Multi-source data driven aerial material demand prediction method, system and equipment and storage medium

Abstract

The invention discloses a multi-source data driven aerial material demand prediction method, system and device and a storage medium. The method comprises the following steps: obtaining multi-source data of an aerial material demand according to obtained related information of the aerial material demand; the multi-source data comprises maintenance record data, aircraft operation data, sensor data, environment data and supply chain data; then preprocessing the multi-source data; carrying out feature extraction on the preprocessed multi-source data to obtain feature data; the feature data comprises historical feature data, operation feature data, real-time feature data and environment feature data; and finally, constructing an aerial material demand prediction model according to the feature data. According to the method, through preprocessing and feature extraction of the multi-source data, various complex factors influencing the aerial material demand can be known more deeply, the accuracy of the prediction result of the aerial material demand prediction model can be improved, and thus the operation cost of an airline company is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of aviation material management, and in particular to a method, system, device and storage medium for predicting aviation material demand driven by multi-source data. Background Art

[0002] As the scale of airline operations continues to expand and the number of aircraft increases year by year, the importance of aviation material (aviation materials and parts) management has become increasingly prominent. The demand management of aviation materials directly affects the operating efficiency, maintenance costs and flight safety of airlines. Insufficient aviation material supply will lead to problems such as aircraft grounding and maintenance delays, which in turn affect the operating income of airlines; while excessive aviation material inventory will increase inventory costs and cause capital waste. How to accurately predict the demand for aviation materials and ensure the reasonable inventory and timely supply of aviation materials has become a key issue for airlines to improve operational efficiency and reduce costs.

[0003] Traditional methods of forecasting demand for aviation materials mainly rely on historical data and empirical rules. They estimate future demand for aviation materials by analyzing past maintenance records and using time series models or simple statistical methods such as mean and regression.

[0004] The inventors have found through research that the aviation material demand forecasting method in the prior art still has at least the following defects:

[0005] The accuracy of aviation material demand forecast is low, which affects the operational efficiency of airlines.

[0006] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgment or any form of suggestion that the information constitutes the prior art already known to a person skilled in the art. Summary of the invention

[0007] The purpose of the present invention is to improve the accuracy of aviation material demand forecasting.

[0008] The present invention provides a method for forecasting aviation material demand driven by multi-source data, comprising the steps of:

[0009] S11, obtaining multi-source data of aviation material demand according to the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data;

[0010] S12, preprocessing the multi-source data; and extracting features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data;

[0011] S13. Constructing an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

[0012] Preferably, in the embodiment of the present invention,

[0013] The historical characteristic data include: frequency of aviation material consumption, maintenance cycle and usage time;

[0014] The operational characteristic data include: the number of flights, flight duration and flight environment of the aircraft;

[0015] The real-time characteristic data includes: real-time status information of aviation materials;

[0016] The environmental characteristic data include: environmental factors related to aircraft flight.

[0017] Preferably, in an embodiment of the present invention, constructing the aviation material demand forecasting model includes:

[0018] Time series models, machine learning models, and deep learning models.

[0019] Preferably, in the embodiment of the present invention,

[0020] The time series model includes: ARIMA model or exponential smoothing model;

[0021] The machine learning model includes: a random forest model or an XGBoost model;

[0022] The deep learning model includes: a long short-term memory network model.

[0023] Preferably, in the embodiment of the present invention,

[0024] The ARIMA model includes autoregression, difference and moving average, which is specifically the following formula:

[0025]

[0026] In the formula, Xt is the demand at the current time point; p is the order of the autoregressive term; q is the order of the moving average term; φ i is the autoregression coefficient; θ j is the moving average coefficient; ∈t is the error term;

[0027] The exponential smoothing model includes the following formula:

[0028] St=αXt+(1-α)St-1

[0029] In the formula, St is the smoothing value at time point t; α is the smoothing coefficient;

[0030] The random forest model includes the following formula:

[0031]

[0032] In the formula, is the predicted value of aviation material demand; n is the number of decision trees; f i (x) is the output of the i-th decision tree;

[0033] The XGBoost model includes the following formula:

[0034]

[0035] In the formula, is the predicted value of aviation material demand; k is the learning rate; f k (x) is the predicted value of the kth tree; K is the total number of trees.

