CN116362355A - Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm - Google Patents

Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm Download PDF

Info

Publication number
CN116362355A
CN116362355A CN202310370279.1A CN202310370279A CN116362355A CN 116362355 A CN116362355 A CN 116362355A CN 202310370279 A CN202310370279 A CN 202310370279A CN 116362355 A CN116362355 A CN 116362355A
Authority
CN
China
Prior art keywords
learning model
xgboost
random forest
integrated learning
temperature
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310370279.1A
Other languages
Chinese (zh)
Inventor
林莉
王静芝
曹雅裴
孔宪宇
范米
林雨露
谭惠丹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University
Original Assignee
Chongqing University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chongqing University filed Critical Chongqing University
Priority to CN202310370279.1A priority Critical patent/CN116362355A/en
Publication of CN116362355A publication Critical patent/CN116362355A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

A method for predicting the temperature of a cable joint guide core by adopting XGBoost and random forest algorithm comprises the following steps: s1, acquiring historical data affecting the temperature of a cable joint guide core, cleaning the historical data, and S2, constructing an XGBoost serial integrated learning model and a random forest parallel integrated learning model; s3, inputting the cleaned historical data into an XGBoost integrated learning model and a random forest integrated learning model for training; s4, judging whether the XGBoost serial learning model and the random forest parallel learning model are trained, if yes, entering a step S5, if not, updating parameters of the XGBoost serial learning model and the random forest parallel learning model, and returning to the step S3; s5, acquiring real-time data affecting the temperature of the cable joint guide core, inputting the real-time data into the XGBoost serial learning model and the random forest parallel learning model after training is completed, and performing prediction processing to obtain a predicted value y i And F i The method comprises the steps of carrying out a first treatment on the surface of the S6, predicting value y i And F i Weighted summation is carried out to obtain the final cable guide coreTemperature predictions.

