Nano-silver algae growth inhibition level prediction method based on random forest

Abstract

The invention discloses a nano-silver algae growth inhibition level prediction method based on a random forest, and the method comprises the following steps: collecting the physical property, chemical property, experimental method and environmental condition data of nano-silver, and obtaining a data set comprising a descriptor of the nano-silver and algae growth inhibition data; aiming at missing values existing in the data set, preprocessing the data, and dividing the data set into a training set and a verification set; constructing a nano-silver algae growth inhibition random forest prediction model, and training the random forest prediction model by using the training set data until the random forest prediction model is stably fitted; and inputting verification set data into the trained random forest prediction model, and outputting a nano-silver algae growth inhibition level prediction value by the random forest prediction model for evaluating the nano-silver ecological risk. The algae growth inhibition level of the nano-silver is rapidly predicted, the method is simple and rapid, the cost is low, and manpower, material resources and financial resources required by experimental determination are saved.

Description

Technical Field

[0001] The present invention relates to the technical field of high-throughput ecological risk assessment and prediction of nanomaterials, and in particular to a method for predicting the growth inhibition level of nanosilver algae based on random forests for ecological risk assessment. Background Art

[0002] The aquatic toxicity assessment of artificial nanomaterials mainly follows the experimental methods of chemical toxicity assessment. For example, the standard ecotoxicity test methods of the Organization for Economic Co-operation and Development (OECD), including algal growth toxicity (OECD 201), Daphnia magna acute toxicity (OECD 202) and fish acute toxicity (OECD 203). Nanomaterials are defined as materials with one or more external dimensions between 1nm and 100nm. Due to the nano-size effect, nanomaterials have many unique physical and chemical properties, such as large specific surface area and high surface energy.

[0003] These properties not only bring many excellent performances to nanomaterials, but are also important factors affecting the toxicity of nanomaterials. Nanomaterials are not exactly the same as traditional chemicals. For example, silver (Ag) nanomaterials cannot be regarded as silver (Ag) for toxicity assessment, because changes in physical and chemical properties such as size, shape, surface modification or coating can affect the toxicological effects of silver (Ag). It is difficult to complete the toxicity assessment of all forms of artificial nanomaterials using only experimental methods. A faster and more effective method is needed to conduct toxicity assessment of artificial nanomaterials.

[0004] Computational toxicology methods, such as the quantitative structure-activity relationship (QSAR) model, are widely used in the toxicity assessment of chemicals. The quantitative structure-activity relationship model constructed for nanomaterials is called the quantitative nanostructure-activity relationship (Nano-QSAR) model. Similar to the quantitative structure-activity relationship, the quantitative nanostructure-activity relationship also follows the assumption that similar nanomaterials may produce similar biological effects. The construction of the quantitative nanostructure-activity relationship model relies on a set of descriptors that encode the composition, topology or geometric characteristics of the nanomaterials. These descriptors can be obtained either through customized software or through experimental testing, such as elemental composition, Zeta potential, size distribution, shape, etc.

[0005] Machine learning methods are an indispensable analytical tool in computational toxicology. Through machine learning algorithms, not only can the toxicity of nanomaterials be predicted, but also the toxic mechanism of nanomaterials can be explained, and the causal relationship between identification properties and toxic effects can be identified. The current machine learning model of nanosilver toxicity considers the impact of environmental factors on toxicity. However, various factors in the water environment can change the physicochemical properties of the original form of nanosilver, affect its environmental transformation process, and thus change the toxic effects of nanosilver on algae. Therefore, it is necessary to establish a predictive model for nanosilver algae toxicity based on environmental factors. Summary of the invention

[0006] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and provide a method for predicting the algae growth inhibition level of nanosilver based on random forests to predict the half effective concentration (EC50) of the algae growth inhibition level of nanosilver. 50 ) data can achieve more efficient and rapid ecological risk assessment for nanosilver.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0008] A method for predicting the growth inhibition level of nanosilver algae based on random forests comprises the following steps:

[0009] S1. Collect data on the physical properties, chemical properties, experimental methods, and environmental conditions of nanosilver to obtain a data set including descriptors of nanosilver and algae growth inhibition data;

[0010] S2. Preprocess the data for missing values ​​in the data set and divide the data set into a training set and a validation set;

[0011] S3. Construct a random forest prediction model for the inhibition of algae growth by nanosilver, and train the random forest prediction model using the training set data until the random forest prediction model is stable in fitting;

[0012] S4. Input the validation set data into the trained random forest prediction model, and the random forest prediction model outputs the predicted value of the nanosilver algae growth inhibition level for evaluating the ecological risk of nanosilver.

[0013] Further, the descriptors of nanosilver include: source, shape, surface coating, diameter, size test method, hydrated size, solubility, zeta potential, polymer dispersibility index, pH, temperature, light, exposure duration, exposure experiment, and algae growth status.

[0014] Furthermore, the data is preprocessed for missing values ​​in the data set, including:

[0015] For numerical data, the mean value is used to replace the missing values ​​of the corresponding descriptors;

[0016] For textual description data, one-hot encoding is used to convert the textual description data into binary numerical data, that is, convert it into 0 and 1. If there are missing values, the number 2 is used to fill them.

[0017] Furthermore, diameter, hydrated size, solubility, zeta potential, polymer dispersibility index, pH, temperature, light, and exposure time are numerical data.

[0018] Furthermore, the source, shape, surface coating, size test method, exposure experiment, and algae growth status are textual description data.

[0019] Furthermore, the random forest algorithm is used to build a random forest prediction model, including:

[0020] Generate multiple sub-sample sets, i.e., random subsets, by sampling with replacement from the training set data;

[0021] A decision tree is constructed for each sub-sample set. The construction process of each decision tree is:

[0022] First, feature selection is performed, that is, when each node is split, a part of the features are randomly selected for splitting, and then the node is split into sub-nodes according to the optimal split point among the selected features. The above process is repeated until the predetermined stopping condition is reached;

[0023] Finally, based on the average method, the results of each decision tree are averaged to obtain the final prediction value.

[0024] Furthermore, the random forest prediction model was modeled using Python's Scikit-learn package. The data set included 52 data records, and the data set was divided into a training set and a validation set in a ratio of 4:1.

[0025] Furthermore, the random forest algorithm uses the RandomForestRegressor function in the Scikit-learn package. The specific parameters are: the number of decision trees is set to 500, the random number seed is set to 2023, the partitioning criterion uses the mean square error, and the other parameters are the function default values.

[0026] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned methods for predicting the growth inhibition level of nanosilver algae based on random forest are implemented.

[0027] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for predicting the growth inhibition level of nanosilver algae based on random forest as described in any one of the above.

[0028] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0029] 1. The method for predicting the algae growth inhibition level of nanosilver involved in the present invention can quickly predict the algae growth inhibition level of nanosilver through the physical properties, chemical properties, quasi-experimental methods and environmental condition data of nanosilver. The method is not only simple, fast and low-cost, but also saves the manpower, material resources and financial resources required for experimental determination.

[0030] 2. The establishment and verification of the prediction method of the nanosilver algae growth inhibition level of the present invention are strictly in accordance with the OECD guidelines for the establishment and use of QSAR models. Therefore, the prediction results of the nanosilver algae growth inhibition level of the present invention can provide data support for chemical supervision and are of great significance to the ecological risk assessment of chemicals. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Flow chart of the random forest-based nanosilver algae growth inhibition level prediction method.

[0032] Figure 2 Schematic diagram of the process of building a random forest prediction model.

[0033] Figure 3 Schematic diagram of the performance characterization of the random forest prediction model for nanosilver algae growth inhibition. DETAILED DESCRIPTION

[0034] The following is a further description of the method for predicting the growth inhibition level of algae using nanosilver based on random forests in conjunction with the accompanying drawings and specific embodiments.

[0035] See also Figure 1 The present invention discloses a method for predicting the growth inhibition level of nanosilver algae based on random forest, comprising the following steps:

[0036] S1. Collect data on the physical properties, chemical properties, experimental methods, and environmental conditions of nanosilver to obtain a data set including descriptors of nanosilver and algae growth inhibition data;

[0037] S2. Preprocess the data for missing values ​​in the data set and divide the data set into a training set and a validation set;

[0038] S3. Construct a random forest prediction model for the inhibition of algae growth by nanosilver, and train the random forest prediction model using the training set data until the random forest prediction model is stable in fitting;

[0039] S4. Input the validation set data into the trained random forest prediction model, and the random forest prediction model outputs the predicted value of the nanosilver algae growth inhibition level for evaluating the ecological risk of nanosilver.

[0040] The present invention collects the half effective concentration (EC) of the nanosilver to inhibit the growth of algae 50 ) data, the algae growth inhibition level is an important indicator for assessing the ecological risk of chemicals. By establishing a predictive model for the algae growth inhibition level of nanosilver, a more efficient and rapid ecological risk assessment of nanosilver can be achieved.

[0041] Specifically, the descriptors of nanosilver include: source, shape, surface coating, diameter, size test method, hydrated size, solubility, Zeta potential, polymer dispersity index (PDI), pH (Pondus Hydrogenii), temperature, light, exposure time, exposure experiment, algae growth state. Zeta potential refers to the potential of the shear plane, also known as electromotive potential or electromotive potential (ζ-potential or ζ-potential), which is an important indicator to characterize the stability of colloidal dispersions.

[0042] Specifically, for the missing values ​​in the data set, the data is preprocessed, including:

[0043] For numerical data, the average value is used to replace the missing value of the corresponding descriptor; for text description data, one-hot encoding is used to convert the text description data into binary numerical data, that is, convert it into 0 and 1. If there are missing values, the number 2 is used to fill them.

[0044] Among them, diameter, hydrated size, solubility, Zeta potential, polymer dispersibility index, pH, temperature, light, and exposure time are numerical data. Source, shape, surface coating, size test method, exposure experiment, and algae growth status are textual description data.

[0045] Specifically, see Figure 2 , using the random forest algorithm, build a random forest prediction model, including:

[0046] Generate multiple sub-sample sets, i.e., random subsets, by sampling with replacement from the training set data;

[0047] A decision tree is constructed for each sub-sample set. The construction process of each decision tree is as follows: first, feature selection is performed, that is, when each node is split, a part of the features are randomly selected for splitting, and then the node is split into sub-nodes according to the optimal split point among the selected features. The above process is repeated until the predetermined stopping condition is reached;

[0048] Finally, based on the average method, the results of each decision tree are averaged to obtain the final prediction value.

[0049] Specifically, the random forest prediction model was built using Python's Scikit-learn package. The data set included 52 data records, and the data set was divided into a training set and a validation set at a ratio of 4:1. The random forest algorithm used the RandomForestRegressor function in the Scikit-learn package. The specific parameters were: the number of decision trees was set to 500, the random number seed was set to 2023, the partition standard was the mean square error, and the remaining parameters were the function default values.

[0050] See also Figure 3 , the goodness of fit R of the training set 2 tr The goodness of fit R of the validation set is 0.863. 2 test It is 0.805, indicating that the random forest prediction model has good goodness of fit and external prediction ability, and there is no overfitting phenomenon. It can be used to predict the algae growth inhibition level of nanosilver.

[0051] Example

[0052] Given a nanosilver material, assuming that it is from a commercial source, has a coated surface, is spherical in shape, has a diameter of 15 nanometers, and has a hydrated diameter test method of dynamic light scattering (DLS), a hydrated diameter of 2.36 nanometers, a solubility of 5%, a Zeta potential of -26 mV, a polymer dispersibility index of 0.34, a pH of 7.32, an exposure temperature of 24°C, an illumination time of 24 hours, and an exposure time of 72 hours, the growth inhibition level of silver algae (EC 50 ) has a base 10 logarithm of 2.87.

[0053] The specific calculation process is as follows Figure 2As shown in the figure, when the data is input into the random forest prediction model, the random forest prediction model first randomly generates 500 decision trees, each of which will give a prediction result, and the final result is the average of these 500 decision trees. For example, assuming there are only 3 decision trees, and the results given by decision tree 1, decision tree 2, and decision tree 3 are 3, 4, and 5 respectively, then the final prediction result given by the random forest prediction model is the average of these three decision trees, which is 4.

[0054] In summary, the present invention has the following advantages and beneficial effects:

[0055] 1. The method for predicting the algae growth inhibition level of nanosilver involved in the present invention can quickly predict the algae growth inhibition level of nanosilver through the physical properties, chemical properties, quasi-experimental methods and environmental condition data of nanosilver. The method is not only simple, fast and low-cost, but also saves the manpower, material resources and financial resources required for experimental determination.

[0056] 2. The establishment and verification of the prediction method of the nanosilver algae growth inhibition level of the present invention are strictly in accordance with the OECD guidelines for the establishment and use of QSAR models. Therefore, the prediction results of the nanosilver algae growth inhibition level of the present invention can provide data support for chemical supervision and are of great significance to the ecological risk assessment of chemicals.

[0057] The present invention also discloses an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for predicting the growth inhibition level of nanosilver algae based on random forest as described above are implemented. The electronic device of the present invention can execute the method for predicting the growth inhibition level of nanosilver algae based on random forest of the present invention, and can execute any combination of implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.

[0058] The present invention also discloses a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the method for predicting the growth inhibition level of nanosilver algae based on random forest as described above are implemented. The computer-readable storage medium of the present invention can execute the method for predicting the growth inhibition level of nanosilver algae based on random forest of the present invention, and can execute any combination of implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.

[0059] Although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions, and internal relationships of the various functional modules in the system disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0060] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, 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. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0061] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0062] The various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0063] The above description is a detailed description of the preferred feasible embodiments of the present invention, but the embodiments are not intended to limit the scope of the patent application of the present invention. All equivalent changes or modified changes completed under the technical spirit disclosed by the present invention should fall within the patent scope covered by the present invention.

Claims

1. A method for predicting the growth inhibition level of nanosilver algae based on random forest, characterized in that: The following steps are involved: S1. Collect data on the physical properties, chemical properties, experimental methods, and environmental conditions of nanosilver to obtain a data set including descriptors of nanosilver and algae growth inhibition data; S2. Preprocess the data for missing values ​​in the data set and divide the data set into a training set and a validation set; S3. Construct a random forest prediction model for the inhibition of algae growth by nanosilver, and train the random forest prediction model using the training set data until the random forest prediction model is stable in fitting; S4. Input the validation set data into the trained random forest prediction model, and the random forest prediction model outputs the predicted value of the nanosilver algae growth inhibition level for evaluating the ecological risk of nanosilver.

2. The method for predicting the growth inhibition level of algae using nanosilver based on random forest according to claim 1, characterized in that: Descriptors of nanosilver include: source, shape, surface coating, diameter, size test method, hydrated size, solubility, zeta potential, polymer dispersibility index, pH, temperature, light, exposure duration, exposure experiment, and algal growth status.

3. The method for predicting the growth inhibition level of algae using nanosilver based on random forest according to claim 2, characterized in that: According to the missing values ​​in the data set, the data is preprocessed, including: For numerical data, the mean value is used to replace the missing values ​​of the corresponding descriptors; For textual description data, one-hot encoding is used to convert the textual description data into binary numerical data, that is, convert it into 0 and 1. If there are missing values, the number 2 is used to fill them.

4. The method for predicting the growth inhibition level of nanosilver algae based on random forest according to claim 3, characterized in that: Diameter, hydrated size, solubility, zeta potential, polymer dispersibility index, pH, temperature, light, and exposure time are numerical data.

5. The method for predicting the growth inhibition level of algae using nanosilver based on random forest according to claim 3, characterized in that: The source, shape, surface coating, size test method, exposure experiment, and algae growth status are textual description data.

6. The method for predicting the growth inhibition level of nanosilver algae based on random forest according to claim 1, characterized in that: Use the random forest algorithm to build a random forest prediction model, including: Generate multiple sub-sample sets, i.e., random subsets, by sampling with replacement from the training set data; A decision tree is constructed for each sub-sample set. The construction process of each decision tree is: First, feature selection is performed, that is, when each node is split, a part of the features are randomly selected for splitting, and then the node is split into sub-nodes according to the optimal split point among the selected features. The above process is repeated until the predetermined stopping condition is reached; Finally, based on the average method, the results of each decision tree are averaged to obtain the final prediction value.

7. The method for predicting the growth inhibition level of nanosilver algae based on random forest according to claim 6, characterized in that: The random forest prediction model was built using Python's Scikit-learn package. The data set included 52 data records, and the data set was divided into a training set and a validation set in a ratio of 4:

1.

8. The method for predicting the growth inhibition level of nanosilver algae based on random forest according to claim 7, characterized in that: The random forest algorithm uses the RandomForestRegressor function in the Scikit-learn package. The specific parameters are: the number of decision trees is set to 500, the random number seed is set to 2023, the partitioning standard uses the mean square error, and the other parameters are the function default values.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for predicting the growth inhibition level of nanosilver algae based on random forests as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for predicting the growth inhibition level of nanosilver algae based on random forests as described in any one of claims 1 to 8 are implemented.

Images