Digital transformation application recommendation method and system, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a digital transformation application recommendation method and system, electronic equipment and a storage medium, and belongs to the field of digital transformation intelligent recommendation of small and medium-sized enterprises. The method comprises the following steps: obtaining enterprise self-diagnosis data, and processing the enterprise self-diagnosis data to obtain target data; the target data is input into a pre-constructed prediction model for prediction, enterprise real application classification is obtained, and the prediction model is constructed based on a linear regression function; and generating a recommendation scheme according to the actual application classification of the enterprise in combination with the actual situation of the enterprise. By constructing an intelligent application recommendation prediction model, specific diagnostic data and business requirements of an enterprise can be analyzed, and a most suitable application solution is automatically matched and recommended for the enterprise, so that decision errors and resource waste caused by information asymmetry and resource limitation are reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent recommendation technology for digital transformation of small and medium-sized enterprises, and specifically to a digital transformation application recommendation method, system, electronic device and storage medium. Background Art

[0002] In today's world of globalization and rapid development of information technology, digital transformation has become the key to the continued competition and development of enterprises. Especially for small and medium-sized enterprises, traditional business models and operating methods can no longer meet the development needs of enterprises. Effective digital strategies can not only improve their operational efficiency, but also open up new business opportunities and markets. However, faced with a wide range of complex digital technology options and solutions, small and medium-sized enterprises generally face multiple challenges. (1) Information asymmetry: Due to the limitations of scale and resources, small and medium-sized enterprises usually lack channels to obtain the latest digital technology information. This information asymmetry puts enterprises at a disadvantage when choosing digital solutions suitable for their own development, and they often cannot make the most favorable decisions. (2) Resource limitations: Compared with large enterprises, small and medium-sized enterprises have more limited resources in terms of funds, technology and talents. These limitations affect their investment and attempts in digital transformation, making it difficult for them to bear the cost of trial and error. (3) Technical complexity: Contemporary digital technologies cover a wide range, including but not limited to cloud computing, big data, artificial intelligence, etc. The understanding and application of these technologies require deep professional knowledge, and small and medium-sized enterprises often lack such professional talents. (4) Personalized needs: Each small and medium-sized enterprise has its own unique business needs and market positioning, and standardized digital solutions on the market often cannot fully meet their personalized needs. This incomplete match may lead to waste of corporate investment and unsatisfactory transformation results. (5) Market adaptability: Digital transformation is not only the application of technology, but also a market strategy. In a rapidly changing market environment, small and medium-sized enterprises need to quickly adapt and use digital technology to improve their market response speed and customer service quality.

[0003] Traditional transformation consulting services often rely on manual experience, which is expensive and inefficient. In addition, information asymmetry and resource constraints lead to wrong decisions and waste of resources. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a digital transformation application recommendation method, system, electronic device and storage medium, which are used to fully or at least partially solve the technical problems existing in the above-mentioned prior art.

[0005] In order to achieve the above object, an embodiment of the present invention provides a digital transformation application recommendation method, including:

[0006] Acquire enterprise self-diagnosis data, and process the enterprise self-diagnosis data to obtain target data;

[0007] Input the target data into a pre-built prediction model for prediction to obtain the enterprise's real application classification, wherein the prediction model is built based on a linear regression function;

[0008] Generate recommended solutions based on the enterprise's real application classification and actual situation.

[0009] Optionally, obtaining enterprise self-diagnosis data and processing the enterprise self-diagnosis data to obtain target data includes:

[0010] Use Talend tools to connect to data sources, import enterprise self-diagnosis data, and perform data consistency checks to repair or delete inconsistent data records;

[0011] Use Talend tools to remove duplicates, process missing values, and process outliers in the enterprise self-diagnosis data to obtain target data.

[0012] Optionally, the prediction model building process includes:

[0013] Use the torch.utils.data.Dataset class or its subclasses to load the enterprise self-diagnosis data and obtain individual samples and the total number of samples;

[0014] For the image data in the enterprise self-diagnosis data, use the predefined datasets in torchvision.datasets and perform data preprocessing through torchvision.transforms;

[0015] Use torch.utils.data.DataLoader to encapsulate the dataset into an iterable data loader to support batch loading, shuffling data, and multi-process loading.

[0016] Define the parameters of the prediction model and set the loss function and optimizer;

[0017] Traverse multiple epochs, and in each epoch, iteratively load data through DataLoader;

[0018] Perform forward propagation on each batch of data to calculate the output and loss, perform backpropagation to calculate the gradient of the loss with respect to the model parameters, use the optimizer to update the parameters of the prediction model, and obtain the prediction model.

[0019] Optionally, the optimization evaluation process of the prediction model includes:

[0020] Obtain the enterprise self-diagnosis dataset and divide it into feature X and label y;

[0021] Input feature X into the prediction model to obtain the prediction label;

[0022] Use scikit-learn's cross_val_score function or cross_validate function for cross validation; the cross_val_score function returns an array containing the score for each iteration, and the cross_validate function returns the mean, standard deviation, and training time for each score metric;

[0023] The parameters of the prediction model are adjusted according to the results of cross-validation to optimize the performance of the prediction model.

[0024] Optionally, the recommendation logic used to generate the recommendation solution is as follows:

[0025] Sort out all application categories and match them with the corresponding categories of the purchased applications, and sort out the application categories corresponding to each item in the evaluation form;

[0026] According to the evaluation question table, the enterprise evaluation scores are obtained, and the enterprise evaluation scores are matched with the applications provided by the supplier. Priorities are set according to the score weights, and matching applications are arranged according to the priorities.

[0027] On the other hand, the present invention also provides a digital application recommendation system, comprising:

[0028] An acquisition unit, used for acquiring enterprise self-diagnosis data, and processing the enterprise self-diagnosis data to obtain target data;

[0029] A prediction unit, used for inputting the target data into a pre-built prediction model for prediction to obtain the enterprise's real application classification, wherein the prediction model is built based on a linear regression function;

[0030] The recommendation unit is used to generate recommendation solutions based on the enterprise's real application classification and the enterprise's actual situation.

[0031] Optionally, the acquiring unit is specifically used for:

[0032] Use Talend tools to connect to data sources, import enterprise self-diagnosis data, and perform data consistency checks to repair or delete inconsistent data records;

[0033] Use Talend tools to remove duplicates, process missing values, and process outliers in the enterprise self-diagnosis data to obtain target data.

[0034] Optionally, the prediction unit is specifically used for:

[0035] Use the torch.utils.data.Dataset class or its subclasses to load the enterprise self-diagnosis data and obtain individual samples and the total number of samples;

[0036] For the image data in the enterprise self-diagnosis data, use the predefined datasets in torchvision.datasets and perform data preprocessing through torchvision.transforms;

[0037] Use torch.utils.data.DataLoader to encapsulate the dataset into an iterable data loader to support batch loading, shuffling data, and multi-process loading.

[0038] Define the parameters of the prediction model and set the loss function and optimizer;

[0039] Traverse multiple epochs, and in each epoch, iteratively load data through DataLoader;

[0040] Perform forward propagation on each batch of data to calculate the output and loss, perform backpropagation to calculate the gradient of the loss with respect to the model parameters, use the optimizer to update the parameters of the prediction model, and obtain the prediction model.

[0041] On the other hand, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned digital transformation application recommendation method when executing the program.

[0042] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned digital transformation application recommendation method.

[0043] Through the above technical solution, by building an intelligent application recommendation prediction model, it is possible to analyze the company's specific diagnostic data and business needs, automatically match and recommend the most suitable application solutions for the company, thereby reducing decision-making errors and resource waste caused by information asymmetry and resource constraints.

[0044] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0046] Figure 1 It is an implementation flow chart of a digital transformation application recommendation method provided by an embodiment of the present invention;

[0047] Figure 2 It is a structural diagram of a digital application recommendation system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0049] See also Figure 1 As shown, it is a flowchart of an implementation method of a digital transformation application recommendation method provided by an embodiment of the present invention, including the following execution steps:

[0050] Step 100: Acquire enterprise self-diagnosis data, and process the enterprise self-diagnosis data to obtain target data.

[0051] Specifically, when executing step 100, the following steps may be specifically performed:

[0052] S1000: Use Talend tools to connect to data sources, import enterprise self-diagnosis data, and perform data consistency checks to repair or delete inconsistent data records.

[0053] S1001: Use Talend tools to remove duplicates, process missing values, and process outliers in the enterprise self-diagnosis data to obtain target data.

[0054] In some implementations, the code is as follows:

[0055] #Assume that the data has been imported into the Pandas DataFrame via Talend

[0056] import pandas as pd

[0057] #Sample data cleaning

[0058] df=pd.read_csv('diagnosis_data.csv')

[0059] df.dropna(inplace=True)#delete missing values

[0060] df.drop_duplicates(inplace=True)#delete duplicates

[0061] #Abnormal value processing (this is just an example, the specific situation depends on the actual situation)

[0062] df['x1'] = df['x1'].apply(lambda x:x if x<100else np.nan) #Assume that the reasonable range of x1 is less than 100

[0063] df.dropna(inplace=True)#Delete the processed missing values ​​again.

[0064] In some embodiments, selecting features useful for predicting a target may include data standardization or normalization.

[0065] Code Explanation:

[0066] from sklearn.preprocessing import StandardScaler

[0067] scaler = StandardScaler()

[0068] df_scaled=df[['x1','x2']].apply(scaler.fit_transform)

[0069] df_scaled=pd.DataFrame(df_scaled, columns=['x1','x2']).

[0070] Step 101: Input the target data into a pre-built prediction model for prediction to obtain the enterprise's real application classification, wherein the prediction model is built based on a linear regression function.

[0071] In some embodiments, the linear regression function is y=β0+β1x+∈, where β0 and β1 represent linear regression coefficients, ∈ represents an error term, x represents target data, and y represents a predicted result.

[0072] In some embodiments, the process of building a prediction model includes the following steps:

[0073] S1: Use the torch.utils.data.Dataset class or its subclasses to load the enterprise self-diagnosis data and obtain individual samples and the total number of samples.

[0074] S2: For the image data in the enterprise self-diagnosis data, use the predefined datasets in torchvision.datasets and perform data preprocessing through torchvision.transforms.

[0075] S3: Use torch.utils.data.DataLoader to encapsulate the dataset into an iterable data loader to support batch loading, shuffled data, and multi-process loading functions.

[0076] S4: Define the parameters of the prediction model and set the loss function and optimizer.

[0077] Code example:

[0078] import torch.nn as nn

[0079] class MyModel(nn.Module):

[0080] def __init__(self):

[0081] super(MyModel,self).__init__()

[0082] self.conv1=nn.Conv2d(1,20,5)#1input channel,20output channels,5x5kernel

[0083] self.pool = nn.MaxPool2d(2,2)

[0084] self.fc2 = nn.Linear(100,10) #Assume there are 10 categories

[0085] def forward(self,x):

[0086] x=self.pool(F.relu(self.conv1(x)))

[0087] x=x.view(-1,20*12*12)#Flatten

[0088] x = F.relu(self.fc1(x))

[0089] x = self.fc2(x)

[0090] return x.

[0091] Preferably, select the cross entropy loss nn.CrossEntropyLoss loss function. Select the SGD optimizer and pass in the model parameters.

[0092] Code Explanation:

[0093] import torch.optim as optim

[0094] loss_fn=nn.CrossEntropyLoss()

[0095] optimizer=optim.Adam(model.parameters(),lr=0.001).

[0096] S5: Traverse multiple epochs, and in each epoch, iteratively load data through DataLoader.

[0097] S6: Perform forward propagation on each batch of data, calculate the output and loss, perform backpropagation, calculate the gradient of the loss with respect to the model parameters, use the optimizer to update the parameters of the prediction model, and obtain the prediction model.

[0098] Code Explanation:

[0099]

[0100] In some embodiments, the optimization evaluation process of the prediction model includes the following steps:

[0101] S1: Obtain the enterprise self-diagnosis dataset and divide it into features X and labels y.

[0102] S2: Input feature X into the prediction model to obtain the predicted label.

[0103] S3: Use the cross_val_score function or cross_validate function of scikit-learn for cross validation; the cross_val_score function returns an array containing the score for each iteration, and the cross_validate function returns the mean, standard deviation, and training time for each scoring metric.

[0104] S4: Adjust the parameters of the prediction model according to the results of cross-validation to optimize the performance of the prediction model.

[0105] In some embodiments, the evaluation pattern:

[0106] Calling the model's evaluate method sets the model into evaluation mode.

[0107] Evaluation process: traverse the test set (or validation set), use the model to make predictions, and calculate evaluation indicators (such as accuracy, precision, recall, etc.).

[0108] Save the model: Use torch.save() to save the model parameters (state_dict) or the entire model object. When saving the model, you can choose to save only the model parameters, which can reduce storage space and be more flexible when loading.

[0109] Code Explanation:

[0110] torch.save(model.state_dict(),'model_weights.pth')

[0111] #Or save the entire model

[0112] torch.save(model,'model.pth')

[0113] Evaluate the model using scikit-learn's evaluation tools such as cross validation.

[0114] In some embodiments, based on the results of cross-validation, the parameters of the model are adjusted (such as using a grid search GridSearchCV or a random search RandomizedSearchCV) or different models are tried to improve the performance of the model.

[0115] d) scikit-learn evaluation code framework:

[0116] from sklearn.model_selection import train_test_split

[0117] from sklearn.metrics import mean_squared_error

[0118] X_train,

[0119] #Assume that model is a trained PyTorch model, and use its prediction function here

[0120] model.eval()

[0121] with torch.no_grad():

[0122] predictions=model(torch.tensor(X_test,

[0123] dtype=torch.float32)).numpy().flatten()

[0124] mse=mean_squared_error(y_test,predictions)

[0125] print(f'Test MSE:{mse}')

[0126] In some embodiments, a trained model is used for prediction, and the specific prediction process is described as follows:

[0127] 1. Load new enterprise diagnostic data

[0128] Use Pandas to load new enterprise diagnostic data. This usually involves reading a CSV file or other type of data file.

[0129] import pandas as pd

[0130] #Assume the data file is named 'new_data.csv'

[0131] new_data=pd.read_csv('new_data.csv')

[0132] 2. Data preprocessing

[0133] New data is preprocessed, including handling missing values, encoding categorical variables (such as using one-hot encoding or label encoding), standardizing or normalizing numerical features, etc.

[0134] #Handling missing values

[0135] new_data.fillna(method='ffill',inplace=True)#Fill missing values ​​forward, the specific method is selected according to the data situation

[0136] #Encode categorical variables (assuming 'category' is the column that needs to be encoded)

[0137] #Note: You need to choose the appropriate encoding method according to the actual situation

[0138] #new_data['category_encoded']=pd.Categorical(new_data['category']).cat.co des

[0139] #If feature scaling is used during training, the same scaling parameters need to be applied

[0140] #from sklearn.preprocessing import StandardScaler

[0141] #scaler = StandardScaler() #Assume this is the scaler used during training

[0142] #new_data_scaled = scaler.transform(new_data[['feature1','feature2']]) #Assume that only these two columns are scaled

[0143] #Note: Here you should only scale the feature columns that were scaled during training

[0144] # If your data preprocessing is more complex, make sure to replicate the training steps here exactly

[0145] 3. Prepare data for prediction

[0146] Make sure your data format matches the format the model was trained in. This usually means you need to select the correct feature columns and possibly convert them to the input format the model expects (such as NumPy arrays).

[0147] #Assume that the model only uses 'feature1', 'feature2' as input

[0148] X_new=new_data[['feature1','feature2']]

[0149] #Convert to NumPy array if necessary

[0150] X_new_array = X_new.values

[0151] 4. Use the trained model to make predictions

[0152] Use the trained model to make predictions on new data.

[0153] #predictions = model.predict(X_new_array) #For classification or regression problems

[0154] #probabilities = model.predict_proba(X_new_array)[:,1]#Assume it is a binary classification problem and get the probability of the positive class.

[0155] 5. Processing prediction results

[0156] Finally, you can convert the predictions back to the original units (if scaled), calculate performance metrics (such as precision, recall, etc., although this is usually done on a test set), or save the predictions back to a CSV file for further analysis.

[0157] Python

[0158] #Example: Save the prediction results back to a CSV file

[0159] #predictions_df=pd.DataFrame(predictions, columns=['Predicted_Class'])

[0160] #predictions_df.to_csv('predictions.csv',index=False)

[0161] Step 102: Generate a recommendation plan based on the enterprise's real application classification and the enterprise's actual situation.

[0162] In some embodiments, the recommendation logic used to generate the recommendation scheme is as follows: sort out all application categories and match the categories corresponding to the procurement applications, and sort out the application categories corresponding to each item in the evaluation table; according to the evaluation question table, obtain the enterprise evaluation score, and match the enterprise evaluation score with the application provided by the supplier, and set the priority according to the score weight, and arrange the matching applications according to the priority.

[0163] In some implementations, a rules engine is used to generate recommendations based on the prediction results and enterprise-specific needs.

[0164] Code framework:

[0165] #Assume that new_data is the new enterprise diagnostic data

[0166] new_data=pd.DataFrame({'x1':[...],'x2':[...]})

[0167] new_data_scaled=scaler.transform(new_data)

[0168] with torch.no_grad():

[0169] new_predictions=model(torch.tensor(new_data_scaled,dtype=torch.float32)).numpy().flatten()

[0170] # Generate recommendations based on new_predictions application rule engine

[0171] #The implementation of the rule engine will be customized according to business logic and requirements

[0172] Examples of recommended solutions:

[0173] 1. Obtain enterprise self-diagnosis data: conduct diagnostic questions based on the enterprise’s own situation and collect enterprise answer data

[0174] 2. Analyze enterprise data: clarify the scores and application categories corresponding to the questions, obtain the scores corresponding to the enterprises and determine the actual application categories of the enterprises;

[0175] 3. Recommend applications to enterprises based on their actual application classification and actual conditions.

[0176] 4. Recommendation logic:

[0177] (1) Sort out all application categories and match them with the corresponding categories of the purchased applications

[0178] (2) Sort out the application categories corresponding to each item in the evaluation table

[0179] (3) Establish the diagnostic platform evaluation logic based on the evaluation question table to obtain the enterprise evaluation score

[0180] (4) Obtain the application score from the supplier application evaluation interface according to the evaluation form

[0181] (5) Matching user evaluation scores with applications provided by suppliers

[0182] (6) Arrange matching applications according to score weights

[0183] (7) Select applications from top to bottom

[0184] (8) When the selected application score + user self-assessment score + value-added service score is greater than 40 points, the total price of the selected application

[0185] (9) When the price of the selected application meets the coupon amount, the above selected application is a required application. Select 2 applications (the added points are less than 60 points) as the secondary package application

[0186] (10) When the price of the selected application does not meet the discount amount, select several applications so that the total price of all applications meets the coupon amount. The above selected applications are required applications, and then select 2 applications (the added points are less than 60 points) as the secondary package.

[0187] (11) When the user's evaluation score exceeds 40 points, it reaches the second level and only the third and fourth level packages are recommended.

[0188] When the user evaluation score exceeds 60 points and reaches level 3, only level 4 packages are recommended.

[0189] In some embodiments, the feedback collection and model optimization steps describe:

[0190] Develop a web interface to collect user feedback.

[0191] Use the AutoML tool to iteratively optimize the model based on feedback data. The specific steps include:

[0192] 1. Choose an AutoML tool

[0193] Choose an AutoML tool like Auto-sklearn that fits your problem and dataset.

[0194] 2. Configure AutoML tasks

[0195] Set task type: Configure the AutoML task according to the problem type (such as classification, regression).

[0196] Define the search space: Some AutoML tools allow you to customize the search space of the algorithm and parameters.

[0197] Resource constraints: Set constraints such as time budget, computing resources (CPU / GPU), etc.

[0198] 3. Initial model training and evaluation

[0199] Automatic training: The AutoML tool will automatically try different algorithms and parameter combinations to train the model using the training set.

[0200] Initial evaluation: Use the validation set to evaluate the performance of the models and select the best performing model.

[0201] 4. Iterative Optimization

[0202] Collect feedback: Collect feedback on preliminary model results based on business or technical needs.

[0203] Adjust search strategy: Adjust AutoML’s search strategy based on feedback, such as increasing or decreasing the search space and adjusting performance evaluation metrics.

[0204] Retrain: Retrain the model using the updated settings and evaluate the performance.

[0205] 5. Final model evaluation and deployment

[0206] Final evaluation: Evaluate the performance of the final model on the test set to ensure there is no overfitting.

[0207] Model explanation (optional): If needed, explain the reasons for the model's predictions to increase transparency.

[0208] Deployment: Deploy the model to production, monitor its performance, and update it as needed.

[0209] 6. Continuous optimization

[0210] Monitoring: Regularly monitor the performance of the model to ensure it still meets business requirements.

[0211] Update: Update the model or retrain it based on newly collected data or business changes.

[0212] Code framework:

[0213] from flask import Flask,request,jsonify

[0214] app = Flask(__name__)

[0215] @app.route(' / feedback',methods=['POST'])

[0216] def collect_feedback():

[0217] data=request.json

[0218] #Assume data contains {'prediction':model_prediction,'actual':actual_result,'feedback':user_feedback}

[0219] #Here the data is stored in the database or used for model training

[0220] return jsonify({'status':'success'})

[0221] #The AutoML part needs to rely on specific AutoML libraries and processes, such as TPOT, Auto-Keras, etc.

[0222] The beneficial effects achieved by this application are as follows:

[0223] a) Provide personalized digital transformation advice and resource recommendations for small and medium-sized enterprises to reduce their transformation costs and risks.

[0224] b) Accelerate the digital transformation of small and medium-sized enterprises and improve their market competitiveness and sustainable development capabilities.

[0225] c) Improve data utilization efficiency and decision-making quality of small and medium-sized enterprises and optimize operational processes.

[0226] See also Figure 2 FIG. 1 is a schematic diagram of a digital application recommendation system provided by an embodiment of the present invention, including:

[0227] The acquisition unit 20 is used to acquire the enterprise self-diagnosis data and process the enterprise self-diagnosis data to obtain target data;

[0228] A prediction unit 21 is used to input the target data into a pre-built prediction model for prediction to obtain the enterprise's real application classification, wherein the prediction model is built based on a linear regression function;

[0229] The recommendation unit 22 is used to generate a recommendation plan based on the actual application classification of the enterprise and the actual situation of the enterprise.

[0230] In some implementations, the acquisition unit 20 is specifically configured to:

[0231] Use Talend tools to connect to data sources, import enterprise self-diagnosis data, and perform data consistency checks to repair or delete inconsistent data records;

[0232] Use Talend tools to remove duplicates, process missing values, and process outliers in the enterprise self-diagnosis data to obtain target data.

[0233] In some implementations, the prediction unit 21 is specifically configured to:

[0234] Use the torch.utils.data.Dataset class or its subclasses to load the enterprise self-diagnosis data and obtain individual samples and the total number of samples;

[0235] For the image data in the enterprise self-diagnosis data, use the predefined datasets in torchvision.datasets and perform data preprocessing through torchvision.transforms;

[0236] Use torch.utils.data.DataLoader to encapsulate the dataset into an iterable data loader to support batch loading, shuffling data, and multi-process loading.

[0237] Define the parameters of the prediction model and set the loss function and optimizer;

[0238] Traverse multiple epochs, and in each epoch, iteratively load data through DataLoader;

[0239] Perform forward propagation on each batch of data to calculate the output and loss, perform backpropagation to calculate the gradient of the loss with respect to the model parameters, use the optimizer to update the parameters of the prediction model, and obtain the prediction model.

[0240] On the other hand, the present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned digital transformation application recommendation method when executing the program.

[0241] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned digital transformation application recommendation method.

[0242] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0243] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0244] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0245] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0246] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0247] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0248] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0249] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0250] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A digital transformation application recommendation method, characterized in that: include: Acquire enterprise self-diagnosis data, and process the enterprise self-diagnosis data to obtain target data; Input the target data into a pre-built prediction model for prediction to obtain the enterprise's real application classification, wherein the prediction model is built based on a linear regression function; Generate recommended solutions based on the enterprise's real application classification and actual situation.

2. The digital transformation application recommendation method according to claim 1, characterized in that: Acquiring enterprise self-diagnosis data and processing the enterprise self-diagnosis data to obtain target data, including: Use Talend tools to connect to data sources, import enterprise self-diagnosis data, and perform data consistency checks to repair or delete inconsistent data records; Use Talend tools to remove duplicates, process missing values, and process outliers in the enterprise self-diagnosis data to obtain target data.

3. The digital transformation application recommendation method according to claim 1, characterized in that: The process of building a prediction model includes: Use the torch.utils.data.Dataset class or its subclasses to load the enterprise self-diagnosis data and obtain individual samples and the total number of samples; For the image data in the enterprise self-diagnosis data, use the predefined datasets in torchvision.datasets and perform data preprocessing through torchvision.transforms; Use torch.utils.data.DataLoader to encapsulate the dataset into an iterable data loader to support batch loading, shuffling data, and multi-process loading. Define the parameters of the prediction model and set the loss function and optimizer; Traverse multiple epochs, and in each epoch, iteratively load data through DataLoader; Perform forward propagation on each batch of data to calculate the output and loss, perform backpropagation to calculate the gradient of the loss with respect to the model parameters, use the optimizer to update the parameters of the prediction model, and obtain the prediction model.

4. The digital transformation application recommendation method according to claim 3, characterized in that: The optimization evaluation process of the prediction model includes: Obtain the enterprise self-diagnosis dataset and divide it into feature X and label y; Input feature X into the prediction model to obtain the prediction label; Use scikit-learn's cross_val_score function or cross_validate function for cross validation; the cross_val_score function returns an array containing the score for each iteration, and the cross_validate function returns the mean, standard deviation, and training time for each score metric; The parameters of the prediction model are adjusted according to the results of cross-validation to optimize the performance of the prediction model.

5. The digital transformation application recommendation method according to claim 1, characterized in that: The recommendation logic used to generate the recommendation solution is as follows: Sort out all application categories and match them with the corresponding categories of the purchased applications, and sort out the application categories corresponding to each item in the evaluation form; According to the evaluation question table, the enterprise evaluation scores are obtained, and the enterprise evaluation scores are matched with the applications provided by the supplier. Priorities are set according to the score weights, and matching applications are arranged according to the priorities.

6. A digital application recommendation system, characterized in that: include: An acquisition unit, used for acquiring enterprise self-diagnosis data, and processing the enterprise self-diagnosis data to obtain target data; A prediction unit, used for inputting the target data into a pre-built prediction model for prediction to obtain the enterprise's real application classification, wherein the prediction model is built based on a linear regression function; The recommendation unit is used to generate recommendation solutions based on the enterprise's real application classification and the enterprise's actual situation.

7. The digital application recommendation system according to claim 6, characterized in that: The acquisition unit is specifically used for: Use Talend tools to connect to data sources, import enterprise self-diagnosis data, and perform data consistency checks to repair or delete inconsistent data records; Use Talend tools to remove duplicates, process missing values, and process outliers in the enterprise self-diagnosis data to obtain target data.

8. The digital application recommendation system according to claim 6, characterized in that: The prediction unit is specifically used for: Use the torch.utils.data.Dataset class or its subclasses to load the enterprise self-diagnosis data and obtain individual samples and the total number of samples; For the image data in the enterprise self-diagnosis data, use the predefined datasets in torchvision.datasets and perform data preprocessing through torchvision.transforms; Use torch.utils.data.DataLoader to encapsulate the dataset into an iterable data loader to support batch loading, shuffling data, and multi-process loading. Define the parameters of the prediction model and set the loss function and optimizer; Traverse multiple epochs, and in each epoch, iteratively load data through DataLoader; Perform forward propagation on each batch of data to calculate the output and loss, perform backpropagation to calculate the gradient of the loss with respect to the model parameters, use the optimizer to update the parameters of the prediction model, and obtain the prediction model.

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 program, the steps of the digital transformation application recommendation method as described in any one of claims 1 to 5 are implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the digital transformation application recommendation method as described in any one of claims 1 to 5 are implemented.

Images