Logistics order evaluation data processing method and device, equipment and storage medium
Through deep learning models and attention mechanisms, the problem of inefficiency of traditional methods is solved, and service problems are efficiently discovered and service processes are optimized, and user satisfaction is improved.
Patent Information
- Application Number
- CN202410886258.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-03
- Publication Date
- 2024-12-06
AI Technical Summary
The traditional logistics order evaluation data processing method is inefficient, and it is difficult to accurately capture the deep meaning and nuances of user evaluation. It is impossible to detect problems in logistics services in a timely manner, resulting in a decline in service levels and a decrease in user satisfaction.
By collecting and cleaning logistics order evaluation data, establishing a deep learning network model, using emotion classification data sets for training, combining attention mechanisms to extract key information from the evaluation text, generating improvement suggestions strategies, and optimizing service processes.
It improves the efficiency of logistics order evaluation data processing, promptly discover service problems, provides improvement basis, and improves user satisfaction and service level.
Smart Images

Figure CN119090322A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of information processing technology, and in particular to a logistics order evaluation data processing method, device, equipment and storage medium. Background Art
[0002] With the rapid rise of global e-commerce, consumers are increasingly dependent on online shopping, which has prompted the logistics industry to undertake unprecedented cargo transportation tasks. In such a highly competitive market environment, the quality of logistics services is directly related to the reputation of e-commerce platforms and user loyalty. Therefore, the evaluation of logistics service quality has become crucial. It is not only an important indicator to measure the efficiency of the logistics industry, but also the key to improving user satisfaction and promoting the sustainable development of e-commerce. However, traditional logistics order evaluation data processing methods, such as manual review and simple statistical analysis methods, have been unable to cope with the processing needs of massive order evaluation information in the big data era. These methods are not only inefficient, but also difficult to accurately capture the deep meaning and subtle differences of user evaluations. They cannot promptly discover problems in logistics services, take timely measures and optimize service processes, reduce the overall service level, thereby reducing user satisfaction and limiting the further improvement of logistics service quality. Summary of the invention
[0003] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a logistics order evaluation data processing method, device, equipment and storage medium with high evaluation data processing efficiency, which can promptly discover problems in logistics services and provide a basis for improvement for logistics service providers. This helps logistics service providers to take timely measures and optimize service processes, improve the overall service level, and thus improve user satisfaction.
[0004] The first aspect of the present invention provides a method for processing logistics order evaluation data, including: collecting logistics order evaluation data, cleaning the logistics order evaluation data to obtain an evaluation text data set; obtaining a historical evaluation data set, marking the historical evaluation data set to obtain a sentiment classification data set; establishing a deep learning network model, using the sentiment classification data set to train the deep learning network model to obtain a service quality analysis model; inputting the evaluation text data set into the service quality analysis model, and using an attention mechanism to extract key information from the evaluation text data set; analyzing according to the key information to obtain analysis results of specific service quality problems, and generating improvement suggestion strategies based on the analysis results.
[0005] Optionally, in a first implementation method of the first aspect of the present invention, the logistics order evaluation data is collected, the logistics order evaluation data is cleaned, and an evaluation text data set is obtained, including: creating an order evaluation listener, and using the order evaluation listener to listen to the target evaluation information uploaded by the client; when the target evaluation information is listened to, the current target evaluation information is obtained; the target evaluation information is collected according to a preset time interval to obtain logistics order evaluation data; duplicate data in the logistics order evaluation data is deleted to obtain a deduplicated data set; and the deduplicated data set is converted into a standardized format to obtain an evaluation text data set.
[0006] Optionally, in a second implementation method of the first aspect of the present invention, the acquiring of a historical evaluation data set, marking the historical evaluation data set, and obtaining a sentiment classification data set include: acquiring a historical evaluation data set; parsing the historical evaluation data set to obtain a plurality of evaluation records, wherein the evaluation records include text evaluation content and evaluation scores corresponding to the text evaluation content; based on a preset vocabulary, identifying keywords in the text evaluation content, and determining a user feedback type according to the keywords; and marking a plurality of corresponding user feedback types according to a plurality of the evaluation scores to obtain a sentiment classification data set.
[0007] Optionally, in a third implementation method of the first aspect of the present invention, the deep learning network model is established, and the deep learning network model is trained using the sentiment classification data set to obtain a service quality analysis model, including: establishing a deep learning network model; performing feature analysis on the sentiment classification data set to obtain an order feature set; performing feature weight configuration and vector encoding on the order feature set to obtain a historical order evaluation vector; training the deep learning network model using the historical order evaluation vector to obtain an initial model; optimizing the initial model through cross-validation and hyperparameter adjustment to obtain a service quality analysis model.
[0008] Optionally, in a fourth implementation method of the first aspect of the present invention, the evaluation text dataset is input into the service quality analysis model, and key information is extracted from the evaluation text dataset using an attention mechanism, including: preprocessing the evaluation text dataset to obtain a preprocessed dataset; inputting the preprocessed dataset into the service quality analysis model, and calculating an average score of the preprocessed dataset; when the average score is less than or equal to a preset benchmark threshold, extracting key information from the preprocessed dataset using an attention mechanism.
[0009] Optionally, in a fifth implementation of the first aspect of the present invention, analyzing the key information to obtain analysis results of specific service quality problems, and generating improvement suggestion strategies based on the analysis results, includes: identifying intent of the key information through a natural language understanding engine to obtain key intent; analyzing the key intent to obtain analysis results of specific service quality problems; calling a preset strategy library to match the analysis results to obtain improvement suggestion strategies corresponding to the analysis results.