[0036] Preferably, in an embodiment of the present invention, the long short-term memory network model includes: a forget gate, an input gate and an output gate; wherein,

[0037] The forget gate is calculated by the following formula:

[0038] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0039] In the formula, f t For the forget gate; W f is the weight matrix of the forget gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b f is the bias term of the forget gate;

[0040] The input gate is calculated by the following formula:

[0041] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0042] In the formula, i t is the input gate; W i is the weight matrix of the input gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b i is the bias term of the input gate;

[0043] The output gate is calculated by the following formula:

[0044] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0045] In the formula, o t is the output gate; W o is the weight matrix of the output gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b o is the bias term of the output gate.

[0046] Preferably, in an embodiment of the present invention, the prediction results of the aviation material demand prediction model include:

[0047] Demand forecasts, inventory optimization recommendations and visualizations.

[0048] In another aspect of the present invention, a multi-source data-driven aviation material demand forecasting system is also provided, comprising:

[0049] A multi-source data acquisition module is used to obtain multi-source data on aviation material demand based on the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data;

[0050] A multi-source data preprocessing module is used to preprocess the multi-source data; and extract features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data;

[0051] The aviation material demand prediction model building module is used to build an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

[0052] On the other hand, a multi-source data-driven aviation material demand forecasting device is provided; the multi-source data-driven aviation material demand forecasting device includes a computer program stored on a medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the methods described in the above aspects and achieves the same technical effects.

[0053] On the other hand, a storage medium is provided on which a computer program is stored. When the computer program is executed by a processor, each step of the multi-source data-driven aviation material demand forecasting method as described in any one of the above items is implemented.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] In the present invention, firstly, multi-source data of aviation material demand is obtained according to the obtained information related to aviation material demand; then, the multi-source data is preprocessed; and feature extraction is performed on the preprocessed multi-source data to obtain feature data; finally, an aviation material demand prediction model is constructed according to each feature data; by preprocessing and feature extraction of the multi-source data, a deeper understanding of various complex factors affecting aviation material demand can be achieved, and the accuracy of the prediction results of the aviation material demand prediction model can be improved, thereby reducing the operating costs of airlines.

[0056] Furthermore, the present invention dynamically updates the prediction results according to the real-time collected aviation material status data, thereby ensuring the accuracy and timeliness of the aviation material demand prediction.

[0057] In addition, the present invention can accurately predict the remaining life of aircraft materials, helping airlines optimize the use strategy of aircraft materials. Through early warning and regular maintenance, excessive wear and tear of aircraft materials or early replacement of aircraft materials can be avoided, thus extending the actual service life of aircraft materials.

[0058] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, and to make the above and other purposes, technical features and advantages of the present invention easier to understand, one or more preferred embodiments are listed below and described in detail with the accompanying drawings as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solution of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 It is a step diagram of the aviation material demand forecasting method driven by multi-source data in the present invention;

[0061] Figure 2 is a flow chart of the aviation material demand forecasting method driven by multi-source data in the present invention;

[0062] Figure 3 It is a flow chart of constructing the aviation material demand forecasting model described in the present invention;

[0063] Figure 4 It is a structural schematic diagram of the aviation material demand forecasting system driven by multi-source data in the present invention;

[0064] Figure 5 It is a schematic diagram of the structure of the aviation material demand forecasting device driven by multi-source data in the present invention. DETAILED DESCRIPTION

[0065] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific implementation modes.

[0066] Unless explicitly stated otherwise, throughout the specification and claims, the term “comprise” or variations such as “include” or “comprising”, etc., will be understood to include the stated elements or components but not to exclude other elements or components.

[0067] In this document, the terms "first", "second", etc. are used to distinguish two different elements or parts, and are not used to limit a specific position or relative relationship. In other words, in some embodiments, the terms "first", "second", etc. can also be interchangeable.

[0068] Embodiment 1

[0069] In order to improve the accuracy of aviation material demand forecast, Figure 1 and Figure 2 As shown, in an embodiment of the present invention, a method for forecasting aviation material demand driven by multi-source data is provided, comprising the steps of:

[0070] S11, obtaining multi-source data of aviation material demand according to the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data;

[0071] In the embodiment of the present invention, data information related to aviation material demand is obtained from multiple sources, including:

[0072] Maintenance record data: maintenance frequency, maintenance time, maintenance cost, degree of wear and tear, and types of parts repaired or replaced. This type of data is the basic data source for analyzing aviation material demand.