Description

Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm
Technical Field
The invention relates to the field of cable joint guide core temperature detection, in particular to a method for predicting the cable joint guide core temperature by adopting XGBoost and a random forest algorithm.
Background
With the widespread use of urban cables, the number of joints for connecting cables at both ends has also grown. In actual operation, the temperature of the cable connector guide core is higher than that of the cable body guide core, so that the failure occurrence rate of the cable connector is far higher than that of the cable body, and the temperature of the cable connector guide core needs to be predicted in order to ensure timely maintenance of equipment and safe operation of the cable connector.
At present, in the method for predicting the temperature of the cable joint guide core, a serial integration mode is formed by continuous serial superposition of a base learner, but because of adopting a serialization and self-adaptive optimization strategy, each sub-model has strong correlation, so that the variance and the overfitting risk of a prediction model cannot be obviously reduced. The parallel integrated learning method is to randomly sample the original training data, train each basic learner and calculate the average value, and as the input data extracted each time are mutually independent, the variance is effectively reduced, the model overfitting effect is avoided, but the prediction deviation cannot be effectively reduced. If only a single integration mode is adopted to predict the model, a more stable prediction model with stronger generalization capability cannot be obtained.
In order to solve the above technical problems, a new technical means is needed.
Disclosure of Invention
In view of the above, the invention aims to provide a method for predicting the temperature of a cable joint guide core by adopting XGBoost and a random forest algorithm, which fully utilizes historical data, effectively combines the advantages of two integrated learning modes, helps to obtain a more stable prediction model with stronger generalization capability, improves the accuracy of predicting the temperature of the cable joint guide core, and provides an early warning basis for a cable temperature detection system.
The invention provides a method for predicting the temperature of a cable joint guide core by adopting XGBoost and a random forest algorithm, which comprises the following steps:
s1, acquiring historical data affecting the temperature of a cable joint guide core, and cleaning the historical data, wherein the historical data affecting the temperature of the cable joint guide core comprises the ambient temperature, load current and the temperature of the cable joint;
s2, constructing an XGBoost serial integrated learning model and a random forest parallel integrated learning model;
s3, inputting the cleaned historical data into an XGBoost serial integrated learning model and a random forest parallel integrated learning model for training;
s4, judging whether the XGBoost serial integrated learning model and the random forest parallel integrated learning model finish training, if so, entering a step S5, if not, updating parameters of the XGBoost serial integrated learning model and the random forest parallel integrated learning model, and returning to the step S3;
s5, acquiring real-time data affecting the temperature of the cable joint guide core, inputting the real-time data into the XGBoost serial integrated learning model and the random forest parallel integrated learning model after training is finished, and performing prediction processing to obtain a predicted value
Figure BDA0004168369840000021
And F i Wherein->
Figure BDA0004168369840000022
A cable guide core temperature predicted value F output by XGBoost serial integrated learning model i A cable guide core temperature predicted value output by the random forest parallel integrated learning model;
s6, regarding predicted values
Figure BDA0004168369840000023
And F i And carrying out weighted summation to obtain a final cable guide core temperature predicted value.
Further, in step S3, the XGBoost serial ensemble learning model is trained according to the following method:
s311, according to the history data D= { (x) after cleaning i ,y i )}(|D|＝n,x i ∈R m ,y i E, R), training CART tree, the integrated mode of tree is:
Figure BDA0004168369840000024
wherein f= { F (x) =w q(x) }(q:R m →T,w∈R T ) Is the collective space of the tree; x is x i Feature vectors for the ith data point; y is i Each data point corresponds to a true value; q is the index of the leaf corresponding to the sample mapped to each tree structure; t is the number of leaves on the tree, f k For CART tree, each tree f k The weight w of the leaf and the independent tree structure q are corresponding; r is R m Representing a set of data point feature vectors; r is R T Representing a set of data points corresponding to real values;
s312, constructing an objective function based on an XGBoost serial integrated learning model:
Figure BDA0004168369840000031
where i refers to each instance of the process,
Figure BDA0004168369840000032
representing predicted value, y i Representing the true value, i is the loss function, the first part is the predicted value +.>
Figure BDA0004168369840000033
And target true value y i Training errors between; omega (f) k ) Representing the complexity of each tree, the second part is the sum of the complexity of each tree, a canonical term for controlling the complexity of the model, the complexity being:
Figure BDA0004168369840000034
wherein, gamma and lambda represent adjustable parameters, T is the number of leaves on the tree, and W represents the score of leaf nodes;
s313, training an objective function by adopting an incremental training method, adding a new function to the model on the basis of preserving the original model each time, and reducing the objective function to the greatest extent possible by each added incremental function, wherein the process is as follows:
Figure BDA0004168369840000035
wherein,,
Figure BDA0004168369840000036
is the model predictive value of the ith sample at the t-th round,>
Figure BDA0004168369840000037
model predictive value of t-1 round is reserved +.>
Figure BDA0004168369840000038
After that, a new function f is added t (x i ) Until the objective function is minimal.
Further, in step S3, the random forest parallel integrated learning model is trained according to the following method:
s321, randomly extracting a data subset from the cleaned historical data by adopting a Bootstrap method;
s322, modeling a decision tree for each subset respectively;
s323, historical data X of environment temperature, load current, joint surface temperature, body surface temperature, surface temperature difference and the like k The decision tree input to the random forest parallel integrated learning model is tested to obtain a prediction result sequence { f ] of each decision tree model k1 (X k ),f k2 (X k ),…,f kw (X k )}；
