Bacterial colony abundance prediction method and device based on environmental ecological indexes, and electronic equipment
By constructing the environmental ecology-colony abundance dataset and using the timing neural network to train a multi-scale prediction model, the existing colony abundance detection problems are solved, and efficient and low-cost colony abundance prediction and dynamic change monitoring are achieved.
Patent Information
- Application Number
- CN202510439725.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-09
- Publication Date
- 2025-07-11
AI Technical Summary
The existing colony abundance detection methods are inefficient and costly, and cannot grasp dynamic changes in time, and rely on high-precision instruments and professional testing personnel.
The environmental ecology-colony abundance data set was constructed, and the colony abundance multi-scale prediction model was trained using the timing neural network, and the colony abundance prediction model was predicted through environmental ecology indicators, combined with the weighted averages of multiple prediction models, so as to achieve no need for regular sampling and laboratory testing.
It improves the efficiency of colony abundance detection, reduces the cost of equipment maintenance and detection difficulty, can timely grasp the dynamic changes in colony abundance, and achieves early identification of potential threats.
Smart Images
Figure CN120299528A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of ecological environment detection, and particularly relates to a method and device for predicting colony abundance based on environmental ecological indicators, and an electronic device. Background Art
[0002] In ecological environment detection, colony abundance, as an important indicator for characterizing the changes of microbial communities, can reflect the health status of ecosystems. Accurately predicting the colony abundance and its change trend in the environment is of positive significance for ecological environment detection and protection. For example, the abundance change of specific bacterial groups (such as functional microorganisms like Pseudomonas) can reflect the types of pollutants and their potential toxicity; the abnormal ratio of nitrogen-fixing bacteria / nitrifying bacteria can warn of soil acidification or fertility degradation (for example, in an environment with a pH lower than 5.5, the nitrogen-fixing activity will significantly decrease, affecting the nitrogen cycle).
[0003] Traditional methods for detecting colony abundance mainly rely on regular sampling and laboratory testing. For example, conventional methods such as plate counting method and MPN method are used, or molecular biology methods such as real-time quantitative PCR and high-throughput sequencing analysis are adopted. However, due to the cumbersome operation steps of sampling and detection, long cycle, and low efficiency, these methods cannot meet the needs of daily detection; and they require high-precision instruments and rely on professional testers, with high equipment maintenance costs, strict detection condition limitations, and high requirements for the experience and skills of testers, making it difficult to timely grasp the dynamic changes of colony abundance. Summary of the Invention
[0004] The present invention aims to solve the problems of low efficiency and high cost existing in the existing methods for detecting colony abundance, and proposes a method and device for predicting colony abundance based on environmental ecological indicators, and an electronic device.
[0005] The technical solutions adopted by the present invention to solve the above technical problems are as follows:
[0006] In a first aspect, the present invention provides a method for predicting colony abundance based on environmental ecological indicators, the method comprising:
[0007] Construct an environmental ecology - colony abundance data set, the environmental ecology - colony abundance data set comprising time series data of environmental ecological indicators and colony abundance;
[0008] Train a time series neural network according to the environmental ecology - colony abundance data set to construct a colony abundance multi-scale prediction model, the colony abundance multi-scale prediction model comprising a plurality of colony abundance prediction models;
[0009] According to the environmental ecological indicators of the environmental area to be predicted and based on the colony abundance multi-scale prediction model, obtain the multi-scale prediction result of the colony abundance in the environmental area to be predicted at the corresponding time, and the multi-scale prediction result of the microecological indicators is the weighted average of the colony abundance prediction results of multiple colony abundance prediction models.
[0010] Further, the colony abundance prediction model includes a first colony abundance prediction model and a second colony abundance prediction model;
[0011] The first colony abundance prediction model is used to predict the colony abundance prediction result in the environmental area to be predicted at the corresponding time according to the environmental ecological indicators of the environmental area to be predicted;
[0012] The second colony abundance prediction model is used to predict the colony abundance prediction result in the environmental area to be predicted at the corresponding time according to the environmental ecological indicators and the historical time series data of the colony abundance in the environmental area to be predicted.