[0073] Aircraft operation data: data such as the aircraft's flight duration, number of flights, routes, flight speed, and flight altitude. Different flight modes and flight environments will affect the consumption of aviation materials, so these data are crucial to the construction of an aviation material demand forecasting model.

[0074] Real-time sensor data: IoT sensors installed on aircraft can monitor the status of aviation materials in real time, such as wear, temperature, pressure, vibration and other parameters. Sensor data provides a key basis for real-time prediction of aviation material status and demand.

[0075] Environmental data: including meteorological information in the flight environment, such as temperature, humidity, wind speed, air pressure and other external environmental factors during the flight. Environmental conditions directly affect the service life and consumption patterns of aviation materials and are important input variables for the aviation material demand forecasting model.

[0076] Supply chain data: Contains information such as supplier delivery cycles, inventory levels, shipping times, and supplier performance. This data helps optimize aviation material procurement and inventory management.

[0077] In addition, it can also include: timestamp, that is, the time point information of various types of data, which will also play a key role in building an aviation material demand forecasting model.

[0078] S12, preprocessing the multi-source data; and extracting features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data;

[0079] In order to ensure the accuracy and integrity of the data, in the embodiment of the present invention, the collected multi-source data is first preprocessed, including data cleaning, outlier detection, missing value filling and data format conversion; wherein,

[0080] Data cleaning removes duplicate, invalid and noisy data, unifies data formats, and ensures data accuracy and consistency. It is applicable to situations where there may be duplicate records, inconsistent formats or invalid data when collecting data from multiple sources, such as negative flight time, inconsistent date formats, etc. By deleting unreasonable records, formatting dates or units, etc., the original data is cleaned and standardized to ensure that subsequent analysis can be performed on a clean and structured data set, avoiding interference from invalid data when training the aviation material demand forecasting model.

[0081] Outlier detection is used to identify and process abnormal points in a data set that are significantly different from normal values, ensuring that the model is not disturbed by extreme values. It is applicable to situations where there may be extreme deviations in the data, such as extreme values ​​read by sensors or extreme weather conditions in climate data. In the embodiments of the present invention, ZScore, IQR (interquartile range) and machine learning models (such as Isolation Forest, LOF) can be used. These methods can effectively filter out data points that are far from the normal distribution, thereby improving the stability of the model and the prediction accuracy.

[0082] Missing value filling fills the gaps in the data through appropriate strategies to ensure data integrity. It is applicable to data missing problems in cases of incomplete data collection or sensor failure, such as incomplete flight environment monitoring data or incomplete aviation material maintenance records. Depending on the scenario, you can choose to delete missing values, or fill them through mean / median, interpolation, KNN algorithm or model-based prediction methods. Reasonable filling of missing values ​​can prevent deviations caused by incomplete data and ensure effective learning of the model.

[0083] Data format conversion ensures the consistency of all input data and facilitates model processing. It is suitable for scenarios where multi-source data formats are not unified or categorical variables cannot be directly input into the model, such as inconsistent date formats, different temperature units, or categorical variables such as "normal / abnormal" need to be processed numerically. The format and unit of the data are unified through the standardization of numerical and date types, and the encoding of categorical variables (such as one-hot encoding or label encoding). In addition, through normalization or standardization operations, the consistency of feature dimensions is ensured, so that the model can better understand the data characteristics and improve the training effect. The preprocessing process ensures data quality and lays a solid foundation for the construction of subsequent models.

[0084] Further, in an embodiment of the present invention, after the multi-source data preprocessing is completed, feature data highly related to the demand for aviation materials is extracted from the multi-source data, including: historical feature data, operational feature data, real-time feature data and environmental feature data;

[0085] Historical feature data includes: extracting historical features such as aviation material consumption frequency, maintenance cycle, and usage time from aircraft maintenance record data and operation data;

[0086] Operational characteristic data include: aircraft flight times, flight duration, flight environment and other operational characteristic data that affect aviation material demand;

[0087] Real-time characteristic data includes: extracting real-time status information of aviation materials, such as wear degree, temperature, pressure, vibration, etc.;

[0088] Environmental characteristic data include: environmental factors related to flight, such as weather conditions encountered during flight.