S324, synthesizing test results of all decision trees, and obtaining a final cable connector guide core temperature prediction model through voting in the following manner:
Figure BDA0004168369840000041
wherein F is k Is G k The temperature prediction model f of the cable joint guide core ki A single decision tree prediction model; i is an indication function, Y k The result is predicted for a single decision tree.
Further, in step S6, based on step S5, specifically including:
determining a predicted value weight w of an XGBoost serial integrated learning model 1 Predicted value weight w of parallel integrated learning model of random forest 2 ；
From weighted sum formulas
Figure BDA0004168369840000042
And calculating a final cable guide core predicted value.
Further, the predicted value weight w of the XGBoost serial integrated learning model is determined by the following method 1 Predicted value weight w of parallel integrated learning model of random forest 2 ：
Constructing an optimization function aiming at the minimum prediction error:
Figure BDA0004168369840000043
wherein L is i The temperature true value of the guide core is corresponding to the cable joint historical data;
Figure BDA0004168369840000044
the cable joint guide core temperature predicted value based on XGBoost serial integrated learning; f (F) i The cable joint guide core temperature predicted value is based on random forest parallel integrated learning; solving the optimization function by using a commercial optimizer Gurobi, and when the optimization function reaches the minimum value, corresponding w 1 And w 2 Is the optimized weight.
The invention has the beneficial effects that: according to the invention, a cable joint guide core temperature prediction model with higher stability and generalization capability can be obtained, the accuracy of the prediction model can be improved, and an early warning basis is provided for a cable temperature detection system.
Drawings
The invention is further described below with reference to the accompanying drawings and examples:
FIG. 1 is a flow chart of the present invention
FIG. 2 is a schematic diagram of a XGBoost-based serial ensemble learning mechanism provided in an embodiment of the present application
FIG. 3 is a schematic diagram of a parallel integrated learning mechanism based on random forests according to an embodiment of the present application
Fig. 4 is a graph showing the effect of predicting the temperature of a cable joint guide core according to an embodiment of the present application
Detailed Description
The present invention is further described in detail below:
the invention provides a method for predicting the temperature of a cable joint guide core by adopting XGBoost and a random forest algorithm, which comprises the following steps:
s1, acquiring historical data affecting the temperature of a cable joint guide core, and cleaning the historical data, wherein the historical data affecting the temperature of the cable joint guide core comprises the ambient temperature, load current and the temperature of the cable joint;
s2, constructing an XGBoost serial integrated learning model and a random forest parallel integrated learning model;
s3, inputting the cleaned historical data into an XGBoost serial integrated learning model and a random forest parallel integrated learning model for training;
s4, judging whether the XGBoost serial integrated learning model and the random forest parallel integrated learning model finish training, if so, entering a step S5, if not, updating parameters of the XGBoost serial integrated learning model and the random forest parallel integrated learning model, and returning to the step S3;
s5, acquiring real-time data affecting the temperature of the cable connector guide core, and inputting the real-time data into the XGBoost serial integrated learning model and the random after trainingIn the forest parallel integrated learning model, prediction processing is carried out to obtain a predicted value
Figure BDA0004168369840000051
And F i Wherein->
Figure BDA0004168369840000052
A cable guide core temperature predicted value F output by XGBoost serial integrated learning model i A cable guide core temperature predicted value output by the random forest parallel integrated learning model;
s6, regarding predicted values
Figure BDA0004168369840000053
And F i And carrying out weighted summation to obtain a final cable guide core temperature predicted value. The XGBoost serial integrated learning model and the random forest parallel integrated learning model are all achieved by the prior art, and through the method, the cable joint guide core temperature prediction model with higher stability and generalization capability can be obtained, the accuracy of the prediction model can be improved, and an early warning basis is provided for a cable temperature detection system.
In this embodiment, in step S1, history data is obtained according to the following method:
collecting load current, ambient temperature and cable joint temperature data with a finite element simulation technique, wherein the cable joint temperature data comprises: cable joint surface temperature, body surface temperature, and surface temperature difference; cleaning the data, removing abnormal data and complementing incomplete data, wherein the abnormal data comprises: error data, duplicate data, etc., and data cleaning procedures are known in the art. By the method, the accuracy of the final prediction result can be ensured.
In this embodiment, in step S3, based on step S2, the XGBoost serial integrated learning model is trained according to the following method:
s311, according to the history data D= { (x) after cleaning i ,y i )}(|D|＝n,x i ∈R m ,y i E, R), training CART tree, the integrated mode of tree is:
Figure BDA0004168369840000061
wherein f= { F (x) =w q(x) }(q:R m →T,w∈R T ) Is the collective space of the tree; x is x i Feature vectors for the ith data point; y is i Each data point corresponds to a true value; q is the index of the leaf corresponding to the sample mapped to each tree structure; t is the number of leaves on the tree, f k For CART tree, each tree f k The weight w of the leaf and the independent tree structure q are corresponding; r is R m Representing a set of data point feature vectors; r is R T Representing a set of data points corresponding to real values;
s312, constructing an objective function based on an XGBoost serial integrated learning model:
Figure BDA0004168369840000062
where i refers to each instance of the process,
Figure BDA0004168369840000063
representing predicted value, y i Representing the true value, i is the loss function, the first part is the predicted value +.>
Figure BDA0004168369840000064
And target true value y i Training errors between; omega (f) k ) Representing the complexity of each tree, the second part is the sum of the complexity of each tree, a canonical term for controlling the complexity of the model, the complexity being:
Figure BDA0004168369840000065
where γ, λ represent the adjustable parameters, T is the number of leaves on the tree, and W represents the score of the leaf node.
S312, training an objective function by adopting an incremental training method, wherein each time, a new function is added to the model on the basis of preserving the original model, and each added incremental function reduces the objective function to the greatest extent as much as possible, and the process is as follows:
Figure BDA0004168369840000071
wherein,,
Figure BDA0004168369840000072
is the model predictive value of the ith sample at the t-th round,>
Figure BDA0004168369840000073
model predictive value of t-1 round is reserved +.>
Figure BDA0004168369840000074
After that, a new function f is added t (x i ) Until the objective function is minimal. By the method, the predicted value can be calculated rapidly, the weight of the training sample points with higher learning rate is improved, and the prediction deviation is reduced effectively.
In this embodiment, in step S3, based on step S2, a random forest parallel integrated learning model is trained according to the following method:
s321, randomly extracting a data subset from the cleaned historical data by adopting a Bootstrap method;
s322, modeling a decision tree for each subset.
S323, historical data X of environment temperature, load current, joint surface temperature, body surface temperature, surface temperature difference and the like k The decision tree input to the random forest parallel integrated learning model is tested to obtain a prediction result sequence { f ] of each decision tree model k1 (X k ),f k2 (X k ),…,f kw (X k )}；
S324, synthesizing test results of all decision trees, and obtaining a final cable connector guide core temperature prediction model through voting in the following manner:
Figure BDA0004168369840000075
wherein F is k Is G k The temperature prediction model f of the cable joint guide core ki A single decision tree prediction model; i is an indication function, Y k The result is predicted for a single decision tree. By the method, variance is effectively reduced, and model overfitting effect is avoided.
In this embodiment, in step S4, it is determined whether the XGBoost serial integrated learning model and the random forest parallel integrated learning model are trained according to the following method:
according to the relative mean square error e RMSE Determining whether the XGBoost serial integrated learning model is trained, wherein the relative mean square error e RMSE The method comprises the following steps:
Figure BDA0004168369840000081
according to average relative error e MAPE Determining whether the random forest parallel integrated learning model is trained, wherein the average relative error e MAPE The method comprises the following steps:
Figure BDA0004168369840000082
wherein L is i Representing the corresponding actual value of the temperature of the guide core of the historical cable joint under the conditions of the same environment temperature, load current, joint surface temperature, body surface temperature and surface temperature difference with the predicted value at the same moment; y is as follows i Predicting cable connector guide core temperature value for XGBoost serial integrated learning model, F i Predicting a cable joint guide core temperature value for a random forest parallel integrated learning model;
when the relative mean square error e RMSE When the minimum value is reached, the XGBoost serial integrated learning model is trained, and when the average relative error e MAPE And when the training of the random forest parallel integrated learning model is completed.By the method, the error between the predicted value and the true value can be ensured to be minimum.
In this embodiment, in step S5, real-time data affecting the temperature of the cable connector guide core is obtained, and the real-time data is input into the XGBoost serial integrated learning model and the random forest parallel integrated learning model after training is completed, and prediction processing is performed to obtain a predicted value
Figure BDA0004168369840000083
And F i Wherein->
Figure BDA0004168369840000084
A cable guide core temperature predicted value F output by XGBoost serial integrated learning model i And outputting a cable guide core temperature predicted value for the random forest parallel integrated learning model.
In this embodiment, in step S6, based on step S5, the method specifically includes:
determining a predicted value weight w of an XGBoost serial integrated learning model 1 Predicted value weight w of parallel integrated learning model of random forest 2 ；
From weighted sum formulas
Figure BDA0004168369840000085
And calculating a final cable guide core predicted value.
By the method, whether the temperature of the cable joint guide core exceeds the safety limit value can be judged according to the predicted value, and the graph of the temperature prediction effect of the cable joint guide core is shown in fig. 4, so that the prediction deviation is small, the prediction effect is stable, and an example test shows that the serial-parallel integrated learning model based on XGBoost and a random forest algorithm has the characteristics of high accuracy, difficult fitting, strong generalization and the like, so that whether the temperature of the cable joint guide core exceeds the safety limit value is judged reliably according to the predicted temperature of the cable joint guide core, and whether the temperature of the cable joint guide core exceeds the safety limit value can be judged according to the predicted value.
In this embodiment, the predicted value weight of the XGBoost serial integrated learning model is determined by the following methodWeight w 1 Predicted value weight w of parallel integrated learning model of random forest 2 ：
Constructing an optimization function aiming at the minimum prediction error:
Figure BDA0004168369840000091
and has a constraint on the range of weights as follows:
Figure BDA0004168369840000092
wherein L is i The temperature true value of the guide core is corresponding to the cable joint historical data;
Figure BDA0004168369840000093
the cable joint guide core temperature predicted value based on XGBoost serial integrated learning; f (F) i The cable joint guide core temperature predicted value is based on random forest parallel integrated learning; solving the optimization function by using a commercial optimizer Gurobi, wherein w corresponds to the optimization function reaching the minimum value 1 And w 2 Is the optimized weight. By the method, the serial-parallel integration model error based on XGBoost and random forests can be minimized.
Finally, it is noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made thereto without departing from the spirit and scope of the technical solution of the present invention, which is intended to be covered by the scope of the claims of the present invention.