[0010] Optionally, in a sixth implementation method of the first aspect of the present invention, after analyzing the key information to obtain analysis results of specific service quality problems and generating improvement suggestion strategies based on the analysis results, it also includes: generating an improvement suggestion report based on the improvement suggestion strategy, and recording the generation time; uploading the improvement suggestion report and the generation time to the blockchain; and sending the improvement suggestion report to a management terminal so that the management terminal generates and displays an improvement suggestion page based on the improvement suggestion report.
[0011] The second aspect of the present invention provides a logistics order evaluation data processing device, including: a collection and cleaning module, used to collect logistics order evaluation data, clean the logistics order evaluation data, and obtain an evaluation text data set; an acquisition and marking module, used to obtain a historical evaluation data set, mark the historical evaluation data set, and obtain a sentiment classification data set; a training module, used to establish a deep learning network model, use the sentiment classification data set to train the deep learning network model, and obtain a service quality analysis model; an input extraction module, used to input the evaluation text data set into the service quality analysis model, and use the attention mechanism to extract key information from the evaluation text data set; an analysis generation module, used to analyze according to the key information, obtain analysis results of specific service quality problems, and generate improvement suggestion strategies according to the analysis results.
[0012] Optionally, in a first implementation method of the second aspect of the present invention, the collection and cleaning module includes: creating a monitoring unit, used to create an order evaluation listener, and using the order evaluation listener to monitor the target evaluation information uploaded by the client; a first acquisition unit, used to obtain the current target evaluation information when the target evaluation information is monitored; a collection unit, used to collect the target evaluation information according to a preset time interval to obtain logistics order evaluation data; a deletion unit, used to delete duplicate data in the logistics order evaluation data to obtain a deduplicated data set; a conversion unit, used to convert the deduplicated data set into a standardized format to obtain an evaluation text data set.
[0013] Optionally, in a second implementation of the second aspect of the present invention, the acquisition and marking module includes: a second acquisition unit, used to acquire a historical evaluation data set; a parsing unit, used to parse the historical evaluation data set to obtain multiple evaluation records, wherein the evaluation records include text evaluation content and evaluation scores corresponding to the text evaluation content; an identification and determination unit, used to identify keywords in the text evaluation content based on a preset vocabulary, and determine the user feedback type according to the keywords; and a marking unit, used to mark multiple corresponding user feedback types according to the multiple evaluation scores to obtain a sentiment classification data set.
[0014] Optionally, in a third implementation method of the second aspect of the present invention, the establishment of the training module includes: an establishment unit, used to establish a deep learning network model; a first analysis unit, used to perform feature analysis on the sentiment classification data set to obtain an order feature set; an encoding unit, used to perform feature weight configuration and vector encoding on the order feature set to obtain a historical order evaluation vector; a training unit, used to train the deep learning network model using the historical order evaluation vector to obtain an initial model; and an optimization unit, used to optimize the initial model through cross-validation and hyperparameter adjustment to obtain a service quality analysis model.
[0015] Optionally, in a fourth implementation of the second aspect of the present invention, the input extraction module includes: a preprocessing unit, used to preprocess the evaluation text data set to obtain a preprocessed data set; an input calculation unit, used to input the preprocessed data set into the service quality analysis model and calculate the average score of the preprocessed data set; an extraction unit, used to extract key information from the preprocessed data set using an attention mechanism when the average score is less than or equal to a preset benchmark threshold.
[0016] Optionally, in a fifth implementation of the second aspect of the present invention, the analysis generation module includes: an identification unit, used to identify the intent of the key information through a natural language understanding engine to obtain the key intent; a second analysis unit, used to analyze according to the key intent to obtain an analysis result of a specific problem of service quality; and a matching unit, used to call a preset strategy library to match the analysis result to obtain an improvement suggestion strategy corresponding to the analysis result.
[0017] Optionally, in a sixth implementation manner of the second aspect of the present invention, it also includes: a record generation module, used to generate an improvement suggestion report according to the improvement suggestion strategy, and record the generation time; an upload module, used to upload the improvement suggestion report and the generation time to the blockchain; a sending module, used to send the improvement suggestion report to the management terminal, so that the management terminal generates and displays an improvement suggestion page based on the improvement suggestion report.
[0018] The third aspect of the present invention provides a logistics order evaluation data processing device, which includes: a memory and at least one processor, wherein instructions are stored in the memory; at least one of the processors calls the instructions in the memory to enable the logistics order evaluation data processing device to execute each step of the logistics order evaluation data processing method described in any one of the above items.
[0019] A fourth aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, and when the instructions are executed by a processor, the various steps of any of the above-mentioned logistics order evaluation data processing methods are implemented.
[0020] In the technical solution of the present invention, by marking the historical evaluation data set, a sentiment classification data set is obtained, a deep learning network model is established, the sentiment classification data set is used to train the deep learning network model, and a service quality analysis model is obtained. The evaluation text data set is input into the service quality analysis model, and the attention mechanism is used to extract key information from the evaluation text data set. The evaluation data processing efficiency is high, and analysis is performed based on the key information to obtain analysis results of specific service quality problems. Improvement suggestion strategies are generated based on the analysis results, which can timely discover problems in logistics services and provide improvement basis for logistics service providers. This helps logistics service providers to take timely measures and optimize service processes, improve the overall service level, and thus enhance user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 A first flow chart of a method for processing logistics order evaluation data provided by an embodiment of the present invention;
[0022] Figure 2 A second flow chart of the logistics order evaluation data processing method provided by an embodiment of the present invention;
[0023] Figure 3 A third flow chart of the logistics order evaluation data processing method provided by an embodiment of the present invention;
[0024] Figure 4 A fourth flow chart of the logistics order evaluation data processing method provided by an embodiment of the present invention;
[0025] Figure 5 A schematic diagram of the structure of a logistics order evaluation data processing device provided by an embodiment of the present invention;
[0026] Figure 6 Another structural schematic diagram of a logistics order evaluation data processing device provided by an embodiment of the present invention;
[0027] Figure 7 A schematic diagram of the structure of a logistics order evaluation data processing device provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0028] The present invention provides a logistics order evaluation data processing method, device, equipment and storage medium. The evaluation data processing efficiency is high, problems in logistics services can be discovered in a timely manner, and a basis for improvement can be provided for logistics service providers, which helps logistics service providers to take timely measures and optimize service processes, improve the overall service level, and thus improve user satisfaction.
[0029] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0030] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of the method for processing logistics order evaluation data in an embodiment of the present invention includes:
[0031] 101. Collect logistics order evaluation data, clean the logistics order evaluation data, and obtain an evaluation text data set;
[0032] In this embodiment, the specific source from which the evaluation data is to be collected is determined, the evaluation data is captured from the data source using an API interface, a crawler, or other data capture tools, the captured data is stored in a database or file, duplicate evaluation records are checked and deleted, and for some data that is missing evaluation content or ratings, deletion or filling is selected according to the specific situation, and information not related to the logistics order evaluation, such as labels, non-text content, etc., is deleted, and the evaluation text is converted into a unified format, such as lowercase letters, removal of special characters, etc., and necessary corrections or deletions are made to the evaluation text containing spelling errors, garbled characters, or format errors.
[0033] 102. Obtain a historical evaluation data set, mark the historical evaluation data set, and obtain a sentiment classification data set;
[0034] In this embodiment, the source of the evaluation data is clarified to ensure the reliability and relevance of the data, the evaluation data is extracted from the data source, the extracted data is stored in an appropriate database or data structure, the sentiment labels to be used are determined, such as positive, negative or neutral, and these labels should be able to accurately reflect the sentiment tendency of the evaluation text, and sentiment labeling is performed using a pre-trained model or an automated tool. The labeled data is verified to ensure the accuracy of the labeling. After the above steps, an evaluation data set with sentiment labels is obtained, namely, a sentiment classification data set. This data set can be used to train a sentiment classification model to predict the sentiment tendency of new evaluation texts.
[0035] 103. Establish a deep learning network model, use the sentiment classification data set to train the deep learning network model, and obtain a service quality analysis model;
[0036] In this embodiment, a suitable deep learning model architecture is selected according to the task requirements, such as a recurrent neural network (RNN), a long short-term memory network (LSTM), a convolutional neural network (CNN) or a Transformer, etc., the number of network layers, the number of neurons in each layer, the activation function and other parameters are determined, initial values are set for parameters such as the weight and bias of the model, and necessary preprocessing is performed on the sentiment classification data set, such as text cleaning, word segmentation, stop word removal, vocabulary building, etc. For text data, the text needs to be converted into a numerical form that the model can understand, such as a word embedding vector, and the preprocessed data set is divided into a training set, a validation set and a test set. The training set is used to train the model, the validation set is used to adjust the hyperparameters of the model, and the test set is used to evaluate the performance of the model. The number of training iterations (epochs), batch size, and learning rate (learning) are set. rate) and other parameters, and use the training set and the set training parameters to train the deep learning network model. During the training process, the model will adjust its parameters such as weights and biases according to the input data and labels to minimize the loss function. At the end of each training cycle, the validation set is used to verify the performance of the model, and the test set is used to evaluate the performance of the model, and related evaluation indicators such as accuracy, recall rate, F1 score, etc. are calculated, and the trained model is deployed to the actual application environment.
[0037] 104. Input the evaluation text dataset into the service quality analysis model, and use the attention mechanism to extract key information from the evaluation text dataset;
[0038] In this embodiment, the preprocessed evaluation text data set is input into the trained service quality analysis model. For each word or phrase in the text, the model calculates an attention weight, which reflects the importance of the word or phrase to the service quality analysis. According to the calculated attention weight, the words or phrases in the text are weighted and summed to obtain a weighted average value representing the entire text vector. This weighted average value is the key information extracted by the model from the text.
[0039] 105. Analyze the key information to obtain the analysis results of the specific problems of service quality, and generate improvement suggestions and strategies based on the analysis results;
[0040] In this embodiment, the extracted key information is carefully read to understand its meaning and context, and points directly related to service quality are identified from the key information, such as service attitude, response speed, product quality, etc. The identified service quality problems are sorted according to different categories, such as service process problems, employee quality problems, etc. Each problem is deeply analyzed to find out the possible causes, such as insufficient training, cumbersome processes, etc. The impact of each problem on the overall service quality is evaluated to determine the priority of the problem, and the specific goals of improving service quality are clarified, such as improving customer satisfaction, reducing complaint rates, etc. According to the problem analysis and impact assessment, targeted improvement strategies are formulated. For example, if the problem lies in employee quality, a training plan can be formulated to improve the professional quality of employees. If the problem lies in the service process, the process can be optimized to improve efficiency. Detailed implementation steps and timetables are formulated for each improvement strategy to ensure the effective implementation of the strategy, evaluate the resources required to implement the improvement strategy, such as human, material, and financial resources, and formulate corresponding resource allocation plans.