[0089] In order to select the key features that best reflect the changes in aviation material demand, a variety of machine learning algorithms are used in the embodiment of the present invention. Feature selection is performed through random forests and support vector machines to screen out features that have a significant impact on aviation material demand forecasting. These key feature data are used as input variables for constructing an aviation material demand forecasting model.

[0090] S13. Constructing an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

[0091] In the embodiment of the present invention, the characteristic data is used as the input variable to construct an aviation material demand prediction model, specifically as follows: Figure 3 shown.

[0092] Construct an aviation material demand forecasting model, including:

[0093] Time series models, machine learning models, and deep learning models.

[0094] The time series model predicts future demand for aviation materials by analyzing the trends and periodicity of historical data. It is suitable for aviation materials with stable demand and obvious periodicity.

[0095] Includes: ARIMA model or exponential smoothing model;

[0096] The ARIMA (Autoregressive Integrated Moving Average) model is used to process stationary time series data and can capture trends and seasonality in the data; it includes autoregression, difference, and moving average, specifically the following formula:

[0097]

[0098] In the formula, Xt is the demand at the current time point; p is the order of the autoregressive term; q is the order of the moving average term; φ i is the autoregression coefficient; θ j is the moving average coefficient; ∈t is the error term;

[0099] The exponential smoothing model predicts future demand by weighting historical data. It is suitable for short-term demand forecasting and includes the following formula:

[0100] St=αXt+(1-α)St-1

[0101] In the formula, St is the smoothing value at time point t; α is the smoothing coefficient.

[0102] In constructing the time series model, the characteristic data is decomposed into time series to separate the trend, seasonality and residual, the ARIMA model or exponential smoothing model is selected, the model parameters (such as the p value, d value, and q value in ARIMA) are adjusted, and then the model is used to fit the historical aviation material demand. The model is adjusted by minimizing the error term, and the fitted model is used to predict the future aviation material demand and generate the demand curve within the time period.

[0103] In the embodiment of the present invention, the machine learning model is suitable for scenarios where there are complex nonlinear relationships in the data and historical data is combined with real-time data; the machine learning model includes: a random forest model or an XGBoost (distributed gradient boosting library) model; wherein,

[0104] The random forest model includes the following formula:

[0105]

[0106] In the formula, is the predicted value of aviation material demand; n is the number of decision trees; f i (x) is the output of the i-th decision tree;

[0107] The XGBoost model is an enhanced decision tree model that reduces prediction errors by gradually building a tree, including the following formulas:

[0108]

[0109] In the formula, is the predicted value of aviation material demand; k is the learning rate; f k (x) is the predicted value of the kth tree; K is the total number of trees.

[0110] In the machine learning model, the feature data is divided into a training set and a test set, and then a random forest or XGBoost model is selected, and hyperparameters (such as the number of decision trees, maximum depth, etc.) are set. The model is trained with the training set, and the decision tree is gradually optimized to reduce errors. Then the model is evaluated with the test set, and the model performance is further improved by adjusting hyperparameters (such as learning rate). Finally, the trained model is used to perform regression prediction on future aviation material demand and output the predicted value.

[0111] In order to capture long-term dependencies and complex nonlinear patterns, a deep learning model is used in an embodiment of the present invention. This model is suitable for dynamic and complex flight environments and aviation material consumption patterns.

[0112] Deep learning models include: Long Short-Term Memory Network (LSTM) model; The Long Short-Term Memory Network model is a deep learning model specially designed to process time series. It can remember data features over a long time span and is used to predict future aviation material demand.

[0113] Long short-term memory network model, including:

[0114] Forget gate, input gate and output gate; among them,

[0115] The forget gate is calculated by the following formula:

[0116] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0117] In the formula, f t For the forget gate; Wf is the weight matrix of the forget gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b f is the bias term of the forget gate;

[0118] The input gate is calculated by the following formula:

[0119] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0120] In the formula, i t is the input gate; W i is the weight matrix of the input gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b i is the bias term of the input gate;

[0121] The output gate is calculated by the following formula:

[0122] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0123] In the formula, o t is the output gate; W o is the weight matrix of the output gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b o is the bias term of the output gate.

[0124] In the deep learning model, the historical time series data is first normalized to ensure that the data range is consistent. Secondly, the structure of the LSTM network is designed, including the input layer, the hidden layer (LSTM layer), and the output layer, and the appropriate time window and number of hidden units are selected. Then, the LSTM model is trained using historical data, and the model parameters are optimized through back propagation to minimize the loss function (such as mean square error). Finally, the trained LSTM model is used to predict future aviation material demand, which is particularly suitable for long-term demand change prediction. In the embodiment of the present invention, the LSTM model is used to capture long-term dependencies and dynamic changes, and performs well in predicting aviation material demand in complex flight environments.

[0125] In the embodiment of the present invention, by introducing IoT sensors and real-time data collection, the aviation material demand forecasting model can be updated in real time; when the aircraft's operating status or external environment changes, the parameters of the forecasting model can be automatically adjusted and the forecast results can be updated according to the latest sensor data and environmental data. Specifically:

[0126] Real-time update: Regularly obtain real-time data from aircraft and dynamically adjust model parameters to ensure that the prediction model always reflects the latest operating status and environmental changes.

[0127] Adaptive algorithm: In order to improve the robustness of the model, an adaptive algorithm is introduced in the embodiment of the present invention to adjust the structure and parameters of the model according to real-time data, so that the prediction model can self-learn and optimize and adapt to changing actual needs.

[0128] Preferably, in the embodiment of the present invention, the aviation material demand forecasting model can output forecasting results in various forms to provide strong support for airlines in aviation material management decisions. Specifically, it includes:

[0129] Demand forecast value: output the demand for specific aviation materials in a certain period of time in the future, including the expected replacement time, emergency demand, etc.;

[0130] Inventory optimization suggestions: Based on the forecast results, optimization suggestions for aviation material inventory management can be generated, including optimal procurement quantity, inventory scheduling and other information.

[0131] Visual display: By displaying the forecast results in a visual way, users can intuitively understand the demand trends and changes of aviation materials through charts and other forms, which can assist in decision-making.

[0132] To summarize, in an embodiment of the present invention, first, multi-source data on aviation material demand is obtained based on the acquired information related to aviation material demand; then, the multi-source data is preprocessed; and feature extraction is performed on the preprocessed multi-source data to obtain feature data; finally, an aviation material demand prediction model is constructed based on each feature data; by preprocessing and feature extraction of the multi-source data, a deeper understanding of the various complex factors affecting aviation material demand can be achieved, and the accuracy of the prediction results of the aviation material demand prediction model can be improved, thereby reducing the operating costs of airlines.

[0133] Furthermore, the present invention dynamically updates the prediction results according to the real-time collected aviation material status data, thereby ensuring the accuracy and timeliness of the aviation material demand prediction.

[0134] In addition, the present invention can accurately predict the remaining life of aircraft materials, helping airlines optimize the use strategy of aircraft materials. Through early warning and regular maintenance, excessive wear and tear of aircraft materials or early replacement of aircraft materials can be avoided, thus extending the actual service life of aircraft materials.

[0135] Embodiment 2

[0136] Corresponding to the method embodiment, another aspect of the present invention embodiment further provides an aviation material demand forecasting system driven by multi-source data. Figure 4 The schematic diagram of the structure of the multi-source data driven aviation material demand forecasting system provided by the embodiment of the present invention is shown. Figure 1 The system corresponding to the multi-source data driven aviation material demand forecasting method described in the corresponding embodiment, that is, implemented by a virtual system Figure 1 In the corresponding embodiment, the aviation material demand forecasting method based on multi-source data drive, each virtual module constituting the high-precision, low-latency, high-bandwidth communication scheduling system based on timing flow control can be executed by electronic devices, such as network devices, terminal devices or servers. Specifically, the aviation material demand forecasting system driven by multi-source data in the embodiment of the present invention includes:

[0137] The multi-source data acquisition module 01 is used to obtain multi-source data of aviation material demand based on the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data;

[0138] In the embodiment of the present invention, the multi-source data includes:

[0139] Maintenance record data: maintenance frequency, maintenance time, maintenance cost, degree of wear and tear, and types of parts repaired or replaced. This type of data is the basic data source for analyzing aviation material demand.

[0140] Aircraft operation data: data such as the aircraft's flight duration, number of flights, routes, flight speed, and flight altitude. Different flight modes and flight environments will affect the consumption of aviation materials, so these data are crucial to the construction of an aviation material demand forecasting model.

[0141] Real-time sensor data: IoT sensors installed on aircraft can monitor the status of aviation materials in real time, such as wear, temperature, pressure, vibration and other parameters. Sensor data provides a key basis for real-time prediction of aviation material status and demand.

[0142] Environmental data: including meteorological information in the flight environment, such as temperature, humidity, wind speed, air pressure and other external environmental factors during the flight. Environmental conditions directly affect the service life and consumption patterns of aviation materials and are important input variables for the aviation material demand forecasting model.

[0143] Supply chain data: Contains information such as supplier delivery cycles, inventory levels, shipping times, and supplier performance. This data helps optimize aviation material procurement and inventory management.

[0144] In addition, it can also include: timestamp, that is, the time point information of various types of data, which will also play a key role in building an aviation material demand forecasting model.

[0145] The multi-source data preprocessing module 02 is used to preprocess the multi-source data; and extract features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data;

[0146] In order to ensure the accuracy and completeness of the data, in the embodiment of the present invention, the collected multi-source data is first preprocessed, including data cleaning, outlier detection, missing value filling and data format conversion.

[0147] Further, in an embodiment of the present invention, after the multi-source data preprocessing is completed, feature data highly related to the demand for aviation materials is extracted from the multi-source data, including: historical feature data, operational feature data, real-time feature data and environmental feature data;

[0148] Historical feature data includes: extracting historical features such as aviation material consumption frequency, maintenance cycle, and usage time from aircraft maintenance record data and operation data;

[0149] Operational characteristic data include: aircraft flight times, flight duration, flight environment and other operational characteristic data that affect aviation material demand;

[0150] Real-time characteristic data includes: extracting real-time status information of aviation materials, such as wear degree, temperature, pressure, vibration, etc.;

[0151] Environmental characteristic data include: environmental factors related to flight, such as weather conditions encountered during flight.

[0152] In order to select the key features that best reflect the changes in aviation material demand, a variety of machine learning algorithms are used in the embodiment of the present invention. Feature selection is performed through random forests and support vector machines to screen out features that have a significant impact on aviation material demand forecasting. These key feature data are used as input variables for constructing an aviation material demand forecasting model.

[0153] The aviation material demand prediction model building module 03 is used to build an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

[0154] In an embodiment of the present invention, building an aviation material demand forecasting model includes:

[0155] Time series models, machine learning models, and deep learning models.

[0156] The time series model predicts future demand for aviation materials by analyzing the trends and periodicity of historical data. It is suitable for aviation materials with stable demand and obvious periodicity.

[0157] Includes: ARIMA model or exponential smoothing model;

[0158] The ARIMA (Autoregressive Integrated Moving Average) model is used to process stationary time series data and can capture trends and seasonality in the data; it includes autoregression, difference, and moving average, specifically the following formula:

[0159]

[0160] In the formula, Xt is the demand at the current time point; p is the order of the autoregressive term; q is the order of the moving average term; φ i is the autoregression coefficient; θ j is the moving average coefficient; ∈ t is the error term;

[0161] The exponential smoothing model predicts future demand by weighting historical data. It is suitable for short-term demand forecasting and includes the following formula:

[0162] St=αXt+(1-α)St-1

[0163] In the formula, St is the smoothing value at time point t; α is the smoothing coefficient.

[0164] In constructing the time series model, the characteristic data is decomposed into time series to separate the trend, seasonality and residual, the ARIMA model or exponential smoothing model is selected, the model parameters (such as the p value, d value, and q value in ARIMA) are adjusted, and then the model is used to fit the historical aviation material demand. The model is adjusted by minimizing the error term, and the fitted model is used to predict the future aviation material demand and generate the demand curve within the time period.

[0165] In the embodiment of the present invention, the machine learning model is suitable for scenarios where there are complex nonlinear relationships in the data and historical data is combined with real-time data; the machine learning model includes: a random forest model or an XGBoost (distributed gradient boosting library) model; wherein,

[0166] The random forest model includes the following formula:

[0167]

[0168] In the formula, is the predicted value of aviation material demand; n is the number of decision trees; f i (x) is the output of the i-th decision tree;

[0169] The XGBoost model is an enhanced decision tree model that reduces prediction errors by gradually building a tree, including the following formulas:

[0170]

[0171] In the formula, is the predicted value of aviation material demand; k is the learning rate; f k (x) is the predicted value of the kth tree; K is the total number of trees.

[0172] In the machine learning model, the feature data is divided into a training set and a test set, and then a random forest or XGBoost model is selected, and hyperparameters (such as the number of decision trees, maximum depth, etc.) are set. The model is trained with the training set, and the decision tree is gradually optimized to reduce errors. Then the model is evaluated with the test set, and the model performance is further improved by adjusting hyperparameters (such as learning rate). Finally, the trained model is used to perform regression prediction on future aviation material demand and output the predicted value.

[0173] In order to capture long-term dependencies and complex nonlinear patterns, a deep learning model is used in an embodiment of the present invention. This model is suitable for dynamic and complex flight environments and aviation material consumption patterns.

[0174] Deep learning models include: Long Short-Term Memory Network (LSTM) model; The Long Short-Term Memory Network model is a deep learning model specially designed to process time series. It can remember data features over a long time span and is used to predict future aviation material demand.

[0175] Long short-term memory network model, including:

[0176] Forget gate, input gate and output gate; among them,

[0177] The forget gate is calculated by the following formula:

[0178] f t =σ(W f ·[h t-1 ,x t ]+b f )

[0179] In the formula, f t For the forget gate; W f is the weight matrix of the forget gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b f is the bias term of the forget gate.

[0180] The input gate is calculated by the following formula:

[0181] i t =σ(W i ·[h t-1 ,x t ]+b i )

[0182] In the formula, i t is the input gate; W i is the weight matrix of the input gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b i is the bias term of the input gate.

[0183] The output gate is calculated by the following formula:

[0184] o t =σ(W o ·[h t-1 ,x t ]+b o )

[0185] In the formula, o t is the output gate; W o is the weight matrix of the output gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b o is the bias term of the output gate.

[0186] It should be noted that the specific implementation and technical effects of the multi-source data driven aviation material demand forecasting system in the embodiment of the present invention can be referred to in Figure 1 The corresponding multi-source data-driven aviation material demand forecasting method will not be elaborated here.

[0187] Embodiment 3

[0188] Corresponding to the method embodiment, in the embodiment of the present invention, a multi-source data-driven aviation material demand forecasting device is also provided, such as a terminal, a server, etc. Among them, the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto.

[0189] An example of a hardware structure block diagram of a multi-source data driven aviation material demand forecasting device provided by an embodiment of the present invention is shown in FIG. Figure 5As shown, it may include:

[0190] Processor 1, communication interface 2, memory 3 and communication bus 4;

[0191] The processor 1, the communication interface 2, and the memory 3 communicate with each other via the communication bus 4;

[0192] Optionally, the communication interface 2 may be an interface of a communication module, such as an interface of a GSM module;

[0193] The processor 1 may be a central processing unit CPU, or an application-specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0194] The memory 3 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0195] The processor 1 is specifically used to execute the computer program stored in the memory 3 to perform the following steps:

[0196] S11, obtaining multi-source data of aviation material demand according to the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data;

[0197] S12, preprocessing the multi-source data; and extracting features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data;

[0198] S13. Constructing an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

[0199] The above product can execute the method provided by the embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not described in detail in this embodiment, please refer to the multi-source data driven aviation material demand forecasting method provided by the embodiment of the present invention.

[0200] Embodiment 4

[0201] In an embodiment of the present invention, a storage medium is further provided, which may store a program suitable for execution by a processor, wherein the program is used to:

[0202] S11, obtaining multi-source data of aviation material demand according to the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data;

[0203] S12, preprocessing the multi-source data; and extracting features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data;

[0204] S13. Constructing an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

[0205] Optionally, the detailed functions and extended functions of the program may refer to the above description.

[0206] The above product can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the methods provided by other embodiments of the present invention.

[0207] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0208] In the several embodiments provided in this application, it should be understood that the disclosed systems, systems and methods can be implemented in other ways. Another point is that the coupling or direct coupling or communication connection between each other shown or discussed can be an indirect coupling or communication connection through some interface, system or unit, which can be electrical, mechanical or other forms.

[0209] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0210] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0211] It should be understood that in the embodiments of the present application, the various embodiments and features can be combined with each other to solve the aforementioned technical problems.

[0212] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0213] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-source data-driven aviation material demand forecasting method, characterized in that: Includes steps: S11, obtaining multi-source data of aviation material demand according to the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data; S12, preprocessing the multi-source data; and extracting features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data; S13. Constructing an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

2. The multi-source data-driven aviation material demand forecasting method according to claim 1, characterized in that: The historical characteristic data include: frequency of aviation material consumption, maintenance cycle and usage time; The operational characteristic data include: the number of flights, flight duration and flight environment of the aircraft; The real-time characteristic data includes: real-time status information of aviation materials; The environmental characteristic data include: environmental factors related to aircraft flight.

3. The multi-source data-driven aviation material demand forecasting method according to claim 2, characterized in that: Constructing the aviation material demand forecasting model includes: Time series models, machine learning models, and deep learning models.

4. The multi-source data-driven aviation material demand forecasting method according to claim 3, characterized in that: The time series model includes: ARIMA model or exponential smoothing model; The machine learning model includes: a random forest model or an XGBoost model; The deep learning model includes: a long short-term memory network model.

5. The multi-source data-driven aviation material demand forecasting method according to claim 4, characterized in that: The ARIMA model includes autoregression, difference and moving average, which is specifically the following formula: In the formula, Xt is the demand at the current time point; p is the order of the autoregressive term; q is the order of the moving average term; φ i is the autoregression coefficient; θ j is the moving average coefficient; ∈t is the error term; The exponential smoothing model includes the following formula: St=αXt+(1-α)St-1 In the formula, St is the smoothing value at time point t; α is the smoothing coefficient; The random forest model includes the following formula: In the formula, is the predicted value of aviation material demand; n is the number of decision trees; f i (x) is the output of the i-th decision tree; The XGBoost model includes the following formula: In the formula, is the predicted value of aviation material demand; η k is the learning rate; f k (x) is the predicted value of the kth tree; K is the total number of trees.

6. The multi-source data-driven aviation material demand forecasting method according to claim 5, characterized in that: The long short-term memory network model comprises: Forget gate, input gate and output gate; among them, The forget gate is calculated by the following formula: f t =σ(W f ·[h t-1 ,x t ]+b f ) In the formula, f t For the forget gate; W f is the weight matrix of the forget gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b f is the bias term of the forget gate; The input gate is calculated by the following formula: i t =σ(W i ·[h t-1 ,x t ]+b i ) In the formula, i t is the input gate; W i is the weight matrix of the input gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b i is the bias term of the input gate; The output gate is calculated by the following formula: the t =σ(W o ·[h t-1 ,x t ]+b o ) In the formula, o t is the output gate; W o is the weight matrix of the output gate; h t-1 is the hidden state of the previous time step; x t is the input of the current time step; σ is the activation function; b o is the bias term of the output gate.

7. The multi-source data-driven aviation material demand forecasting method according to claim 1, characterized in that: The prediction results of the aviation material demand prediction model include: Demand forecasts, inventory optimization recommendations and visualizations.

8. A multi-source data-driven aviation material demand forecasting system, characterized in that: include: A multi-source data acquisition module is used to obtain multi-source data on aviation material demand based on the acquired information related to aviation material demand; the multi-source data includes: maintenance record data, aircraft operation data, sensor data, environmental data and supply chain data; A multi-source data preprocessing module is used to preprocess the multi-source data; and extract features from the preprocessed multi-source data to obtain feature data; the feature data includes: historical feature data, operation feature data, real-time feature data and environmental feature data; The aviation material demand prediction model building module is used to build an aviation material demand prediction model according to each of the characteristic data; the aviation material demand prediction model is used to predict the demand for different types of aviation materials.

9. A multi-source data-driven aviation material demand forecasting device, characterized in that: include: Memory for storing computer programs; A processor is used to call and execute the computer program to implement the steps of the multi-source data driven aviation material demand forecasting method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The method comprises a software program, wherein the software program is suitable for executing, by a processor, the steps of the aviation material demand forecasting method driven by multi-source data as claimed in any one of claims 1 to 7.

Images