Claims (5)

1. A method for predicting the temperature of a cable joint guide core by adopting XGBoost and random forest algorithm is characterized in that: the method comprises the following steps:
s1, acquiring historical data affecting the temperature of a cable joint guide core, and cleaning the historical data, wherein the historical data affecting the temperature of the cable joint guide core comprises the ambient temperature, load current and the temperature of the cable joint;
s2, constructing an XGBoost serial integrated learning model and a random forest parallel integrated learning model;
s3, inputting the cleaned historical data into an XGBoost serial integrated learning model and a random forest parallel integrated learning model for training;
s4, judging whether the XGBoost serial integrated learning model and the random forest parallel integrated learning model finish training, if so, entering a step S5, if not, updating parameters of the XGBoost serial integrated learning model and the random forest parallel integrated learning model, and returning to the step S3;
s5, acquiring real-time data affecting the temperature of the cable joint guide core, inputting the real-time data into the XGBoost serial integrated learning model and the random forest parallel integrated learning model after training is finished, and performing prediction processing to obtain a predicted value
Figure FDA0004168369820000011
And F i Wherein->
Figure FDA0004168369820000012
A cable guide core temperature predicted value F output by XGBoost serial integrated learning model i A cable guide core temperature predicted value output by the random forest parallel integrated learning model;
s6, regarding predicted values
Figure FDA0004168369820000013
And F i And carrying out weighted summation to obtain a final cable guide core temperature predicted value.
2. The method for predicting cable joint core temperatures using XGBoost and random forest algorithm of claim 1, wherein: in step S3, based on step S2, the XGBoost serial integrated learning model is trained according to the following method:
s311, according to the history after cleaningData d= { (x i ,y i )}(|D|＝n,x i ∈R m ,y i E, R), training CART tree, the integrated mode of tree is:
Figure FDA0004168369820000014
wherein f= { F (x) =w q(x) }(q:R m →T,w∈R T ) Is the collective space of the tree; x is x i Feature vectors for the ith data point; y is i Each data point corresponds to a true value; q is the index of the leaf corresponding to the sample mapped to each tree structure; t is the number of leaves on the tree, f k For CART tree, each tree f k The weight w of the leaf and the independent tree structure q are corresponding; r is R m Representing a set of data point feature vectors; r is R T Representing a set of data points corresponding to real values;
s312, constructing an objective function based on an XGBoost serial integrated learning model:
Figure FDA0004168369820000021
where i refers to each instance of the process,
Figure FDA0004168369820000022
representing predicted value, y i Representing the true value, i is the loss function, the first part is the predicted value +.>
Figure FDA0004168369820000023
And target true value y i Training errors between; omega (f) k ) Representing the complexity of each tree, the second part is the sum of the complexity of each tree, a canonical term for controlling the complexity of the model, the complexity being:
Figure FDA0004168369820000024
wherein, gamma and lambda represent adjustable parameters, T is the number of leaves on the tree, and W represents the score of leaf nodes;
s313, training an objective function by adopting an incremental training method, adding a new function to the model on the basis of preserving the original model each time, and reducing the objective function to the greatest extent possible by each added incremental function, wherein the process is as follows:
Figure FDA0004168369820000025
wherein,,
Figure FDA0004168369820000026
is the model predictive value of the ith sample at the t-th round,>
Figure FDA0004168369820000027
model predictive value of t-1 round is reserved +.>
Figure FDA0004168369820000028
After that, a new function f is added t (x i ) Until the objective function is minimal.
3. The method for predicting cable joint core temperatures using XGBoost and random forest algorithm of claim 1, wherein: in step S3, based on step S2, the random forest parallel integrated learning model is trained according to the following method:
s321, randomly extracting a data subset from the cleaned historical data by adopting a Bootstrap method;
s322, modeling a decision tree for each subset respectively;
s323, historical data X of environment temperature, load current, joint surface temperature, body surface temperature, surface temperature difference and the like k Inputting the prediction result sequences into decision trees of random forest parallel integrated learning models for testing to obtain prediction result sequences of the decision tree models{f k1 (X k ),f k2 (X k ),…,f kw (X k )}；
S324, synthesizing test results of all decision trees, and obtaining a final cable connector guide core temperature prediction model through voting in the following manner:
Figure FDA0004168369820000031
wherein F is k Is G k The temperature prediction model f of the cable joint guide core ki A single decision tree prediction model; i is an indication function, Y k The result is predicted for a single decision tree.
4. The method for predicting cable joint core temperatures using XGBoost and random forest algorithm of claim 1, wherein: in step S6, based on step S5, specifically including:
determining a predicted value weight w of an XGBoost serial integrated learning model 1 Predicted value weight w of parallel integrated learning model of random forest 2 ；
From weighted sum formulas
Figure FDA0004168369820000032
And calculating a final cable guide core predicted value.
5. The method for predicting cable joint core temperatures using XGBoost and random forest algorithm of claim 1, wherein: the predicted value weight w of the XGBoost serial integrated learning model is determined by the following method 1 Predicted value weight w of parallel integrated learning model of random forest 2 ：
Constructing an optimization function aiming at the minimum prediction error:
Figure FDA0004168369820000033
wherein L is i The temperature true value of the guide core is corresponding to the cable joint historical data;
Figure FDA0004168369820000034
the cable joint guide core temperature predicted value based on XGBoost serial integrated learning; f (F) i The cable joint guide core temperature predicted value is based on random forest parallel integrated learning; solving the optimization function by using a commercial optimizer Gurobi, and when the optimization function reaches the minimum value, corresponding w 1 And w 2 Is the optimized weight. />
CN202310370279.1A 2023-04-07 2023-04-07 Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm Pending CN116362355A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310370279.1A CN116362355A (en) 2023-04-07 2023-04-07 Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310370279.1A CN116362355A (en) 2023-04-07 2023-04-07 Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm

Publications (1)

Publication Number Publication Date
CN116362355A true CN116362355A (en) 2023-06-30

Family

ID=86908441

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310370279.1A Pending CN116362355A (en) 2023-04-07 2023-04-07 Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm

Country Status (1)

Country Link
CN (1) CN116362355A (en)

Similar Documents

Publication Publication Date Title
Boussabaine The use of artificial neural networks in construction management: a review
CN116757534B (en) Intelligent refrigerator reliability analysis method based on neural training network
CN109858714B (en) Tobacco shred quality inspection index prediction method, device and system based on improved neural network
Zhou et al. A new type of recurrent fuzzy neural network for modeling dynamic systems
CN116150897A (en) Machine tool spindle performance evaluation method and system based on digital twin
CN106934242B (en) The health degree appraisal procedure and system of equipment under multi-mode based on Cross-Entropy Method
CN116029221B (en) Power equipment fault diagnosis method, device, equipment and medium
CN106610584A (en) Remanufacturing process quality control method based on neural network and expert system
CN108961460B (en) Fault prediction method and device based on sparse ESGP (Enterprise service gateway) and multi-objective optimization
CN112651519A (en) Secondary equipment fault positioning method and system based on deep learning theory
CN104732067A (en) Industrial process modeling forecasting method oriented at flow object
CN114548494B (en) Visual cost data prediction intelligent analysis system
CN115902642A (en) Battery state of charge estimation method and device, electronic equipment and storage medium
CN116303786A (en) Block chain financial big data management system based on multidimensional data fusion algorithm
JP3812358B2 (en) Plant operation support method and program thereof
CN116362355A (en) Method for predicting cable joint guide core temperature by adopting XGBoost and random forest algorithm
CN112257893A (en) Complex electromechanical system health state prediction method considering monitoring error
CN116484904A (en) Supervision data processing implementation method based on artificial intelligence deep learning
CN115659201A (en) Gas concentration detection method and monitoring system for Internet of things
Alrubaie Cascade-Forward neural network for volterra integral equation solution
CN113987904A (en) Method, device, equipment and storage medium for measuring and calculating repair cost of power transmission project
CN112651168B (en) Construction land area prediction method based on improved neural network algorithm
Ahmed et al. Study of genetic algorithm to fully-automate the design and training of artificial neural network
CN115330064A (en) Human-machine decision logic online optimization method for highly automatic driving
JP4267726B2 (en) Device for determining relationship between operation signal and operation amount in control device, control device, data generation device, input / output characteristic determination device, and correlation evaluation device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination