Method, device and equipment for managing call completing rate of small express delivery personnel and storage medium

Abstract

The invention relates to the technical field of logistics management, in particular to a method, a device, equipment and a storage medium for managing the call completing rate of a small courier. According to the method for managing the call completing rate of the widget, the call completing data of the widget is obtained firstly, and the multi-modal data identification technology is adopted to analyze the call completing data so as to obtain the comprehensive call completing data; then setting an analysis dimension rule, and carrying out statistics and analysis on the comprehensive connection data according to the analysis dimension rule to obtain multiple pieces of scene analysis data; analyzing the scene analysis data by adopting an anomaly analysis model to obtain anomaly data; and finally, calculating the correlation between the abnormal data and each feature data in the call answering data by adopting a correlation analysis method to obtain influence factors, associating the influence factors with the abnormal data, and outputting the scene analysis data, the abnormal data and the influence factors after performing visualization processing on the scene analysis data, the abnormal data and the influence factors. And through intelligent management and multi-dimensional analysis, the management efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics management, and in particular to a method, device, equipment and storage medium for managing the connection rate of small parcel handlers. Background Art

[0002] In the logistics industry, the quality of last-mile delivery services is a core element of customer experience and brand image. As a key link in direct customer contact, the call connection rate of courier delivery personnel directly impacts delivery efficiency and customer satisfaction. However, traditional methods for managing courier call connection rates present numerous challenges, making them unable to meet the efficiency and precision demands of modern logistics services.

[0003] First, data collection is incomplete and lacks real-time performance. Traditional management models rely primarily on manual recording of call information, a method prone to omissions and errors. It also lacks real-time visibility into the delivery drivers' phone connections. Furthermore, call data lacks effective integration with other systems (such as delivery and work order systems), creating information silos that make it difficult to fully reflect actual conditions.

[0004] Secondly, limited analytical capabilities make it difficult to pinpoint issues. Traditional methods typically only measure overall call completion rates and lack multi-dimensional data analysis tools. For example, segmented analysis by region, time period, or customer group is impossible, making it difficult to identify service weaknesses.

[0005] It can be seen that the existing technology still needs to be improved and enhanced. Summary of the Invention

[0006] In view of the above-mentioned deficiencies in the prior art, the purpose of the present invention is to provide a method, device, equipment and storage medium for managing the connection rate of small parcel handlers, aiming to improve the connection rate of small parcel handlers and improve the level of analysis and monitoring, thereby improving customer satisfaction.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] A first aspect of the present invention provides a method for managing the connection rate of small parcel workers, comprising the following steps: obtaining the telephone connection data of the small parcel workers, analyzing the telephone connection data using multimodal data recognition technology to obtain comprehensive connection data; setting analysis dimension rules, and statistically analyzing the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data; analyzing the scenario analysis data using an abnormal analysis model to obtain abnormal data; using a correlation analysis method to calculate the correlation between the abnormal data and each feature data in the telephone connection data to obtain influencing factors, associating the influencing factors with the abnormal data, and visually processing the scenario analysis data, abnormal data and influencing factors and then outputting them.

[0009] Optionally, in a first implementation method of the first aspect of the present invention, the method of obtaining the courier's telephone connection data and analyzing the telephone connection data using multimodal data recognition technology to obtain comprehensive connection data specifically includes: obtaining the courier's telephone connection data, the telephone connection data including active call test data and customer call back data, and extracting call recording data from the telephone connection data; using speech-to-text technology to obtain call text records based on the call recording data; evaluating customer emotions and courier services based on the call text records and call recording data to obtain customer emotion scores and courier communication scores; integrating customer emotion scores and courier communication scores into the telephone connection data to obtain comprehensive connection data.

[0010] Optionally, in a second implementation method of the first aspect of the present invention, the use of speech-to-text technology to obtain call text records based on call recording data specifically includes: obtaining call recording data, using a speech recognition model to identify the language type in the call recording data, and obtaining language type data; based on the language type data, using corresponding speech-to-text technology to convert the call recording data into call text records; using a correction model based on deep learning to correct the call text records in combination with the language type data.

[0011] Optionally, in a third implementation of the first aspect of the present invention, the correction model based on deep learning is used to correct the call text records, specifically including: obtaining language texts from multiple regions and countries, training the Transformer model with the language texts from multiple regions and countries as training sets to obtain a correction model; using the correction model to correct the call text records according to the language type data to obtain a corrected text; translating the corrected text into the official language version to obtain a translated text, and associating the translated text, the corrected text, and the call recording data.

[0012] Optionally, in a fourth implementation of the first aspect of the present invention, the customer emotion and courier service are evaluated based on the call text records and call recording data to obtain the customer emotion score and the courier communication score, specifically including: using blind source separation technology to separate the customer voice and the courier voice in the call recording data to obtain the customer call recording and the courier call recording; using sentiment analysis technology to evaluate the customer emotion and the courier service based on the customer call recording and the courier call recording, respectively, to obtain the customer's first score and the courier's first score; extracting keywords in the call text records, evaluating the customer emotion and the courier service based on the keywords, to obtain the customer's second score and the courier's second score; summarizing the customer's first score, the courier's first score, the customer's second score and the courier's second score according to preset rules to obtain the customer's emotion score and the courier communication score.

[0013] Optionally, in a fifth implementation of the first aspect of the present invention, the analysis dimension rules are set, and the comprehensive connection data is counted and analyzed according to the analysis dimension rules to obtain multiple scenario analysis data, specifically including: setting analysis dimension rules, and the analysis dimensions include time period, region, and call type; for the time period dimension, the telephone connection data is split and analyzed separately according to the time period, and the time period with peak calls is marked; for the regional dimension, separate data analysis is performed at the levels of group, region, business province, distribution, grid and outlet; for call type, analysis is performed separately according to the type of active call test and customer call.

[0014] Optionally, in a sixth implementation of the first aspect of the present invention, the use of an anomaly analysis model to analyze the scene analysis data to obtain anomaly data specifically includes: obtaining historical scene analysis data and call connection data, preprocessing the data and marking the outliers therein to obtain preprocessed data; using the preprocessed data to train an isolation forest model to obtain an anomaly analysis model; associating the scene analysis data with the call connection data to obtain associated data, and inputting the associated data into the anomaly analysis model to obtain anomaly data.

[0015] The second aspect of the present invention provides a device for managing the connection rate of small parcel workers, including: a multimodal analysis module for obtaining the telephone connection data of small parcel workers, and analyzing the telephone connection data using multimodal data recognition technology to obtain comprehensive connection data; a statistical analysis module for setting analysis dimension rules, and performing statistics and analysis on the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data; an abnormality analysis module for analyzing the scenario analysis data using an abnormality analysis model to obtain abnormal data; a correlation analysis module for calculating the correlation between the abnormal data and each feature data in the telephone connection data using a correlation analysis method to obtain influencing factors, associating the influencing factors with the abnormal data, and visually processing the scenario analysis data, abnormal data and influencing factors and outputting them.

[0016] Optionally, in a first implementation method of the second aspect of the present invention, the multimodal analysis module includes: an extraction submodule for obtaining the courier's telephone connection data, the telephone connection data including active call test data and customer callback data, and extracting call recording data from the telephone connection data; a conversion submodule for using speech-to-text technology to obtain call text records based on call recording data; a scoring submodule for evaluating customer emotions and courier services based on call text records and call recording data, and obtaining customer emotion scores and courier communication scores; and an integration submodule for integrating customer emotion scores and courier communication scores into the telephone connection data to obtain comprehensive connection data.

[0017] Optionally, in a second implementation of the second aspect of the present invention, the conversion submodule includes: an identification unit, used to obtain call recording data, and use a speech recognition model to identify the language type in the call recording data to obtain language type data; a conversion unit, used to convert the call recording data into a call text record based on the language type data and using corresponding speech-to-text technology; and a correction unit, used to use a deep learning-based correction model in combination with the language type data to correct the call text record.