[0041] In an embodiment of the present invention, a sentiment classification data set is obtained by marking a historical evaluation data set, a deep learning network model is established, the deep learning network model is trained using the sentiment classification data set, and a service quality analysis model is obtained. The evaluation text data set is input into the service quality analysis model, and key information is extracted from the evaluation text data set using an attention mechanism. The evaluation data processing efficiency is high, and analysis is performed based on the key information to obtain analysis results of specific service quality problems. Improvement suggestion strategies are generated based on the analysis results, so that problems in logistics services can be discovered in a timely manner, and a basis for improvement can be provided for logistics service providers, which helps logistics service providers to take timely measures and optimize service processes, improve the overall service level, and thus enhance user satisfaction.
[0042] See also Figure 2 The second embodiment of the method for processing logistics order evaluation data in the embodiment of the present invention includes:
[0043] 201. Create an order evaluation listener, and use the order evaluation listener to monitor the target evaluation information uploaded by the client;
[0044] In this embodiment, the event type to be monitored is determined. The event type is an HTTP request sent by the client, such as a POST request for uploading evaluation information. A suitable programming language and framework are selected according to the actual situation of the project, such as Java's Spring framework, etc. Listener code is written in the selected framework to receive client requests. In the request processing method, code is written to parse the data sent by the client, which involves deserializing the request body and converting it into an object that can be processed by the program. The parsed evaluation information is verified to ensure the integrity and validity of the data, and the verified evaluation information is stored in the database.
[0045] 202. When target evaluation information is monitored, current target evaluation information is obtained;
[0046] In this embodiment, the order evaluation listener will continue to wait until it receives a matching request from the client and parse the received request, which includes parsing the request body to obtain the sent data. In the parsed data, the listener will check whether there is target evaluation information. Once the listener recognizes that the request contains target evaluation information, it will extract the target evaluation information from the parsed request.
[0047] 203. Collect target evaluation information according to preset time intervals to obtain logistics order evaluation data;
[0048] In this embodiment, according to business needs and the frequency of data updates, a suitable time interval is determined, such as every hour, every day, every week or every month, and a scheduled task is configured in the server or application (such as using Cron Job, Windows Task Scheduler or the scheduled task function in the application) to automatically trigger the collection process of the evaluation information when the set time interval arrives. According to the determined data source, the data source containing the evaluation information is accessed through API calls, database queries or file reading, and data that meets the scope of the target evaluation information is filtered out from the data source. The filtered target evaluation information is stored in a suitable storage medium to obtain logistics order evaluation data.
[0049] 204. Delete duplicate data in the logistics order evaluation data to obtain a deduplicated data set;
[0050] In this embodiment, SQL query statements (such as GROUP BY and HAVING clauses) are used to find duplicate items. Once duplicate data is detected, and if the duplicate data is completely redundant without any additional information, the duplicate items can be directly deleted and only one record is retained. Specifically, the data set is traversed, a set or dictionary is used to track the records that have appeared, and subsequent duplicate items are deleted to obtain a deduplicated data set.
[0051] 205. Convert the deduplicated data set into a standardized format to obtain an evaluation text data set;
[0052] In this embodiment, the target format, that is, the structure, fields and format requirements of the converted data set are clarified, the current structure of the data set, such as the organization of rows and columns, the fields in the data set are identified, and the meaning and purpose of each field are understood. The differences between the current format and the target format are compared to determine what conversions and adjustments are needed. According to the size and complexity of the data set and the requirements of the target format, a suitable conversion tool or method is selected. ETL (Extract, Transform, Load) tools, such as Talend, Apache NiFi, etc., can be used to automate the conversion process. According to the requirements of the target format, the extracted data is converted, including field renaming, data type conversion, format adjustment, etc. The converted data is stored in a new data set to obtain an evaluation text data set.
[0053] In the embodiment of the present invention, the evaluation information is captured in real time by a listener and processed actively, thereby ensuring the immediacy and currency of the data. Operations such as batch processing of data and deletion of duplicate data can improve the efficiency and accuracy of data processing. Operations such as deduplication and standardized format conversion can improve the quality and consistency of data. The preset time interval and standardized data set format can be adjusted and changed according to actual needs to meet different business scenarios and needs.
[0054] See also Figure 3 The third embodiment of the method for processing logistics order evaluation data in the embodiment of the present invention includes:
[0055] 301. Obtain a historical evaluation data set;
[0056] In this embodiment, the historical evaluation data set is obtained through an API interface.
[0057] 302. Parse the historical evaluation data set to obtain multiple evaluation records, where the evaluation records include text evaluation content and evaluation scores corresponding to the text evaluation content;
[0058] In this embodiment, a historical evaluation data set is read using a selected tool or programming language to identify fields in the historical evaluation data set that contain text evaluation content and evaluation scores. During the parsing process, the data is cleaned, such as removing null values, outliers, duplicate values, etc., to ensure that the text evaluation content does not contain HTML tags, special characters, or unnecessary spaces, etc., and each row of the data set is traversed to extract the text evaluation content and the corresponding evaluation scores, and each evaluation record is stored as a dictionary, object, or other data structure to obtain multiple evaluation records.
[0059] 303. Based on a preset word library, identify keywords in the text evaluation content, and determine the user feedback type according to the keywords;
[0060] In this embodiment, each word in the preprocessed text is matched with a preset vocabulary to find out the keywords appearing in the text. A string matching algorithm (such as exact matching, fuzzy matching) or a more complex natural language processing algorithm (such as rule-based matching, model-based matching) can be used to improve the matching accuracy. For multiple keywords appearing in the text, different weights are assigned according to their importance or relevance in the text. According to the results of keyword matching and the established rules, the type of user feedback is determined. The feedback type can be a specific category (such as good reviews, bad reviews, suggestions, complaints, etc.) or a more fine-grained label (such as product quality issues, service attitude issues, etc.).
[0061] 304. Marking a plurality of corresponding user feedback types according to a plurality of evaluation scores to obtain a sentiment classification data set;
[0062] In this embodiment, positive, negative, and neutral sentiment classifications are defined according to business needs. For example, an evaluation score between 4 and 5 may be considered positive, between 1 and 2 as negative, and 3 as neutral. Code or script is written to map the evaluation score to the corresponding sentiment label according to the defined sentiment classification. This mapping process is performed for each user feedback type, and the mapped sentiment label is assigned to the corresponding user feedback type to ensure that each user feedback type has a corresponding sentiment label. All labeled user feedback types are integrated into a data set to ensure that the format and structure of the data set meet the requirements of subsequent analysis or model training. The data set is saved in an appropriate file format, such as CSV, Excel, or a database table to obtain a sentiment classification data set.
[0063] 305. Establish a deep learning network model;
[0064] In this embodiment, according to the nature of the problem and the characteristics of the data, an appropriate deep learning model architecture is selected, and a deep neural network (DNN), convolutional neural network (CNN), recurrent neural network (RNN), etc. can be used to design the structure and parameters of the deep learning network model, including the number of hidden layers, the number of neurons, activation functions, loss functions, and optimization algorithms.
[0065] 306. Perform feature analysis on the sentiment classification data set to obtain an order feature set;
[0066] In this embodiment, relevant features in the sentiment classification data set are extracted, and the features include order amount, sentiment words, keywords, and sentence structure. The total amount of the order may affect the user's satisfaction. Sentiment words in the text are identified using a sentiment dictionary or sentiment analysis tool. These words may directly express the user's emotional tendencies. Keywords related to orders or products, such as "delivery speed" and "product quality", may imply the focus of the user's evaluation. Sentence structure, for example, sentences using negative words (such as "no" and "no") may express negative emotions. After extracting a large number of features, feature selection is required to screen out the features that are most helpful for sentiment classification and obtain an order feature set.
[0067] 307. Perform feature weight configuration and vector encoding on the order feature set to obtain a historical order evaluation vector;
[0068] In this embodiment, the historical data is first analyzed, and the relationship between each feature and the target variable (such as sentiment tendency) is statistically analyzed. The feature importance score automatically calculated during the model training process is used to configure the weights of the features, and the weights of all features are standardized, for example, by normalization or maximum and minimum scaling to ensure that the weights are within a comparable range. The features are then encoded, and finally the feature values of each historical order are combined into one or more numerical vectors according to the corresponding weights and encoding methods, namely, the historical order evaluation vector.
[0069] 308. Use the historical order evaluation vector to train the deep learning network model to obtain an initial model;
[0070] In this embodiment, a training function or loop in a deep learning framework is used to train the model. During the training process, the model will use the training data to update its weights and bias terms to minimize the loss function and optimize the performance. The training process will go through multiple epochs (i.e., the process in which the entire training set is traversed once), and the performance of the model on the validation set is evaluated after each epoch. After the training is completed, the initial model is obtained.
[0071] 309. The initial model is optimized through cross-validation and hyperparameter adjustment to obtain a service quality analysis model;
[0072] In this embodiment, in cross-validation, the training set is further subdivided into K subsets (K folds), one subset is selected as the validation set, and the remaining K-1 subsets are used as training sets. The model is trained using these K-1 subsets, and the performance of the model is evaluated using the selected validation set. The performance of the model after K validations is evaluated, and then the hyperparameters that need to be adjusted, such as learning rate, regularization parameter, number of hidden layer neurons, etc., are determined. Grid Search or Random Search and other methods are used to search in the hyperparameter space. For each hyperparameter combination, cross-validation is used to evaluate the performance of the model, so as to optimize the initial model and obtain a service quality analysis model.
[0073] In the embodiment of the present invention, the entire process from data acquisition to model optimization adopts automated and intelligent methods, which greatly improves the efficiency and accuracy of data processing and model training, makes full use of historical evaluation data, and constructs a high-quality service quality analysis model through deep learning technology. The service quality analysis model performs self-optimization and improvement based on new data, further improving the performance and stability of the model, so that the model can better adapt to actual application scenarios. Through the powerful learning ability of the service quality analysis model, it can handle complex nonlinear relationships and accurately analyze and predict service quality.
[0074] See also Figure 4 The fourth embodiment of the method for processing logistics order evaluation data in the embodiment of the present invention includes:
[0075] 401. Preprocess the evaluation text data set to obtain a preprocessed data set;
[0076] In this embodiment, the evaluation text dataset is cleaned to remove content irrelevant to the task, and a stop word list is loaded, which is a predefined list of common stop words. Each word in the evaluation text dataset is traversed, and if the word is in the stop word list, it is deleted. The evaluation text dataset is converted into a unified format to obtain a preprocessed dataset for subsequent processing.
[0077] 402. Input the preprocessed data set into the service quality analysis model, and calculate the average score of the preprocessed data set;
[0078] In this embodiment, the preprocessed data set is used as the input of the service quality analysis model, columns or features related to the service quality score are extracted from the preprocessed data set, and the average value of the score data is calculated using appropriate statistical methods (such as arithmetic mean, weighted average, etc.). Based on the calculated score average, the service quality level of the data set is interpreted. For example, if the score average is very high, it may indicate that the service quality represented by the data set is good. If the score average is low, further investigation and improvement of the service quality may be required.
[0079] 403. When the average score is less than or equal to the preset benchmark threshold, the attention mechanism is used to extract key information from the preprocessed data set;
[0080] In this embodiment, a baseline threshold of service quality is set according to business needs or historical data. This baseline threshold is used to determine whether the service quality meets the standard. The calculated average score is compared with the preset baseline threshold. If the average score is less than or equal to the baseline threshold, it means that there may be problems with the service quality. In this way, the attention mechanism is used to extract key information from the preprocessed data set.
[0081] 404. The key information is identified through a natural language understanding engine to obtain the key intent;
[0082] In this embodiment, the key information is input into a natural language understanding engine, which parses and recognizes the input text to identify the key intent therein.
[0083] 405. Analyze according to the key intent and obtain the analysis results of the specific problems of service quality;
[0084] In this embodiment, the identified key intent is interpreted in depth, and the interpreted key intent is associated with the preset service quality standard. By comparing the intent with the standard, it is possible to preliminarily determine whether there is a problem with the service quality and in which aspect the problem may occur. After determining the area where the problem may exist, a more detailed analysis is performed. By analyzing this data, the specific causes of the service quality problems can be identified, such as system errors, process defects, personnel operation problems, etc., to obtain analysis results.
[0085] 406. Calling a preset strategy library to match the analysis result to obtain an improvement suggestion strategy corresponding to the analysis result;
[0086] In this embodiment, relevant policies or rules are retrieved from the policy library based on the keywords, categories or attributes in the analysis results, and the analysis results are matched with the policies in the policy library using appropriate matching algorithms (such as fuzzy matching, exact matching, rule-based reasoning, etc.). If there are multiple matching policies, they need to be selected based on the priority or applicable conditions of the policy to obtain improved recommended policies corresponding to the analysis results.
[0087] 407. Generate an improvement suggestion report according to the improvement suggestion strategy and record the generation time;
[0088] In this embodiment, specific suggestion content is extracted from the previously matched improvement suggestion strategy, and an improvement suggestion report is generated based on the extracted improvement suggestion strategy. The improvement suggestion report contains key information such as a detailed description of the problem, the cause of the problem, suggested improvement measures, expected effects, implementation steps, responsible persons, and a timetable. The generation time of the report is recorded at the end or in a specific location of the report.
[0089] 408. Upload the improvement suggestion report and generation time to the blockchain;
[0090] In this embodiment, a suitable blockchain platform is determined to store data, and a smart contract is created. The smart contract is a piece of code that runs on the blockchain, which can automatically execute predefined rules and operations. Through the smart contract, it is ensured that only authorized users can access and improve the suggestion report. The hash value and generation time of the improvement suggestion report are uploaded to the blockchain using the API or tools provided by the blockchain platform.
[0091] 409. Sending the improvement suggestion report to the management terminal, so that the management terminal generates and displays an improvement suggestion page based on the improvement suggestion report;
[0092] In this embodiment, according to the characteristics of the management terminal and the requirements of the network environment, a suitable data transmission method is selected, which includes email, FTP (File Transfer Protocol), HTTP (Hypertext Transfer Protocol), etc. According to the selected transmission method, the improvement suggestion report is sent to the management terminal. After the management terminal receives the improvement suggestion report, it needs to parse and process it. Based on the information extracted from the improvement suggestion report, the management terminal generates and displays an improvement suggestion page. This page clearly displays information such as the description of the problem, the scope of impact, the recommended improvement measures, and the expected effects of implementing these measures.
[0093] In the embodiment of the present invention, the entire process improves the efficiency and accuracy of service quality analysis through automation and intelligent means. The system can automatically extract key information from the evaluation text, identify user intentions, analyze service quality issues, and provide targeted improvement suggestions. At the same time, blockchain technology is used to ensure the credibility and security of data, enhance the transparency and traceability of the entire process, and send the improvement suggestion report to the management terminal, so that the management terminal generates and displays the improvement suggestion page based on the improvement suggestion report. This enables organizations to respond to and resolve service quality issues more quickly and effectively, thereby improving customer satisfaction and overall service quality.
[0094] The above describes the logistics order evaluation data processing method in the embodiment of the present invention. The following describes the logistics order evaluation data processing device in the embodiment of the present invention. Figure 5 In one embodiment of the present invention, a logistics order evaluation data processing device includes:
[0095] A collection and cleaning module 501 is used to collect logistics order evaluation data, clean the logistics order evaluation data, and obtain an evaluation text data set;
[0096] The labeling module 502 is used to obtain a historical evaluation data set, label the historical evaluation data set, and obtain a sentiment classification data set;
[0097] Establishing a training module 503 for establishing a deep learning network model, and training the deep learning network model using a sentiment classification data set to obtain a service quality analysis model;
[0098] An input extraction module 504 is used to input the evaluation text data set into the service quality analysis model and extract key information from the evaluation text data set using an attention mechanism;
[0099] The analysis generation module 505 is used to perform analysis based on key information to obtain analysis results of specific service quality issues and generate improvement suggestion strategies based on the analysis results.