[0013] Further, the environmental ecological indicators include multiple of the daily maximum temperature, daily minimum temperature, daily average temperature, daily average humidity, daily cumulative rainfall, daily average air pressure, and daily average wind speed;
[0014] The colony abundance includes at least the relative abundance or absolute abundance of bacteria and fungi;
[0015] The sampling frequencies of the time series data of the environmental ecological indicators and the colony abundance are the same.
[0016] Further, the corresponding time of the colony abundance prediction result is the same as the corresponding time of the environmental ecological indicators of the environmental area to be predicted, or is the future time of the corresponding time of the environmental ecological indicators of the environmental area to be predicted.
[0017] Further, the training method of the colony abundance prediction model includes:
[0018] Divide the environmental ecology-colony abundance data set into a training set and a validation set according to a preset ratio, use the supervised learning method to train the time series neural network according to the training set, optimize the network through backpropagation, and use the validation set to determine the prediction accuracy. When the loss function of the validation set is less than the loss function threshold and the prediction accuracy of the validation set is greater than the accuracy threshold, use the corresponding time series neural network as the corresponding colony abundance prediction model.
[0019] Further, the time series neural network is a long short-term memory network model, a gated recurrent unit model, or a time series model based on the attention mechanism; the loss function is the mean square error, mean absolute error, L1 norm loss function, or L2 norm loss function.
[0020] Further, the method further includes:
[0021] After constructing the environmental ecology - colony abundance dataset, standardize the environmental ecology indicators in the environmental ecology - colony abundance dataset.
[0022] Further, the method further includes:
[0023] Obtain new environmental ecology indicator data and colony abundance data according to a preset period, update and expand the environmental ecology - colony abundance dataset, and update the colony abundance multi - scale prediction model according to the preset period.
[0024] In a second aspect, the present invention provides a device for predicting colony abundance based on environmental ecology indicators, the device includes:
[0025] A construction module, configured to construct an environmental ecology - colony abundance dataset, where the environmental ecology - colony abundance dataset includes time - series data of environmental ecology indicators and colony abundance;
[0026] A training module, configured to train a time - series neural network according to the environmental ecology - colony abundance dataset to construct a colony abundance multi - scale prediction model, where the colony abundance multi - scale prediction model includes multiple colony abundance prediction models;
[0027] A prediction module, configured to obtain a multi - scale prediction result of the colony abundance in the to - be - predicted environmental area at the corresponding time according to the environmental ecology indicators of the to - be - predicted environmental area and based on the colony abundance multi - scale prediction model, where the multi - scale prediction result of the micro - ecological indicators is the weighted average of the colony abundance prediction results of multiple colony abundance prediction models.
[0028] In a third aspect, the present invention provides an electronic device, the electronic device includes a processor, a memory, and a communication bus;
[0029] The communication bus is used to realize the connection and communication between the processor and the memory;
[0030] The processor is configured to execute one or more programs in the memory to implement the steps of the method for predicting colony abundance based on environmental ecology indicators as described in the first aspect.
[0031] The beneficial effects of the present invention are as follows: The method, device, and electronic equipment for predicting colony abundance based on environmental ecological indicators provided by the present invention utilize the environmental ecological indicators of the environmental area to be predicted and, based on the multi-scale prediction model of colony abundance, can achieve the prediction of the colony abundance in the environmental area to be predicted. After constructing the multi-scale prediction model of colony abundance, it is no longer necessary to rely on regular sampling and laboratory testing of the colony abundance in the environmental area to be predicted, thereby improving the detection efficiency of colony abundance, meeting the requirements of daily detection of colony abundance, and not relying on high-precision instruments and professional testers, reducing the equipment maintenance cost and detection difficulty, and being able to promptly grasp the dynamic changes of colony abundance. And the present invention trains the prediction model based on a time series neural network, can achieve the advance prediction of colony abundance, and thus can identify potential threats in the environment in advance. In addition, the present invention introduces a multi-scale time series neural network model, comprehensively utilizes the time series data of environmental ecological indicators and colony abundance, and improves the accuracy of colony abundance prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1 It is a schematic flowchart of the method for predicting colony abundance based on environmental ecological indicators provided by the embodiment;
[0033] Figure 2 It is a schematic diagram of the principle of the multi-scale prediction model of colony abundance provided by the embodiment;
[0034] Figure 3 It is a schematic diagram of the loss functions of the training set and the validation set in the training process provided by the embodiment;
[0035] Figure 4 It is a schematic diagram of the prediction effect of the colony abundance prediction model provided by the embodiment;
[0036] Figure 5 It is a schematic diagram of the structure of the device for predicting colony abundance based on environmental ecological indicators provided by the embodiment;
[0037] Figure 6 It is a schematic diagram of the structure of the electronic equipment provided by the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0038] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments will be clearly and completely described below with reference to the accompanying drawings in the embodiments.
[0039] In some of the processes described in the specification of the present invention and the above-mentioned drawings, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations are only used to distinguish the different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel.
[0040] The technical solution of the present invention is applicable to the prediction of colony abundance in areas such as industrial parks or buildings, for example, agricultural parks, industrial parks, medical facility buildings, office facility buildings, etc.
[0041] In order to improve the efficiency of colony abundance detection and reduce the detection cost, the technical solution of the present invention is proposed. In the present invention, an environmental ecology-colony abundance data set is constructed, and the environmental ecology-colony abundance data set includes environmental ecological indicators and time series data of colony abundance; a time series neural network is trained according to the environmental ecology-colony abundance data set to construct a colony abundance multi-scale prediction model, and the colony abundance multi-scale prediction model includes a plurality of colony abundance prediction models; according to the environmental ecological indicators of the environment area to be predicted and based on the colony abundance multi-scale prediction model, a multi-scale prediction result of the colony abundance in the environment area to be predicted at the corresponding time is obtained, and the multi-scale prediction result of the microecological index is the weighted average of the colony abundance prediction results of a plurality of colony abundance prediction models.
[0042] It can be understood that there is a multi-scale interaction mechanism between environmental ecological indicators and colony abundance. Environmental factors jointly shape the structure and function of microbial communities through direct physiological stress and indirect ecological chain feedback. Among various environmental ecological indicators, temperature and humidity dominate the metabolic rate and environmental adaptability of microorganisms; air pressure regulates the activity of aerobic bacteria through oxygen supply; rainfall affects water distribution and the migration and diffusion of microorganisms; wind speed may have an indirect impact by changing microenvironmental conditions (such as water evaporation and ventilation). Based on this, the present application uses a time series neural network to capture the complex non-linear relationship between environmental ecological indicators and colony abundance, realizes the accurate mapping from environmental ecological indicator data to colony abundance data, completes the accurate prediction of colony abundance, and further does not rely on high-precision instruments and professional testers, reduces the equipment maintenance cost and detection difficulty, improves the detection efficiency, and can timely master the dynamic changes of colony abundance.
[0043] The following will clearly and completely describe the technical solutions in the present embodiment with reference to the drawings in the present embodiment. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.
[0044] Figure 1 A method for predicting colony abundance based on environmental ecological indicators is shown. Please refer toFigure 1 , the method includes:
[0045] Step 1: Construct an environmental ecology - colony abundance dataset, where the environmental ecology - colony abundance dataset includes time - series data of environmental ecology indicators and colony abundance.
[0046] In practical applications, environmental ecology indicator data and colony abundance data of the environmental area to be predicted within a historical time period can be collected at the same sampling frequency to obtain historical time - series data of environmental ecology indicators and colony abundance of the environmental area to be predicted, and then an environmental ecology - colony abundance dataset is constructed. In the environmental ecology - colony abundance dataset, the historical time - series data of environmental ecology indicators and the historical time - series data of colony abundance correspond one by one according to the sampling time.
[0047] In this embodiment, the environmental ecology indicators include multiple of daily maximum temperature, daily minimum temperature, daily average temperature, daily average humidity, daily cumulative rainfall, daily average air pressure, and daily average wind speed; the colony abundance includes at least the relative abundance or absolute abundance of bacteria and fungi. Among them, the types of bacteria include but are not limited to Sphingomonas, Deinococcus, Pantoea, Bacillus, and the types of fungi include but are not limited to Aspergillus, Nigrospora, Mrakia, Alternaria.
[0048] In this embodiment, sampling is carried out on a daily basis, and environmental ecology indicators and colony abundance in the past year are collected, including the daily average temperature (unit: °C), daily average humidity (unit: %), daily average wind speed (unit: m / s), daily average air pressure (unit: hPa), daily cumulative rainfall (unit: mm) of the environmental area to be predicted (such as a certain park), and the colony abundance of the environmental area to be predicted. Taking bacteria as an example in this embodiment, the colony abundances of 9 types of bacteria including Sphingomonas, Deinococcus, Pantoea, Bacillus, etc. are considered.
[0049] Taking May 15, 2023 as an example, the data is shown as: 24, 68.8, 0.1, 964.3, 0, 0.088, 0.195, 0.023, 0.063, 0.276, 0.044, 0.027, 0.020, 0.264;
[0050] Among them, the first 5 columns are daily average temperature, daily average humidity, daily average wind speed, daily average air pressure, and daily cumulative rainfall, and the last nine columns are the relative abundances of the colonies of 9 types of bacteria, that is, the sum of the last nine columns is 1.
[0051] In this embodiment, the method further includes: after constructing the environmental ecology - colony abundance dataset, performing standardization processing on the ecological indicators in the environmental ecology - colony abundance dataset. Standardization can be performed using the preprocessing.MinMaxScaler function in the scikit - learn library, which is min - max standardization. The standardization of data can be achieved through the fit_transform function command. By standardizing the data, the data quality and consistency can be improved, the model performance can be optimized, and the prediction accuracy can be further enhanced.
[0052] Step 2: Train a time - series neural network based on the environmental ecology - colony abundance dataset to construct a multi - scale prediction model for colony abundance. The multi - scale prediction model for colony abundance includes multiple colony abundance prediction models.
[0053] It can be understood that there is a multi - scale interaction relationship between environmental ecological indicators and colony abundance. In this embodiment, time - series neural networks are trained respectively using the environmental ecology - colony abundance dataset to obtain corresponding colony abundance prediction models. The time - series neural networks are used to capture the complex non - linear relationship between environmental ecological indicators and colony abundance, realizing an accurate mapping from environmental ecological indicator data to colony abundance data. At the same time, combined prediction is performed using multiple colony abundance prediction models to achieve multi - scale prediction of the colony abundance in the environmental area to be predicted.
[0054] Please refer to Figure 2 , in this embodiment, the multiple colony abundance prediction models include a first colony abundance prediction model and a second colony abundance prediction model.
[0055] Among them, the first colony abundance prediction model is used to predict the colony abundance prediction result at the corresponding time in the environmental area to be predicted according to the environmental ecological indicators in the environmental area to be predicted.
[0056] In practical applications, based on the environmental ecology - colony abundance dataset, a time - series neural network is trained using environmental ecological indicators and their corresponding colony abundances, and then the first colony abundance prediction model is obtained. By inputting the environmental ecological indicators of the environmental area to be predicted into the first colony abundance prediction model, the colony abundance prediction result at the corresponding time in the environmental area to be predicted can be obtained.
[0057] The second colony abundance prediction model is used to predict the colony abundance prediction result at the corresponding time in the environmental area to be predicted according to the environmental ecological indicators and the historical time - series data of colony abundance in the environmental area to be predicted.
[0058] In practical applications, based on the environmental ecology - colony abundance dataset, the environmental ecology indicators, the historical time - series data of colony abundance, and the colony abundance at the corresponding time in the corresponding area are used to train a time - series neural network, and then a second colony abundance prediction model is obtained. By inputting the environmental ecology indicators and the historical time - series data of colony abundance in the area to be predicted into the second colony abundance prediction model, the predicted result of the colony abundance at the corresponding time in the area to be predicted can be obtained.
[0059] In this embodiment, the corresponding time of the predicted result of the colony abundance is the same as the corresponding time of the environmental ecology indicators in the area to be predicted, or is the future time of the corresponding time of the environmental ecology indicators in the area to be predicted, which can be flexibly set according to requirements in practical applications. For example, the first colony abundance prediction model can predict the colony abundance at the current time based on the environmental ecology indicators at the current time, or can predict the colony abundance at a future time based on the environmental ecology indicators within a historical time period; the second colony abundance prediction model can predict the colony abundance at the current time based on the environmental ecology indicators at the current time and the historical time - series data of colony abundance, or can predict the colony abundance at a future time based on the environmental ecology indicators within a historical time period and the historical time - series data of colony abundance. Among them, the colony abundance at a future time can be the colony abundance at a certain future moment or within a future time period.
[0060] In this embodiment, the training method of the colony abundance prediction model includes: dividing the environmental ecology - colony abundance dataset into a training set and a validation set according to a preset ratio, training a time - series neural network according to the training set using a supervised learning method, optimizing the network through backpropagation, and determining the prediction accuracy using the validation set. When the loss function of the validation set is less than the loss function threshold and the prediction accuracy of the validation set is greater than the accuracy threshold, the corresponding time - series neural network is used as the corresponding colony abundance prediction model.
[0061] In this embodiment, the time - series neural network is a long short - term memory network model (LSTM), a gated recurrent unit model (GRU), or a time - series model based on an attention mechanism. The loss function of the time - series neural network is the mean squared error, the mean absolute error, the L1 - norm loss function, the L2 - norm loss function, or other error metric functions suitable for time - series prediction. The loss function is defined using the nn.MSELoss() function. The long short - term memory network model screens and retains key information through a gating mechanism, effectively solving the problem of gradient disappearance in traditional RNNs, thereby capturing long - term dependencies in time series. And the time - series neural network can learn complex non - linear patterns, without the need for manual feature design. The model can directly learn implicit features from the original time - series data, simplifying the modeling process.
[0062] The preset ratio for dividing the training set and the validation set can be set as needed, such as 7:3, 4:1, 8:2, or 9:1. In this embodiment, 80% of the data is taken as the training set, and 20% of the data is taken as the validation set. During the training process, the change of the loss function of the training set and the validation set is shown as Figure 3 shown; after the training is completed, the comparison between the predicted results and the true results of the colony relative abundances of 5 groups of samples in the test set is as Figure 4 shown, where different depths of color represent the relative abundances of different colonies.
[0063] Step 3: According to the environmental ecological indicators of the environmental area to be predicted and based on the multi-scale prediction model of colony abundance, obtain the multi-scale prediction results of the colony abundance in the environmental area to be predicted at the corresponding time. The multi-scale prediction results of the microecological indicators are the weighted average of the colony abundance prediction results of multiple colony abundance prediction models.
[0064] After the construction of the multi-scale prediction model of colony abundance is completed, the prediction of the colony abundance in the environmental area to be predicted at the corresponding time can be realized according to the multi-scale prediction model of colony abundance. Furthermore, it does not need to rely on high-precision instruments and professional testers, reduces the equipment maintenance cost and the detection difficulty, improves the detection efficiency, and can timely grasp the dynamic changes of the colony abundance.
[0065] The following takes a certain park in a certain city as the environmental area to be predicted, and uses the environmental ecological indicators of the park in the past 10 days and the colony abundances in the past 10 days to predict the relative abundances of 9 kinds of bacteria in the park in the next 1 day as an example to illustrate the specific process of colony abundance prediction:
[0066] First, input the environmental ecological indicators of the park in the past 10 days into the first colony abundance prediction model to obtain the first prediction results of the relative abundances of 9 kinds of bacteria in the next 1 day. Then, input the environmental ecological indicators of the park in the past 10 days and the relative abundances of 9 kinds of bacteria in the park in the past 10 days into the second colony abundance prediction model to obtain the second prediction results of the relative abundances of 9 kinds of bacteria in the park in the next 1 day. Finally, perform a weighted average on the first prediction results and the second prediction results to obtain the multi-scale prediction results of the relative abundances of 9 kinds of bacteria in the park in the next 1 day.
[0067] By performing multi-scale prediction on the colony abundance, the complex relationship between the environmental ecological indicators and the colony abundance can be fully captured, reducing the misjudgment of the complex ecological relationship by a single model and improving the accuracy of the colony abundance prediction.
[0068] In this embodiment, since the relative abundance is used to represent the colony abundance, the sum of the colony abundance prediction results is 1. By adding a Softmax layer in the neural network model, it can be ensured that the sum of the output relative abundances is 1.
[0069] In this embodiment, the method further includes: obtaining new environmental ecological index data and colony abundance data according to a preset period, updating and expanding the environmental ecology - colony abundance data set, and updating the multi - scale prediction model of colony abundance according to the preset period.
[0070] Regularly updating the data set can correct incorrect data, supplement missing information, and avoid prediction biases caused by data lags. Updating the multi - scale prediction model of colony abundance can adapt to changes in data relationships, prevent the model from becoming invalid due to outdated training data, and further improve the accuracy of prediction.
[0071] In summary, the method for predicting colony abundance based on environmental ecological indicators provided in this embodiment does not need to rely on regular sampling and laboratory detection of the colony abundance in the environment area to be predicted, thereby improving the detection efficiency of colony abundance, meeting the needs of daily detection of colony abundance, and not relying on high - precision instruments and professional testers, reducing the equipment maintenance cost and detection difficulty, and being able to timely grasp the dynamic changes of colony abundance; and this embodiment trains a prediction model based on a time - series neural network, which can realize the advance prediction of colony abundance, and then can identify potential threats in the environment in advance; in addition, this embodiment introduces a multi - scale time - series neural network model, comprehensively utilizes the time - series data of environmental ecological indicators and colony abundance, and improves the accuracy of colony abundance prediction.
[0072] Based on the above technical solution, this embodiment also proposes a device for predicting colony abundance based on environmental ecological indicators. Please refer to Figure 5 , the device includes:
[0073] A construction module for constructing an environmental ecology - colony abundance data set, where the environmental ecology - colony abundance data set includes time - series data of environmental ecological indicators and colony abundance;
[0074] A training module for training a time - series neural network according to the environmental ecology - colony abundance data set to construct a multi - scale prediction model of colony abundance, where the multi - scale prediction model of colony abundance includes multiple colony abundance prediction models;
[0075] A prediction module for obtaining a multi - scale prediction result of colony abundance in the environment area to be predicted at the corresponding time based on the environmental ecological indicators of the environment area to be predicted and the multi - scale prediction model of colony abundance, where the multi - scale prediction result of the micro - ecological index is the weighted average of the colony abundance prediction results of multiple colony abundance prediction models.
[0076] Based on the above technical solution, this embodiment also proposes an electronic device. Please refer to Figure 6 , the electronic device includes a processor, a memory, and a communication bus;
[0077] The communication bus is used to implement the connection and communication between the processor and the memory;
[0078] The processor is used to execute one or more programs in the memory to implement the steps of the method for predicting colony abundance based on environmental ecological indicators as described in this embodiment.
[0079] It can be understood that since the device and the electronic device for predicting colony abundance based on environmental ecological indicators described in this embodiment are used to implement the method for predicting colony abundance based on environmental ecological indicators described in the embodiment, for the device and the electronic device disclosed in the embodiment, since they correspond to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, refer to the partial description of the method and no further details will be provided here.
Claims
1. A method for predicting colony abundance based on environmental ecological indicators, characterized in that, The method includes: Constructing an environmental ecology - colony abundance dataset, where the environmental ecology - colony abundance dataset includes time - series data of environmental ecology indicators and colony abundance; Training a time - series neural network according to the environmental ecology - colony abundance dataset to construct a multi - scale prediction model for colony abundance, where the multi - scale prediction model for colony abundance includes multiple colony abundance prediction models; Based on the environmental ecology indicators of the environmental area to be predicted and using the multi - scale prediction model for colony abundance, obtaining the multi - scale prediction results of the colony abundance in the corresponding time of the environmental area to be predicted, and the multi - scale prediction results of the micro - ecology indicators are the weighted average of the colony abundance prediction results of multiple colony abundance prediction models.
2. The method for predicting colony abundance based on environmental ecological indicators according to claim 1, wherein The colony abundance prediction models include a first colony abundance prediction model and a second colony abundance prediction model; The first colony abundance prediction model is used to predict the colony abundance prediction result in the corresponding time of the environmental area to be predicted according to the environmental ecology indicators of the environmental area to be predicted; The second colony abundance prediction model is used to predict the colony abundance prediction result in the corresponding time of the environmental area to be predicted according to the environmental ecology indicators and the historical time - series data of colony abundance of the environmental area to be predicted.
3. The method for predicting colony abundance based on environmental ecological indicators according to claim 2, wherein The environmental ecology indicators include multiple ones among the daily maximum temperature, daily minimum temperature, daily average temperature, daily average humidity, daily cumulative rainfall, daily average air pressure, and daily average wind speed; The colony abundance includes at least the relative abundance or absolute abundance of bacteria and fungi; The sampling frequencies of the time - series data of the environmental ecology indicators and colony abundance are the same.
4. The method for predicting colony abundance based on environmental ecological indicators according to claim 2, wherein The corresponding time of the colony abundance prediction result is the same as the corresponding time of the environmental ecology indicators of the environmental area to be predicted, or is the future time of the corresponding time of the environmental ecology indicators of the environmental area to be predicted.
5. The method for predicting colony abundance based on environmental ecological indicators according to claim 1, wherein The training method of the colony abundance prediction model includes: Dividing the environmental ecology - colony abundance dataset into a training set and a validation set according to a preset ratio, training a time - series neural network according to the training set using a supervised learning method, optimizing the network through backpropagation, and using the validation set to determine the prediction accuracy. When the loss function of the validation set is less than the loss function threshold and the prediction accuracy of the validation set is greater than the accuracy threshold, the corresponding time - series neural network is used as the corresponding colony abundance prediction model.
6. The method for predicting colony abundance based on environmental ecological indicators according to claim 5, wherein The time - series neural network is a long short - term memory network model, a gated recurrent unit model, or a time - series model based on an attention mechanism; the loss function is a mean square error, mean absolute error, L1 - norm loss function, or L2 - norm loss function.
7. The method for predicting colony abundance based on environmental ecological indicators according to claim 1, characterized in that The method further includes: After constructing the environmental ecology - colony abundance dataset, performing standardization processing on the environmental ecology indicators in the environmental ecology - colony abundance dataset.
8. The method for predicting colony abundance based on environmental ecological indicators according to claim 1, wherein The method further includes: Obtaining new environmental ecology indicator data and colony abundance data according to a preset period, updating and expanding the environmental ecology - colony abundance dataset, and updating the multi - scale prediction model for colony abundance according to a preset period.
9. A device for predicting colony abundance based on environmental ecological indicators, characterized in that, The device includes: A construction module for constructing an environmental ecology - colony abundance dataset, where the environmental ecology - colony abundance dataset includes time - series data of environmental ecology indicators and colony abundance; A training module, configured to train a time series neural network according to the environmental ecology - colony abundance dataset, and construct a colony abundance multi - scale prediction model, where the colony abundance multi - scale prediction model includes multiple colony abundance prediction models; A prediction module, configured to obtain a multi - scale prediction result of the colony abundance in the to - be - predicted environmental area at the corresponding time according to the environmental ecology indicators of the to - be - predicted environmental area and based on the colony abundance multi - scale prediction model, and the multi - scale prediction result of the micro - ecological indicators is the weighted average of the colony abundance prediction results of multiple colony abundance prediction models.
10. An electronic device, characterized in that, The electronic device includes a processor, a memory, and a communication bus; The communication bus is used to implement connection communication between the processor and the memory; The processor is configured to execute one or more programs in the memory to implement the steps of the colony abundance prediction method based on environmental ecology indicators according to any one of claims 1 to 8.