[0011] S2.2. Update the predicted value of cumulative pregnancy rate through a dynamic adjustment mechanism;

[0012] S3, generation of phased cumulative pregnancy rate curve;

[0013] S3.1. Generate a cumulative pregnancy rate curve for each cycle based on the dynamic prediction results;

[0014] S3.2, the cumulative pregnancy rate curve is generated based on the dynamic prediction adjustment mechanism of S2, and is a graphical quantitative expression of the overall prediction;

[0015] S4. Intelligent decision-making based on curve changes;

[0016] S4.1. Combine the changing trend of the cumulative pregnancy rate curve to provide intelligent decision support;

[0017] S4.2. By analyzing the inflection points, upward or downward trends of the curve, key risk points of the next cycle can be predicted, and personalized treatment recommendations can be provided to doctors and patients, including adjusting drug dosages or changing treatment plans.

[0018] S5. Analysis of multidimensional influencing factors for cumulative pregnancy rate prediction;

[0019] S5.1. Through multi-dimensional analysis of the cumulative pregnancy rate prediction results, key factors affecting pregnancy success can be identified, including patient age, embryo quality, and effectiveness of treatment plans.

[0020] Furthermore, the cumulative pregnancy rate prediction model construction includes:

[0021] S1. First, integrate the patient's physiological indicators in each treatment cycle, including follicle development and embryo quality, as well as multidimensional factors including the patient's psychological state, living habits and external environment; multidimensional integration is expressed by the formula:

[0022] Φ(x,t)=∫0 T [x(t)·exp(-∫0 t γ(τ)dτ)+η(t)·cos(∫0 t δ(τ)dτ)]dt

[0023] Among them, x(t) represents the multidimensional data collected in each cycle, γ(τ) reflects the attenuation effect of time on different data items, and δ(τ) is used to capture the fluctuation characteristics of periodic data;

[0024] S2. Adjust the weight of each cycle data in the model according to the pregnancy results of each cycle, including the success rate of embryo implantation and the stability of pregnancy maintenance; the formula is used to express it:

[0025]

[0026] Among them, ω i Represents the weight value of each cycle, y i is the pregnancy outcome data of this cycle; by sin(y i ) captures the nonlinear characteristics of the data, The term reflects the rate of change of periodic data, combined with exp(-λy i )’s weight adjustment;

[0027] S3. Finally, the personalized adjustment factor is incorporated; the patient's physiological basic data and previous treatment response are combined; the accuracy of personalized prediction is improved when predicting the cumulative pregnancy rate through the Θ(z) function:

[0028]

[0029] Among them, z(t) represents the patient's personalized characteristic variables, the exp(-κt) term is used to simulate the changing trend of these characteristics over time, and cos(βt) is combined with The items reflect the dynamic changes of individual differences of patients during the treatment process.

[0030] Furthermore, the method used to update the predicted value of the cumulative pregnancy rate includes:

[0031] S1. First, the prediction model results are compared and adjusted according to the new periodic data through the real-time feedback mechanism. The adaptive adjustment function Φ(t) is used to dynamically update the model according to the real-time data:

[0032]

[0033] Among them, α(t) represents the time series data of pregnancy outcomes, β(t) is the adjustment factor, and γ(t) reflects the dynamic characteristics of outcome changes;

[0034] S2. Next, according to the actual situation of the cycle, the weights of each data feature in the model are readjusted; and the weights are optimized according to the changes in the cycle results; the dynamic adjustment mechanism uses the weight adjustment function Ψ(ω i ,y i ) to achieve:

[0035]

[0036] Among them, ω i Represents the weight value of each cycle, y i is the pregnancy outcome data of this cycle; i ) captures the periodic variation characteristics of the data. It reflects the dynamic adjustment process of weights, combined with exp(-λy i )\ exponential decay;

[0037] S3. Finally, after the prediction model adjusts the overall weight, it dynamically adjusts the personalized prediction path according to the patient's specific response and treatment results; through the personalized path adjustment function Θ(z), a more personalized cumulative pregnancy rate prediction is provided after each data update:

[0038]

[0039] Among them, z(t) represents the patient's personalized characteristic variable, exp(-κt) is used to reflect the changing trend of the characteristic over time, and It is used to capture the dynamic adjustment characteristics of patient responses.

[0040] Furthermore, the cumulative pregnancy rate curve generation and construction includes:

[0041] S1. Guide the new data of each cycle to influence the prediction results of the model through real-time capture mechanism; through adaptive capture and input function Φ(x,t):

[0042]

[0043] Among them, x(t) represents the data input of each period, α controls the decay of data in time, and β is used to adjust the volatility of data;

[0044] S2. According to the pregnancy results of each cycle, the weight of each cycle data and the overall prediction parameters in the model are adjusted in real time; by adaptively adjusting the function Ψ(ω i ,y i ) enables the model to optimize the prediction process:

[0045]

[0046] Among them, ω i Represents the weight of each cycle, y i For pregnancy outcome data, γ and λ regulate the volatility of cycle data and the dynamic adjustment of weights, respectively;

[0047] S3. Quantify the contribution and trend of each cycle to the overall pregnancy rate during treatment by graphical means; use the curve generating function Θ(z):

[0048]

[0049] Where z(t) represents the patient's personalized data, κ is used to control the attenuation of personalized data, and δ adjusts the dynamic change characteristics of the curve.

[0050] Furthermore, the personalized treatment suggestion construction framework comprises the following steps:

[0051] S1. First, by dynamically tracking and analyzing the inflection points, rising or falling trends of the cumulative pregnancy rate curve, the key risk points predicted to occur during the treatment process are identified in real time; the identification process is based on the slope change, curvature and fluctuation amplitude of the curve to monitor sudden changes in the effect of the treatment plan or nonlinear changes in the patient's physiological state;

[0052] S2. Secondly, combining the patient's individual historical data with the curve morphology characteristics including smoothness, amplitude and frequency, a customized risk assessment is used to determine the root cause of the curve change;

[0053] S3. Finally, based on the risk assessment results, intelligent algorithms are used to generate personalized treatment recommendations, including adjusting drug dosage, optimizing treatment rhythm, or adjusting lifestyle;

[0054] Among them, the intelligent algorithm analyzes the effects of different treatment strategies in similar medical records, combines the current patient's curve trend and risk assessment results, and provides multiple options for doctors to choose from.

[0055] Furthermore, the cumulative pregnancy rate curve control method comprises:

[0056] S1. By real-time monitoring of various data of patients during assisted reproductive treatment, including hormone levels, embryo development, and ovarian response, a cumulative pregnancy rate curve is dynamically generated; the curve is controlled by an adaptive acquisition function Φ(x,t):

[0057]

[0058] Among them, x(t) represents the key physiological data collected in real time during each treatment cycle, α is the time decay coefficient, which is used to capture the decay trend of data over time; β is used to adjust the fluctuation characteristics of physiological data in the short term;

[0059] S2. Then, an intelligent algorithm is used to deeply analyze the inflection point, slope change, curvature and fluctuation range of the cumulative pregnancy rate curve, identify and predict key risk points; through the curve analysis function Ψ(ω i ,y i ) to calculate:

[0060]

[0061] Among them, ω i represents the importance weight of each cycle, γ is used to capture the periodic fluctuation of the patient's physiological data in the curve, and λ regulates the dynamic adjustment of the weight;

[0062] S3. Finally, after identifying the key risk points, combined with the patient's individualized physiological data and curve change characteristics, the adaptive risk warning and treatment adjustment function Θ(z) is used to generate real-time risk warning and personalized treatment recommendations:

[0063]

[0064] Among them, z(t) represents the patient-specific physiological and treatment data, κ is used to adjust the sensitivity of risk warning, and δ controls the response speed of treatment adjustment suggestions.

[0065] Furthermore, the method used to determine the root cause of the curve change is:

[0066] S1. A multidimensional data model is constructed by integrating the patient's individualized historical data including previous pregnancy history, hormone levels, and ovarian response with the real-time generated cumulative pregnancy rate curve including the morphological characteristics of smoothness, amplitude, and frequency; it is described by the function Φ(x, t):

[0067]

[0068] Among them, x(t) represents the key physiological data collected in real time during each treatment cycle, α controls the decay characteristics of the data over time, and β is used to capture the fluctuation characteristics of real-time data;

[0069] S2. Secondly, the root cause of the curve change is determined by analyzing the morphological characteristics of the cumulative pregnancy rate curve, including smoothness, amplitude and frequency. The smoothness of the curve reflects the stability of the treatment effect, the amplitude represents the intensity of the treatment response, and the frequency reveals the periodic fluctuation of the physiological response. i ,y i ) These morphological features were evaluated:

[0070]

[0071] Among them, ω i Represents the weights of different morphological features, γ is used to capture the periodic changes of the curve, and λ adjusts the dynamic changes of the weights;

[0072] S3. Finally, the intelligent algorithm is used to determine the cause of the change in the cumulative pregnancy rate curve. Combined with individual historical data and real-time curve characteristics, the function Θ(z) is used for in-depth analysis:

[0073]

[0074] Here, z(t) represents the patient-specific physiological and treatment data, and κ and δ are used to adjust the response speed and sensitivity of intelligent analysis.

[0075] Beneficial effects of the present invention:

[0076] The present invention integrates the patient's individualized historical data and real-time physiological data, and combines intelligent algorithms to analyze the morphological characteristics of the cumulative pregnancy rate curve, which can accurately capture the dynamic changes of each treatment cycle. Through adaptive adjustment and personalized path optimization, this method can provide highly personalized pregnancy rate prediction results, helping doctors to formulate the most appropriate treatment plan according to the specific situation of each patient, thereby significantly improving the success rate of pregnancy.

[0077] The present invention monitors and analyzes various physiological indicators of patients during treatment in real time, and adjusts model parameters in real time according to changes in the cumulative pregnancy rate curve. With the help of adaptive acquisition and adjustment functions, the system can quickly respond to new data changes in each cycle and provide doctors with the latest risk warnings and personalized treatment recommendations. This real-time dynamic adjustment capability ensures the flexibility and timeliness of the treatment plan and effectively reduces the risks during the treatment process. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 This is a flow chart of the optimized system for predicting the cumulative pregnancy rate of the assisted reproductive technology of the present invention.

[0079] Figure 2 A flow chart is constructed for the cumulative pregnancy rate prediction model of the present invention.

[0080] Figure 3 A flow chart of the method used to update the predicted value of cumulative pregnancy rate in the present invention.

[0081] Figure 4 A flow chart was constructed for the generation of the cumulative pregnancy rate curve of the present invention. DETAILED DESCRIPTION

[0082] The specific implementation modes of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0083] Combined with Figure 1 In the optimization system for predicting the cumulative pregnancy rate of assisted reproductive technology of the present invention, the first step is the accumulation of multi-cycle data and initial prediction modeling. First, the patient's reproductive health information is accumulated through data collection of multiple cycles, including key indicators in each assisted reproductive treatment, such as follicle development, embryo quality and endometrial thickness. The accumulation of these data is the basis for predicting the cumulative pregnancy rate. They can provide detailed performance of patients in different treatment cycles and provide sufficient samples and information for the construction of the model. Then, based on the multi-cycle data collected in the early stage, the initial cumulative pregnancy rate prediction model is constructed. This model integrates all historical data to identify the key factors affecting the cumulative pregnancy rate, and converts these factors into parameters that can predict the pregnancy rate of future cycles through mathematical modeling. The construction of the initial model is the core of the cumulative pregnancy rate prediction, which aims to provide a preliminary prediction framework through the integration and analysis of multi-cycle data, laying the foundation for subsequent dynamic adjustment and optimization.

[0084] The second step involves dynamic adjustment of the cumulative prediction of pregnancy rate. After each new treatment cycle, the system compares the success or failure results of that cycle with the results of the initial prediction model. This comparison is to evaluate the accuracy of the model and identify where the model needs to be adjusted. New treatment data will be input into the model to further enrich the data basis of the model so that the model can better reflect the current treatment progress. Then based on the newly input data, the system updates the predicted value of the cumulative pregnancy rate through a dynamic adjustment mechanism. The dynamic adjustment mechanism can adaptively change the parameter settings of the model, so that the model can flexibly respond to new cycle results, thereby providing more accurate and personalized cumulative pregnancy rate predictions.

[0085] The third step is to generate the staged cumulative pregnancy rate curve. First, the system generates a staged cumulative pregnancy rate curve based on the dynamic prediction results after each treatment cycle. This curve can reflect the cumulative pregnancy success rate in the current treatment cycle and is continuously updated as new data is added. This staged curve generation helps to clearly show the patient's pregnancy success trend throughout the treatment process, providing an intuitive reference for doctors and patients. The generation of the cumulative pregnancy rate curve is based on the dynamic prediction adjustment mechanism in the second step mentioned above. This means that each update of the curve takes into account the new cycle results and their impact on the overall prediction. Therefore, this cumulative pregnancy rate curve is not only a simple superposition of the results of each cycle, but also quantifies the data through a dynamic adjustment mechanism, so that the curve can accurately reflect the current predicted status. Finally, this curve generation method can quantify the prediction results in a graphical way, providing doctors with more intuitive decision support.

[0086] The fourth step involves intelligent decision-making based on curve changes. First, the system combines the changing trend of the cumulative pregnancy rate curve to provide intelligent decision support. This process analyzes the trend of the cumulative pregnancy rate curve and identifies potential changing trends, including the smoothness of the curve, slope changes, etc., to help doctors better understand the patient's current treatment status and make more accurate judgments based on this. Through intelligent decision support, doctors can detect potential problems at an early stage and adjust treatment strategies in a timely manner to ensure the effectiveness of treatment plans. The system further predicts the key risk points of the next cycle by analyzing the inflection points, rising or falling trends in the cumulative pregnancy rate curve. By accurately identifying these risk points, the system can provide doctors and patients with personalized treatment recommendations, including adjusting drug dosages, changing treatment plans, or optimizing treatment strategies.

[0087] The fifth step involves a multi-dimensional analysis of influencing factors in the prediction of the cumulative pregnancy rate. First, the system identifies the key factors that affect the success of pregnancy by performing a multi-dimensional analysis of the results of the cumulative pregnancy rate prediction. These key factors include, but are not limited to, patient age, embryo quality, and the effectiveness of the treatment plan. By comprehensively analyzing these multi-dimensional data, the system can more comprehensively understand the extent of the impact of each factor on the cumulative pregnancy rate and evaluate the interactions between the factors. For example, the patient's age has a greater impact on the quality of the embryo, while the effectiveness of the treatment plan directly determines the success rate of embryo implantation and pregnancy maintenance. By identifying and quantifying these key factors, the system can further optimize the prediction model of the cumulative pregnancy rate, thereby providing more accurate prediction results and personalized treatment recommendations.

[0088] Embodiment 1:

[0089] See also Figure 2 In this embodiment, the multi-dimensional factors of the patient in each treatment cycle are first integrated, including physiological indicators such as follicle development and embryo quality, as well as the patient's psychological state, living habits and external environment. This multi-dimensional integration process can be expressed by the following formula:

[0090] Φ(x,t)=∫0 T [x(t)·exp(-∫0 t γ(τ)dτ)+η(t)·cos(∫0 t δ(τ)dτ)]dt

[0091] Among them, x(t) represents the multidimensional data collected in each cycle, such as the patient's follicle development, embryo quality, hormone level, etc. The coefficient γ(τ) reflects the attenuation effect of time on different data items, and its value range can be 0.01 to 0.1, depending on the data type; δ(τ) is used to capture the fluctuation characteristics of cycle data, reflecting the frequency of data changes in different cycles, and its value range can be 0.1 to 1.0; η(t) represents non-physiological factors in the cycle, such as the patient's psychological state, and its influence can be quantified using similar methods.

[0092] Suppose there is a 35-year-old female patient who has received assisted reproductive treatment in the past three cycles. In each cycle, the system collected her follicle development data (x1(t)), embryo quality data (x2(t)), and her psychological state (1). These data are integrated into the above formula to predict her cumulative pregnancy rate.

[0093] The specific data are as follows: In the first cycle, the average value of follicle development x1(t) is 8.0, the score of embryo quality x2(t) is 6.5, and the quantitative score of psychological state x3(t) is 0.8. For these three data items, γ(τ) takes values ​​of 0.05, 0.07, and 0.02, respectively, while δ(τ) takes values ​​of 0.3, 0.4, and 0.5. Substituting these data into the formula, the predicted value of the first cycle is calculated:

[0094] Φ1=∫0 T [8.0·exp(-∫0 t 0.05dτ)+6.5·cos(∫0 t 0.3dτ)+0.8·cos(∫0 t 0.5dτ)]dt

[0095] The total time T of the set cycle is 10 days. After integral calculation, the integrated value of the first cycle is Φ1=25.3.

[0096] Similarly, the data for the second and third cycles are: x1(t) is 7.5, x2(t) is 7.0, x3(t) is 0.9; γ(τ) is 0.06, 0.08 and 0.03, and δ(τ) is 0.35, 0.45 and 0.55. Through the same calculation steps, the integrated value of the second cycle is Φ2 = 26.1, and the integrated value of the third cycle is Φ3 = 27.4.

[0097] Ultimately, the integrated data from these three cycles will be input into a cumulative pregnancy rate prediction model to predict the patient's success rate in the next treatment cycle.

[0098] The patient's data in three treatment cycles have been integrated and the integrated value of each cycle has been calculated, that is, the integrated value of the first cycle is 25.3, the integrated value of the second cycle is 26.1, and the integrated value of the third cycle is 27.4. Next, the weight of these cycle data in the prediction model is adjusted according to the pregnancy results of each cycle. To this end, the formula is used:

[0099]

[0100] to process the data.

[0101] In this formula, ω i Represents the weight value of each cycle, ω i The initial value is set to 0.5, and y i is the pregnancy outcome data of the cycle. For example, the pregnancy outcome can be expressed as a quantitative score. The pregnancy outcome of the first cycle is 0.6, the second cycle is 0.7, and the third cycle is 0.8. i) is used to capture the nonlinear characteristics of the data, and The term reflects the rate of change of periodic data, λ is the weight adjustment factor, and its value range is usually between 0.01 and 0.1. Set λ=0.05.

[0102] Substituting these values ​​into the formula for calculation: For the first cycle, Ψ is calculated as During the calculation process, The value of is determined by the gradient of the weight as the pregnancy outcome changes. The initial gradient is set to 0.02. After calculation, the weight adjustment value for the first cycle is 0.56.

[0103] Similarly, the calculations for the second and third cycles are: and The final weight for the second cycle was adjusted to 0.58, and the weight for the third cycle was adjusted to 0.60.

[0104] Through this method, the system dynamically adjusts the data weight of each cycle so that it can more accurately reflect the changing trend of the cumulative pregnancy rate and ensure the reliability and effectiveness of the model prediction.

[0105] The patient's data for each cycle has been preliminarily integrated through the model, and the weights have been adjusted according to the pregnancy results. Finally, personalized adjustment factors are incorporated to further improve the accuracy of the cumulative pregnancy rate prediction. At this stage, the system will combine the patient's physiological basic data and previous treatment response, and make personalized adjustments through the function Θ(z) to ensure that the model can fully reflect the patient's specific needs and response patterns.

[0106] The formula Θ(z) is as follows:

[0107]

[0108] In this formula, z(t) represents the patient's personalized characteristic variables, including age, ovarian reserve function, hormone level, psychological state, etc. Here, κ is a factor that controls the decay of personalized characteristics over time, and its value range is usually between 0.01 and 0.05; this paper sets κ = 0.03. β is used to adjust the periodic fluctuations during the treatment process, and its value range is usually between 0.1 and 0.5; this paper sets β = 0.2.

[0109] Taking the same 35-year-old patient as an example, her personalized feature variables z(t) include: age 35 years old (z1(t)), ovarian reserve function score 7.0 (z2(t)), hormone level E2 value 200pg / mL (z3(t)). For these three features, the specific values ​​of κ and β are set to 0.03 and 0.2 respectively.

[0110] Substitute these data into the formula to calculate:

[0111]

[0112] In this calculation process, The value of is zero because age is constant in the short term. Therefore, the effect of age is mainly reflected in the prediction through the attenuation factor exp(-0.03t). Similarly, ovarian reserve function and hormone level data are processed in the same way and finally integrated into the model to form a personalized cumulative pregnancy rate prediction value.

[0113] Through these calculations, the system obtains a comprehensive personalized adjustment value Θ, which can further correct the results of the aforementioned cycle weight adjustment, making the prediction model more in line with the patient's actual situation. For example, through analysis, it can be found that the patient's ovarian reserve function has a greater impact on the cumulative pregnancy rate, so the weight of this feature is relatively increased during the calculation process. Ultimately, this personalized adjustment mechanism ensures the accuracy and applicability of the cumulative pregnancy rate prediction results, helping doctors to develop more scientific and reasonable treatment plans for patients.

[0114] Embodiment 2:

[0115] like Figure 3 As shown, in this embodiment, the predicted value of the cumulative pregnancy rate is updated to ensure the accuracy and personalization of the model. To achieve this, the system first compares the results of the prediction model and adjusts it according to the new cycle data through a real-time feedback mechanism. In this process, an adaptive adjustment function Φ(t) is used to achieve dynamic updating of the model.

[0116] The formula Φ(t) is as follows:

[0117]

[0118] In this formula, α(t) represents the time series data of pregnancy outcomes, such as the success or failure of pregnancy after each cycle. The value range is usually between 0.5 and 1.5, depending on the quantitative representation of the pregnancy outcome. For example, a successful pregnancy can be assigned a value of 1.0, while a failed pregnancy is assigned a value of 0.0. The coefficient β(t) is an adjustment factor that controls the responsiveness of the model to new data, and its value range is between 0.01 and 0.1; this paper sets β = 0.05. γ(t) reflects the dynamic characteristics of changes in pregnancy outcomes, such as embryonic development within the cycle, hormone level fluctuations, etc., and its value range is between 0.1 and 0.5. This paper sets γ = 0.2.

[0119] For a 35-year-old patient, after the new cycle, her pregnancy result is set to be successful (α(t) = 1.0), and during this cycle, the embryo develops well and the hormone level is stable (γ(t) = 0.2). Substitute these data into the formula for calculation:

[0120]

[0121] In this calculation process, Indicates the rate of change of embryonic development and hormone levels. Since the cycle is 10 days, the upper limit of the integral is set to 10. After substituting the above values ​​for calculation, the updated cumulative pregnancy rate prediction value is obtained.

[0122] The calculation result is set to Φ(t) = 15.8, which represents the cumulative pregnancy rate prediction value after combining real-time feedback and adjustment factors. Through this dynamic update process, the system can flexibly adapt to new cycle data, ensure continuous optimization of the model, and improve the accuracy of the prediction.

[0123] Next, the system will readjust the weights of each data feature in the model according to the actual situation of the cycle, and optimize the weights according to the changes in the cycle results. To this end, the weight adjustment function Ψ(ω i ,y i ) to implement this process.

[0124] Formula Ψ(ω i ,y i )as follows:

[0125]

[0126] In this formula, ω i Represents the weight value of each cycle. The initial weight is usually set between 0.3 and 0.7. This paper sets ω i =0.5. i The pregnancy result data of this cycle, the value depends on whether the pregnancy is successful or not in the cycle, for example, a successful pregnancy result is 1.0, and a failed result is 0.0. i ) captures the periodic variation characteristics of the data. It reflects the dynamic adjustment process of weights. The coefficient λ is the attenuation factor that controls the weight adjustment, and its value range is between 0.01 and 0.1. In this paper, λ is set to 0.05.

[0127] Continuing with the previous example of the 35-year-old patient, after the new treatment cycle, her pregnancy outcome was again successful (y i =1.0), the system needs to readjust the weights based on this result. Substitute these data into the formula for calculation:

[0128]

[0129] In this calculation process, The value of is updated by the model according to real-time data. If the change rate is set to 0.02, the final weight adjustment result is 0.48.

[0130] Through this process, the system can accurately adjust the weight of each cycle data in the model, making the prediction results more accurate and personalized. The pregnancy results of each cycle will affect the weight. The system ensures the rationality of the weight through a dynamic adjustment mechanism and optimizes the predicted value of the cumulative pregnancy rate.

[0131] Finally, the model’s prediction path is dynamically adjusted through the personalized path adjustment function Θ(z) to make it more in line with the patient’s specific physiological response and treatment outcome. This process ensures that the model is not only optimized in terms of overall weight, but also can provide accurate predictions and treatment recommendations based on individual differences of patients.

[0132] The formula Θ(z) is as follows:

[0133]

[0134] In this formula, z(t) represents the patient's personalized characteristic variables, such as age, hormone level, ovarian function score, psychological state, etc. Here, κ is a factor that controls the decay of these characteristics over time. Its value range is usually between 0.01 and 0.05. This paper sets κ = 0.03. β is used to capture the cyclical changes of patients during treatment. Its value range is usually between 0.1 and 0.3. β is set to 0.2.

[0135] For a 35-year-old patient, in a new treatment cycle, her personalized characteristic variables z(t) include: age 35 (z1(t) = 35), ovarian function score 7.5 (z2(t) = 7.5), hormone level E2 value 220pg / mL (z3(t) = 220). After each cycle, the system substitutes these characteristic variables into the formula for calculation to adjust the personalized path of the prediction model.

[0136] Substitute the data into the formula to calculate:

[0137]

[0138] In this formula, age data is a relatively stable variable, so The change rate of ovarian function score Assuming that it does not change much during the cycle, its change rate can be considered to be 0.1. Through calculation, a personalized path adjustment value can be obtained, which will further affect the prediction results of the cumulative pregnancy rate.

[0139] The final calculation result is set to Θ(z) = 16.5, which shows that the system further optimizes the model's prediction path through in-depth analysis of the patient's personalized characteristics. This result not only reflects the patient's overall physiological state, but also captures her specific reactions and dynamic changes during treatment.

[0140] Embodiment 3:

[0141] Combination Figure 4 In this embodiment, the generation of the cumulative pregnancy rate curve guides the new data of each cycle into the model through a real-time capture mechanism, thereby affecting the prediction results. To achieve this, an adaptive capture and input function Φ(x,t) is used to process the cycle data and generate the cumulative pregnancy rate curve.

[0142] The formula Φ(x,t) is as follows:

[0143]

[0144] In this formula, x(t) represents the data input of each cycle, including follicle development, embryo quality, hormone levels, etc. The coefficient α controls the decay of data over time, and its value range is usually between 0.01 and 0.1. This paper sets it to 1. β is used to adjust the volatility of data, reflecting the fluctuation characteristics of data in the short term. Its value range is usually between 0.1 and 1.0. This paper sets β = 0.3.

[0145] For a 35-year-old patient, her treatment data in the first three cycles are set as follows: in the first cycle, the follicle development data x1(t) is 8.0, the embryo quality data x2(t) is 6.5, and the hormone level data x3(t) is 200pg / mL; in the second cycle, the follicle development data x1(t) is 7.5, the embryo quality data x2(t) is 7.0, and the hormone level data x3(t) is 210pg / mL; in the third cycle, the follicle development data x1(t) is 8.2, the embryo quality data x2(t) is 7.2, and the hormone level data x3(t) is 220pg / mL.

[0146] Substitute these data into the formula for calculation to generate the predicted results of the cumulative pregnancy rate curve. For the first cycle, set the total cycle time to 10 days and substitute the formula for calculation:

[0147]

[0148] set up It represents the rate of change of follicle development data within the cycle, and the result is 0.2. After calculation, the integration value of the first cycle is Φ1=20.3.

[0149] Similarly, the data of the second and third cycles will be substituted into the formula for calculation, and the results will be Φ2 = 21.1 and Φ3 = 22.5. These calculation results will be further used to generate a cumulative pregnancy rate curve to show the trend of successful pregnancy during the treatment process.

[0150] Next, the system needs to adjust the weight of each cycle data and the overall prediction parameters in real time according to the pregnancy results of each cycle. i ,y i ) to optimize the prediction process of the model.

[0151] Formula Ψ(ω i ,y i )as follows:

[0152]

[0153] In this formula, ω i Represents the weight of each cycle. The initial weight is usually set between 0.3 and 0.7. This paper sets ω i =0.5. i is the pregnancy outcome data of the cycle, and the successful pregnancy outcome is set to 1.0 and the failed pregnancy outcome is set to 0.0. Coefficient 1 is used to adjust the volatility of the cycle data, and the value range is usually between 0.1 and 0.5. This paper sets γ = 0.2; λ is used to control the exponential decay of the weight adjustment, and its value range is between 0.01 and 0.1. This paper sets λ = 0.05.

[0154] A 35-year-old patient set her pregnancy result in the third cycle as successful (y i =1.0), the system needs to readjust the weights based on this result. Substitute these data into the formula for calculation:

[0155]

[0156] In this calculation process, The value of is updated by the model based on real-time data, and its change rate is set to 0.02, so the final weight adjustment result is 0.49. Through this process, the system can accurately adjust the weight of each period data in the model, making the prediction results more accurate and personalized.

[0157] Finally, the system quantitatively displays the contribution and trend changes of each cycle to the overall pregnancy rate during the treatment process in a graphical way. This process is achieved through the curve generation function Θ(z), which can intuitively display the trend of the patient's pregnancy success rate in different cycles and provide doctors with more valuable reference information.

[0158] The formula Θ(z) is as follows:

[0159]

[0160] In this formula, z(t) represents the patient's personalized data, such as ovarian function score, hormone level, psychological state, etc. The coefficient κ is used to control the attenuation of personalized data, and its value range is usually between 0.01 and 0.05. This paper sets κ = 0.03. δ is used to adjust the dynamic change characteristics of the curve, and its value range is usually between 0.1 and 0.5. This paper sets δ = 0.2.

[0161] For a 35-year-old patient, her personalized data in the first few cycles are set as follows: in the first cycle, the ovarian function score is 7.5 and the hormone level is 210pg / mL; in the second cycle, the ovarian function score is 7.7 and the hormone level is 220pg / mL; in the third cycle, the ovarian function score is 7.8 and the hormone level is 230pg / mL. These data will be substituted into the formula to generate a prediction curve reflecting the cumulative pregnancy rate.

[0162] Substitute the data into the formula for calculation to generate the prediction curve:

[0163]

[0164] In this calculation process, the rate of change of ovarian function score during the cycle is set is 0.1. After calculation, the curve generation value of the first cycle is Θ1=15.4.

[0165] Similarly, the data of the second and third cycles will be substituted into the calculation, and Θ2=16.2 and Θ3=16.8 will be obtained. These results will be used as a quantitative expression of the contribution of each cycle to the cumulative pregnancy rate in the graphical display, and the generated curve will clearly show the trend change of the pregnancy rate in each cycle. The curve generation function Θ(z) provides a more accurate and intuitive graphical display for the prediction of the cumulative pregnancy rate by combining personalized data and real-time adjustment mechanism.

[0166] Embodiment 4:

[0167] In this embodiment, the cumulative pregnancy rate curve control method dynamically generates a cumulative pregnancy rate curve by real-time monitoring of various physiological data of patients during assisted reproductive treatment, including hormone levels, embryo development, and ovarian response. This process is controlled by an adaptive acquisition function Φ(x, t) to accurately capture and reflect the changing trend of patients during treatment.

[0168] The formula Φ(x,t) is as follows:

[0169]

[0170] In this formula, x(t) represents the key physiological data collected in real time during each treatment cycle, such as hormone levels, embryonic development, and ovarian response. The coefficient α is the time decay coefficient, which is used to capture the decay trend of the data over time. Its value range is usually between 0.01 and 0.05. This paper sets α = 0.03. β is used to adjust the fluctuation characteristics of physiological data in the short term, reflecting the frequency of data changes within the cycle. Its value range is usually between 0.1 and 0.3. This paper sets β = 0.2.

[0171] Take a 35-year-old female patient as an example. She went through multiple cycles of assisted reproductive treatment. During each cycle, the system monitored her key physiological data in real time. In the first cycle, the hormone level data x1(t) was 200pg / mL, the embryo development data x2(t) was 7.0 (score), and the ovarian response data x3(t) was 8.0 (score). Substitute these data into the formula and set the total cycle time to 10 days:

[0172]

[0173] In this calculation process, the rate of change of hormone level data during the cycle is set After calculation, the cumulative pregnancy rate curve of the first cycle generated a value of Φ1=1500 (the set value is used to demonstrate the calculation process).

[0174] Similarly, the data of the second and third cycles will be substituted into the formula to calculate Φ2 = 1550 and Φ3 = 1600. These results are used to dynamically generate a cumulative pregnancy rate curve to show the pregnancy success trend of the patient in different treatment cycles.

[0175] Next, the system uses intelligent algorithms to deeply analyze these cumulative pregnancy rate curves to identify and predict key risk points. This process is done through the curve analysis function Ψ(ω i ,y i ) to help optimize treatment plans and address problems in advance.

[0176] Formula Ψ(ω i ,y i )as follows:

[0177]

[0178] In this formula, ω i Represents the importance weight of each cycle, usually ranging from 0.3 to 0.7. This paper sets ω i=0.5. The coefficient γ is used to capture the periodic fluctuation of the patient's physiological data in the curve. Its value range is usually between 0.1 and 0.5. In this paper, γ is set to 0.2. λ is the dynamic adjustment of the adjustment weight. Its value range is between 0.01 and 0.1. In this paper, λ is set to 0.05.

[0179] For a 35-year-old female patient, during her fourth treatment cycle, the system detected large fluctuations in embryonic development and hormone levels, and an inflection point appeared on the cumulative pregnancy rate curve (the slope changed from positive to negative). At this time, the system needs to use the curve analysis function Ψ(ω i ,y i ) to identify the potential risks of this inflection point and make corresponding predictions and adjustments.

[0180] Set the pregnancy result data of the fourth cycle to y i =0.8, weight ω i =0.5. Substitute these data into the formula for calculation:

[0181]

[0182] set up That is, the weight change rate of the model adjusted according to the change of periodic data. After calculation, the curve analysis value obtained is Ψ=0.49.

[0183] Through this process, the system can identify key change points in the curve and, combined with the patient's historical data, predict key risks for the next cycle. This analysis can not only help doctors adjust treatment plans in advance (such as adjusting drug dosages or changing treatment strategies), but also detect potential problems at the earliest stage, thereby reducing treatment risks and increasing pregnancy success rates.

[0184] Finally, the system combines the patient's individual physiological data and curve change characteristics, and generates real-time risk warnings and personalized treatment recommendations through the adaptive risk warning and treatment adjustment function Θ(z). This process ensures that risks in each treatment cycle can be identified in a timely manner and provides doctors with targeted treatment adjustment recommendations to optimize the patient's pregnancy success rate.

[0185] The formula Θ(z) is as follows:

[0186]

[0187] In this formula, z(t) represents patient-specific physiological and treatment data, including age, hormone levels, ovarian response, etc. The coefficient κ is used to adjust the sensitivity of risk warning, and its value range is usually between 0.01 and 0.05. This paper sets κ = 0.03. δ is used to control the response speed of treatment adjustment suggestions, and its value range is between 0.1 and 0.5. This paper sets δ = 0.2.

[0188] For a 35-year-old female patient, in the first few cycles, the patient's hormone levels and embryo development have been relatively stable, but in the fourth cycle, the system detected fluctuations in hormone levels and slightly delayed embryo development. At this time, the system generates risk warnings by analyzing these data and provides personalized treatment adjustment suggestions.

[0189] The patient's physiological data are set to include: hormone level of 220pg / mL (z1(t) = 220), ovarian response score of 7.8 (z2(t) = 7.8). Substitute these data into the formula for calculation:

[0190]

[0191] In this calculation process, the rate of change of the ovarian response score during the cycle is set After calculation, the personalized path adjustment value is obtained as Θ(z)=18.5.

[0192] Through this process, the system is not only able to identify the risks patients face during the fourth cycle, but is also able to adjust treatment plans based on real-time data, suggesting that doctors optimize drug dosages or change treatment strategies to reduce risks and increase pregnancy success rates.

[0193] Embodiment 5:

[0194] In this embodiment, determining the root cause of the curve change is an important step to ensure accurate adjustment of the treatment strategy. To achieve this goal, the system constructs a multidimensional data model by integrating the patient's individualized historical data (such as previous pregnancy history, hormone levels, ovarian response, etc.) with the real-time generated cumulative pregnancy rate curve morphological characteristics (such as smoothness, amplitude and frequency). It is described by the function Φ(x, t) to capture and analyze the patient's complex response pattern during treatment.

[0195] The formula Φ(x,t) is as follows:

[0196]

[0197] In this formula, x(t) represents the key physiological data collected in real time during each treatment cycle, such as hormone levels, ovarian response scores, embryonic development, etc. The coefficient α controls the decay characteristics of the data over time, and usually ranges from 0.01 to 0.05. This paper sets α = 0.03. β is used to capture the fluctuation characteristics of real-time data and reflect the short-term changes of the data. The value range is between 0.1 and 0.5. This paper sets β = 0.2.

[0198] A 35-year-old female patient has two failed pregnancy experiences in the previous treatment cycle. In the current cycle, the system monitors her hormone level to be 210pg / mL (x1(t)=210), ovarian response score to be 7.5 (x2(t)=7.5), and embryo development score to be 6.8 (x3(t)=6.8). These data will be used to construct and analyze her cumulative pregnancy rate curve to determine the risk points that may occur during treatment.

[0199] Substitute these data into the formula to calculate:

[0200]

[0201] In this calculation, the rate of change of hormone levels during the cycle is set After integral calculation, the value obtained is Φ(x, t) = 1500 (the set value is used to demonstrate the calculation process).

[0202] Through this calculation, the system can identify the patient's physiological response pattern in the current treatment cycle, and combine historical data to determine the impact of these responses on the cumulative pregnancy rate curve. If the calculation result shows that the curve has a large amplitude and an increased frequency, it indicates that the patient is at higher risk in the current treatment cycle, and the system will recommend that the doctor consider adjusting the treatment plan, such as optimizing the drug dosage or adjusting the treatment rhythm.

[0203] Next, the root causes of these changes were further confirmed by in-depth analysis of the morphological characteristics of the cumulative pregnancy rate curve, including smoothness, amplitude, and frequency. The smoothness of the curve reflects the stability of the treatment effect, the amplitude represents the strength of the treatment response, and the frequency reveals the periodic fluctuations of the patient's physiological response. To perform these analyses, the system uses the curve analysis function Ψ(ω i ,y i ) to evaluate these morphological features.

[0204] Formula Ψ(ω i ,y i )as follows:

[0205]

[0206] In this formula, ω i Represents the weight of different morphological features, and the value range is usually 0.3 to 0.7. In this paper, ω is set i =0.5. The coefficient γ is used to capture the periodic changes of the curve. Its value range is usually 0.1 to 0.5. In this paper, γ is set to 0.2. λ adjusts the dynamic changes of the weight. Its value range is usually 0.01 to 0.1. In this paper, λ is set to 0.05.

[0207] For a 35-year-old female patient, the system was set to monitor that the smoothness of the cumulative pregnancy rate curve decreased, the amplitude increased, and the frequency increased significantly in her latest treatment cycle. In order to analyze the specific reasons for these changes, the system first substituted the morphological feature data into the curve analysis function Ψ(ω i ,y i ) for calculation.

[0208] In this cycle, the smoothness weight of the curve is set to 0.5 (ω1), the amplitude weight is set to 0.4 (ω2), and the frequency weight is set to 0.6 (ω3). The corresponding morphological feature data are 0.7 (smoothness), 1.2 (amplitude), and 0.9 (frequency). Substitute these data into the formula for calculation:

[0209]

[0210] set up The calculated curve analysis value is Ψ=1.25 (the set value is used to demonstrate the calculation process).

[0211] Through this calculation, the system identified that the main reason for the decrease in curve smoothness and increase in amplitude in the current treatment cycle is related to the fluctuation of the patient's hormone levels, while the increase in frequency is due to unstable ovarian response. Based on these analyses, the system can suggest that doctors adjust the treatment plan, such as reducing hormone fluctuations or strengthening the regulation of ovarian response, thereby reducing the risk during treatment and increasing the cumulative pregnancy rate.

[0212] Finally, in order to further analyze and accurately determine the actual reasons behind these changes, the system uses intelligent algorithms to combine the patient's individualized historical data and real-time curve characteristics, and performs in-depth analysis through function Θ(z).

[0213] The formula Θ(z) is as follows:

[0214]

[0215] In this formula, z(t) represents patient-specific physiological and treatment data, such as hormone levels, ovarian response scores, embryonic development, etc. The coefficient κ is used to adjust the sensitivity of risk warning, usually ranging from 0.01 to 0.05, and κ = 0.03 is set. δ is used to control the response speed of treatment adjustment recommendations, with a range of 0.1 to 0.5, and δ = 0.2 is set.

[0216] For a 35-year-old female patient, in her latest treatment cycle, the system detected that her hormone levels fluctuated greatly (z1(t) = 215pg / mL), her ovarian response score was 7.7 (z2(t) = 7.7), and her embryonic development score was 6.9 (z3(t) = 6.9). The system uses these data to determine that the patient is at risk and uses the function Θ(z) for in-depth analysis.

[0217] Substitute these data into the formula to calculate:

[0218]

[0219] Set the rate of change of ovarian response score during the cycle After calculation, the system obtains an analysis value of Θ(z)=19.3 (the set value is used to illustrate the calculation process). The analysis value reflects the change trend of the patient's physiological data and historical data in the current cycle.

[0220] Through this calculation, the system can more accurately determine the root cause of the change in the cumulative pregnancy rate curve, and determine whether it is caused by hormone level fluctuations, unstable ovarian response, or embryo development. This intelligent in-depth analysis allows doctors to develop more precise treatment plans based on specific reasons, such as adjusting hormone drug dosages or changing embryo transfer strategies, thereby reducing the patient's risk in subsequent treatment cycles.

[0221] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An optimization system for predicting cumulative pregnancy rate of assisted reproductive technology, characterized by The following steps are involved: S1, multi-period data accumulation and initial forecast modeling; S1.

1. First, accumulate the patient's reproductive health information through multi-cycle data collection, including follicle development, embryo quality, and endometrial thickness during each assisted reproductive treatment; S1.

2. Construct an initial cumulative pregnancy rate prediction model based on multi-cycle data; S2. Dynamic prediction and adjustment of cumulative pregnancy rate; S2.

1. After each new treatment cycle, the success or failure of the cycle is compared with the initial prediction and the new data is input into the model. S2.

2. Update the predicted value of cumulative pregnancy rate through a dynamic adjustment mechanism; S3, generation of stage-by-stage cumulative pregnancy rate curve; S3.

1. Generate a cumulative pregnancy rate curve for each cycle based on the dynamic prediction results. S3.2, the cumulative pregnancy rate curve is generated based on the dynamic prediction adjustment mechanism of S2, which is a graphical quantitative expression of the overall prediction; S4. Intelligent decision-making based on curve changes; S4.

1. Provide intelligent decision support based on the changing trend of the cumulative pregnancy rate curve; S4.

2. By analyzing the inflection points, upward or downward trends of the curve, key risk points of the next cycle can be predicted, and personalized treatment recommendations can be provided to doctors and patients, including adjusting drug dosages or changing treatment plans. S5. Analysis of multidimensional influencing factors for the prediction of cumulative pregnancy rate; S5.

1. Through multi-dimensional analysis of the cumulative pregnancy rate prediction results, identify the key factors affecting pregnancy success, including patient age, embryo quality, and effectiveness of treatment plan.

2. The optimization system for predicting cumulative pregnancy rate of assisted reproductive technology according to claim 1, characterized in that The cumulative pregnancy rate prediction model construction includes: S1. First, integrate the patient's physiological indicators in each treatment cycle, including follicle development and embryo quality, as well as multidimensional factors including the patient's psychological state, lifestyle, and external environment. The multidimensional integration is expressed using the formula: Among them, x(t) represents the multidimensional data collected in each cycle, γ(τ) reflects the attenuation effect of time on different data items, and δ(τ) is used to capture the fluctuation characteristics of periodic data; S2. Adjust the weight of each cycle data in the model based on the pregnancy results of each cycle, including the success rate of embryo implantation and the stability of pregnancy maintenance; the formula is used to express it: Among them, ω i Represents the weight value of each cycle, y i The pregnancy result data of this cycle; by sin(y i ) captures the nonlinear characteristics of the data, The term reflects the rate of change of periodic data, combined with exp(-λy i ) weight adjustment; S3. Finally, personalized adjustment factors are incorporated; combining the patient's physiological baseline data and previous treatment responses.

3. The optimization system for predicting cumulative pregnancy rate of assisted reproductive technology according to claim 1, characterized in that The method used to update the predicted value of the cumulative pregnancy rate includes: S1. First, the prediction model results are compared and then adjusted according to the new periodic data through the real-time feedback mechanism. The adaptive adjustment function Φ(t) is used to achieve this, and the model is dynamically updated according to the real-time data: Among them, α(t) represents the time series data of pregnancy outcomes, β(t) is the adjustment factor, and γ(t) reflects the dynamic characteristics of outcome changes; S2. Next, according to the actual situation of the cycle, the weights of each data feature in the model are readjusted; and the weights are optimized according to the changes in the cycle results; the dynamic adjustment mechanism is implemented through the weight adjustment function Ψ(ω i ,y i ) to achieve: Among them, ω i Indicates the weight value of each cycle, y i is the pregnancy outcome data of this cycle; i ) captures the periodic variation characteristics of the data, It reflects the dynamic adjustment process of weights, combined with exp(-λy i ) exponential decay; S3. Finally, after the prediction model adjusts the overall weight, it dynamically adjusts the personalized prediction path based on the patient's specific response and treatment results.

4. The optimization system for predicting cumulative pregnancy rate of assisted reproductive technology according to claim 1, characterized in that The cumulative pregnancy rate curve generation construction includes: S1, guide the new data of each cycle to influence the prediction results of the model through real-time capture mechanism; S2. Adjust the data weights of each cycle and the overall prediction parameters in the model in real time according to the pregnancy results of each cycle; S3. Quantify the contribution of each cycle to the overall pregnancy rate and trend changes during the treatment process through graphical means.

5. The optimization system for predicting cumulative pregnancy rate of assisted reproductive technology according to claim 1, characterized in that The personalized treatment recommendation construction framework includes the following steps: S1. First, by dynamically tracking and analyzing the inflection points, upward or downward trends of the cumulative pregnancy rate curve, key risk points predicted to occur during the treatment process are identified in real time. The identification process is based on the slope changes, curvature and fluctuation amplitude of the curve to monitor sudden changes in the effectiveness of the treatment plan or nonlinear changes in the patient's physiological state. S2. Secondly, combining the patient's individual historical data with the curve morphological characteristics including smoothness, amplitude and frequency, a customized risk assessment is used to determine the root cause of the curve change; S3. Finally, based on the risk assessment results, intelligent algorithms generate personalized treatment recommendations, including adjusting drug dosage, optimizing treatment rhythm, or adjusting lifestyle; Among them, the intelligent algorithm analyzes the effects of different treatment strategies in similar medical records, combines the current patient's curve trend and risk assessment results, and provides multiple options for doctors to choose from.

6. The optimization system for predicting cumulative pregnancy rate of assisted reproductive technology according to claim 5, characterized in that The cumulative pregnancy rate curve control method comprises: S1. By real-time monitoring of various data during assisted reproductive treatment, including hormone levels, embryo development, and ovarian response, a cumulative pregnancy rate curve is dynamically generated. The curve is controlled by an adaptive acquisition function Φ(x, t): Where x(t) represents the key physiological data collected in real time during each treatment cycle, α is the time decay coefficient, which is used to capture the decay trend of the data over time; β is used to adjust the fluctuation characteristics of physiological data in the short term; S2. Then, intelligent algorithms are used to deeply analyze the inflection point, slope change, curvature and fluctuation range of the cumulative pregnancy rate curve to identify and predict key risk points; through the curve analysis function Ψ(ω i ,y i ) to calculate: Among them, ω i represents the importance weight of each cycle, γ is used to capture the periodic fluctuation of the patient's physiological data in the curve, and λ regulates the dynamic adjustment of the weight; S3. Finally, after identifying the key risk points, combine the patient's individualized physiological data and curve change characteristics.

7. The optimization system for predicting cumulative pregnancy rate of assisted reproductive technology according to claim 5, characterized in that The method used to determine the root cause of the curve change is: S1. A multidimensional data model is constructed by integrating the patient's individualized historical data, including previous pregnancy history, hormone levels, and ovarian response, with the real-time cumulative pregnancy rate curve, including morphological characteristics such as smoothness, amplitude, and frequency. This model is described by the function Φ(x, t): Where x(t) represents the key physiological data collected in real time during each treatment cycle, α controls the decay characteristics of the data over time, and β is used to capture the fluctuation characteristics of the real-time data; S2. Secondly, the root cause of the curve change is determined by analyzing the morphological characteristics of the cumulative pregnancy rate curve, including smoothness, amplitude, and frequency. The smoothness of the curve reflects the stability of the treatment effect, the amplitude represents the intensity of the treatment response, and the frequency reveals the periodic fluctuation of the physiological response. The function Ψ(ω i ,y i ) These morphological characteristics were evaluated: Among them, ω i Represents the weights of different morphological features, γ is used to capture the periodic changes of the curve, and λ adjusts the dynamic changes of the weights; S3. Finally, the intelligent algorithm is used to determine the cause of the change in the cumulative pregnancy rate curve. Combined with individual historical data and real-time curve characteristics, the function Θ(z) is used for in-depth analysis: Here, z(t) represents the patient-specific physiological and treatment data, and κ and δ are used to adjust the response speed and sensitivity of intelligent analysis.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the optimization system for predicting the cumulative pregnancy rate of assisted reproductive technology as described in any one of claims 1 to 7.

Images