[0100] In this embodiment, by marking the historical evaluation data set, a sentiment classification data set is obtained, a deep learning network model is established, and the deep learning network model is trained using the sentiment classification data set to obtain a service quality analysis model. The evaluation text data set is input into the service quality analysis model, and the attention mechanism is used to extract key information from the evaluation text data set. The evaluation data processing efficiency is high, and analysis is performed based on the key information to obtain analysis results of specific service quality problems. Improvement suggestion strategies are generated based on the analysis results, which can timely discover problems in logistics services and provide a basis for improvement for logistics service providers. This helps logistics service providers to take timely measures and optimize service processes, improve the overall service level, and thus enhance user satisfaction.
[0101] See also Figure 6 Another embodiment of the logistics order evaluation data processing device in the embodiment of the present invention includes:
[0102] A collection and cleaning module 501 is used to collect logistics order evaluation data, clean the logistics order evaluation data, and obtain an evaluation text data set;
[0103] The labeling module 502 is used to obtain a historical evaluation data set, label the historical evaluation data set, and obtain a sentiment classification data set;
[0104] Establishing a training module 503 for establishing a deep learning network model, and training the deep learning network model using a sentiment classification data set to obtain a service quality analysis model;
[0105] An input extraction module 504 is used to input the evaluation text data set into the service quality analysis model and extract key information from the evaluation text data set using an attention mechanism;
[0106] Analysis generation module 505, used to analyze the key information, obtain the analysis results of the specific problems of service quality, and generate improvement suggestion strategies according to the analysis results;
[0107] In this embodiment, the collection and cleaning module 501 includes: a monitoring unit 5011 is created, which is used to create an order evaluation listener, and use the order evaluation listener to monitor the target evaluation information uploaded by the client; a first acquisition unit 5012 is used to obtain the current target evaluation information when the target evaluation information is monitored; a collection unit 5013 is used to collect the target evaluation information according to a preset time interval to obtain logistics order evaluation data; a deletion unit 5014 is used to delete duplicate data in the logistics order evaluation data to obtain a deduplicated data set; a conversion unit 5015 is used to convert the deduplicated data set into a standardized format to obtain an evaluation text data set.
[0108] In this embodiment, the acquisition and marking module 502 includes: a second acquisition unit 5021, which is used to acquire a historical evaluation data set; a parsing unit 5022, which is used to parse the historical evaluation data set to obtain multiple evaluation records, and the evaluation records include text evaluation content and evaluation scores corresponding to the text evaluation content; an identification and determination unit 5023, which is used to identify keywords in the text evaluation content based on a preset vocabulary, and determine the user feedback type according to the keywords; a marking unit 5024, which is used to mark multiple corresponding user feedback types according to multiple evaluation scores to obtain a sentiment classification data set.
[0109] In this embodiment, establishing a training module 503 includes: an establishing unit 5031, used to establish a deep learning network model; a first analyzing unit 5032, used to perform feature analysis on the sentiment classification data set to obtain an order feature set; an encoding unit 5033, used to perform feature weight configuration and vector encoding on the order feature set to obtain a historical order evaluation vector; a training unit 5034, used to train the deep learning network model using the historical order evaluation vector to obtain an initial model; an optimizing unit 5035, used to optimize the initial model through cross-validation and hyperparameter adjustment to obtain a service quality analysis model.
[0110] In this embodiment, the input extraction module 504 includes: a preprocessing unit 5041, which is used to preprocess the evaluation text data set to obtain a preprocessed data set; an input calculation unit 5042, which is used to input the preprocessed data set into the service quality analysis model and calculate the average score of the preprocessed data set; an extraction unit 5043, which is used to extract key information from the preprocessed data set using an attention mechanism when the average score is less than or equal to a preset benchmark threshold.
[0111] In this embodiment, the analysis generation module 505 includes: an identification unit 5051, which is used to identify the intent of key information through a natural language understanding engine to obtain the key intent; a second analysis unit 5052, which is used to analyze according to the key intent to obtain the analysis results of the specific problems of the service quality; a matching unit 5053, which is used to call a preset strategy library to match the analysis results to obtain improvement suggestion strategies corresponding to the analysis results.
[0112] In this embodiment, it also includes: a generation record module 506, which is used to generate an improvement suggestion report according to the improvement suggestion strategy and record the generation time; an upload module 507, which is used to upload the improvement suggestion report and the generation time to the blockchain; a sending module 508, which is used to send the improvement suggestion report to the management terminal, so that the management terminal generates and displays an improvement suggestion page based on the improvement suggestion report.
[0113] above Figure 5 and Figure 6The logistics order evaluation data processing device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The logistics order evaluation data processing device in the embodiment of the present invention is described in detail from the perspective of hardware processing.
[0114] Figure 7 6 is a schematic diagram of the structure of a logistics order evaluation data processing device provided by an embodiment of the present invention. The logistics order evaluation data processing device 600 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 610 (for example, one or more processors) and a memory 620, and one or more storage media 630 (for example, one or more mass storage devices) storing application programs 633 or data 632. Among them, the memory 620 and the storage medium 630 can be short-term storage or permanent storage. The program stored in the storage medium 630 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the logistics order evaluation data processing device 600. Furthermore, the processor 610 can be configured to communicate with the storage medium 630, and execute a series of instruction operations in the storage medium 630 on the logistics order evaluation data processing device 600 to implement the steps of the logistics order evaluation data processing method provided by the above-mentioned method embodiments.
[0115] The logistics order evaluation data processing device 600 may also include one or more power supplies 640, one or more wired or wireless network interfaces 650, one or more input and output interfaces 660, and / or one or more operating systems 631, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that Figure 7 The structure of the logistics order evaluation data processing device shown does not constitute a limitation on the logistics order evaluation data processing device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.
[0116] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the logistics order evaluation data processing method.
[0117] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.
[0118] If the integrated unit 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 is essentially or the part that contributes to the prior art or all or 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 a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.
[0119] Finally, it should be noted that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A logistics order evaluation data processing method, It is characterized in that include: Collecting logistics order evaluation data, and cleaning the logistics order evaluation data to obtain an evaluation text data set; Acquire a historical evaluation data set, mark the historical evaluation data set, and obtain a sentiment classification data set; Establishing a deep learning network model, and using the sentiment classification data set to train the deep learning network model to obtain a service quality analysis model; Inputting the evaluation text dataset into the service quality analysis model, and extracting key information from the evaluation text dataset using an attention mechanism; An analysis is performed based on the key information to obtain analysis results of specific service quality issues, and improvement suggestion strategies are generated based on the analysis results.
2. The logistics order evaluation data processing method according to claim 1, It is characterized in that The logistics order evaluation data is collected and cleaned to obtain an evaluation text data set, including: Create an order evaluation listener, and use the order evaluation listener to monitor the target evaluation information uploaded by the client; When the target evaluation information is monitored, the current target evaluation information is obtained; The target evaluation information is collected according to a preset time interval to obtain logistics order evaluation data; Deleting duplicate data in the logistics order evaluation data to obtain a deduplicated data set; The deduplicated data set is converted into a standardized format to obtain an evaluation text data set.
3. The logistics order evaluation data processing method according to claim 1, It is characterized in that The step of acquiring a historical evaluation data set and marking the historical evaluation data set to obtain a sentiment classification data set includes: Obtain historical evaluation data sets; Parsing the historical evaluation data set to obtain a plurality of evaluation records, wherein the evaluation records include text evaluation content and evaluation scores corresponding to the text evaluation content; Based on a preset word library, identifying keywords in the text evaluation content, and determining the user feedback type according to the keywords; A plurality of corresponding user feedback types are respectively marked according to the plurality of evaluation scores to obtain a sentiment classification data set.
4. The logistics order evaluation data processing method according to claim 1, It is characterized in that The step of establishing a deep learning network model and training the deep learning network model using the sentiment classification data set to obtain a service quality analysis model includes: Build deep learning network models; Performing feature analysis on the sentiment classification data set to obtain an order feature set; Performing feature weight configuration and vector encoding on the order feature set to obtain a historical order evaluation vector; Using the historical order evaluation vector to train the deep learning network model to obtain an initial model; The initial model is optimized through cross-validation and hyperparameter adjustment to obtain a service quality analysis model.
5. The logistics order evaluation data processing method according to claim 1, It is characterized in that The step of inputting the evaluation text dataset into the service quality analysis model and extracting key information from the evaluation text dataset using an attention mechanism includes: Preprocessing the evaluation text data set to obtain a preprocessed data set; Inputting the preprocessed data set into the service quality analysis model, and calculating the average score of the preprocessed data set; When the average score is less than or equal to a preset benchmark threshold, the attention mechanism is used to extract key information from the preprocessed data set.
6. The logistics order evaluation data processing method according to claim 1, It is characterized in that The analyzing according to the key information to obtain the analysis results of the specific problems of the service quality and generating improvement suggestion strategies according to the analysis results include: Performing intent recognition on the key information through a natural language understanding engine to obtain the key intent; Analyze according to the key intent to obtain analysis results of specific service quality problems; The preset strategy library is called to match the analysis result to obtain the improvement suggestion strategy corresponding to the analysis result.
7. The logistics order evaluation data processing method according to claim 1, It is characterized in that After analyzing the key information to obtain the analysis results of the specific service quality problems and generating improvement suggestions according to the analysis results, the method further includes: Generate an improvement suggestion report according to the improvement suggestion strategy and record the generation time; Uploading the improvement suggestion report and the generation time to the blockchain; The improvement suggestion report is sent to a management terminal, so that the management terminal generates and displays an improvement suggestion page based on the improvement suggestion report.
8. A logistics order evaluation data processing device, It is characterized in that include: A collection and cleaning module is used to collect logistics order evaluation data, clean the logistics order evaluation data, and obtain an evaluation text data set; An acquisition marking module is used to acquire a historical evaluation data set, mark the historical evaluation data set, and obtain a sentiment classification data set; Establishing a training module for establishing a deep learning network model, and using the sentiment classification data set to train the deep learning network model to obtain a service quality analysis model; An input extraction module, used to input the evaluation text data set into the service quality analysis model, and extract key information from the evaluation text data set using an attention mechanism; The analysis generation module is used to analyze the key information to obtain analysis results of specific service quality problems and generate improvement suggestion strategies based on the analysis results.
9. A logistics order evaluation data processing device, It is characterized in that The logistics order evaluation data processing device comprises: a memory and at least one processor, wherein instructions are stored in the memory; At least one of the processors calls the instructions in the memory so that the logistics order evaluation data processing device executes each step of the logistics order evaluation data processing method as described in any one of claims 1-7.
10. A computer-readable storage medium having instructions stored thereon, It is characterized in that When the instructions are executed by the processor, the various steps of the logistics order evaluation data processing method as described in any one of claims 1-7 are implemented.
Citation Information
Cited By
Government affair service business evaluation method and device, computer equipment and storage medium
CN119809730A
Logistics distribution service evaluation method and device based on voice recognition
CN120766666A
Voice recognition-based logistics distribution service evaluation method and device
CN120766666B