[0018] Optionally, in a third implementation of the second aspect of the present invention, the correction unit includes: a construction subunit, used to obtain language texts of multiple regions and countries, and train the Transformer model with the language texts of multiple regions and countries as training sets to obtain a correction model; a correction subunit, used to use the correction model to correct the call text record according to the language type data to obtain a corrected text; a translation subunit, used to translate the corrected text into the official language version to obtain a translated text, and associate the translated text, the corrected text and the call recording data.

[0019] Optionally, in a fourth implementation of the second aspect of the present invention, the scoring submodule includes: a separation unit, used to use blind source separation technology to separate the customer voice and the courier voice in the call recording data, and obtain the customer call recording and the courier call recording; a first evaluation unit, used to use sentiment analysis technology to evaluate the customer's emotions and the courier's service based on the customer call recording and the courier's call recording, respectively, to obtain the customer's first score and the courier's first score; a second evaluation unit, used to extract keywords in the call text record, evaluate the customer's emotions and the courier's service based on the keywords, and obtain the customer's second score and the courier's second score; a summary calculation unit, used to summarize the customer's first score, the courier's first score, the customer's second score and the courier's second score according to preset rules, and obtain the customer's emotion score and the courier's communication score.

[0020] Optionally, in a fifth implementation of the second aspect of the present invention, the statistical analysis module includes: a setting unit for setting analysis dimension rules, the analysis dimensions including time period, region, and call type; a first analysis unit for splitting and separately analyzing the call connection data according to the time period for the time period dimension, and marking the time period with peak calls; a second analysis unit for performing separate data analysis at the level of group, region, business province, distribution, grid, and outlet for the regional dimension; and a third analysis unit for analyzing the call type separately according to the type of active call test and the type of customer call.

[0021] Optionally, in a sixth implementation of the second aspect of the present invention, the anomaly analysis module includes: a preprocessing unit, used to obtain historical scene analysis data and call connection data, preprocess the data and mark the outliers therein to obtain preprocessed data; a training unit, used to train the isolation forest model using the preprocessed data to obtain an anomaly analysis model; an association unit, used to associate the scene analysis data with the call connection data to obtain associated data, and input the associated data into the anomaly analysis model to obtain anomaly data.

[0022] The third aspect of the present invention provides a device for managing the connection rate of small parcel operators, comprising a memory and at least one processor, wherein the memory stores computer-readable instructions; the at least one processor calls the computer-readable instructions in the memory to execute the various steps of the small parcel operator connection rate management method as described above.

[0023] A fourth aspect of the present invention provides a computer-readable storage medium having computer-readable instructions stored thereon. When the computer-readable instructions are executed by a processor, the various steps of the above-mentioned method for managing the connection rate of small mail operators are implemented.

[0024] Beneficial effect: The present invention provides a method for managing the connection rate of small parcel workers. The method first obtains the telephone connection data of the small parcel workers and analyzes the telephone connection data using multimodal data recognition technology, thereby obtaining comprehensive connection data from multiple angles; then, the comprehensive connection data is statistically analyzed according to analysis dimension rules to obtain multiple scenario analysis data, so that managers can fully and multi-dimensionally understand the data status; then, the scenario analysis data is analyzed using an abnormal analysis model to obtain abnormal data, which is convenient for managers to discover abnormal situations; finally, the correlation between the abnormal data and each feature data in the telephone connection data is calculated using a correlation analysis method to obtain influencing factors, and the scenario analysis data, abnormal data and influencing factors are visualized and output, so that managers can intuitively see the data of each dimension and scenario, and at the same time, conveniently know the abnormal data and possible influencing factors, which is convenient for quickly taking measures to make improvements and improve customer satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A first flow chart of a method for managing the connection rate of small mail handlers provided in an embodiment of the present invention;

[0026] Figure 2 A second flow chart of the method for managing the connection rate of small mail handlers provided in an embodiment of the present invention;

[0027] Figure 3 A third flow chart of the method for managing the connection rate of small mail handlers provided in an embodiment of the present invention;

[0028] Figure 4 A fourth flow chart of the method for managing the connection rate of small mail handlers provided in an embodiment of the present invention;

[0029] Figure 5 A fifth flow chart of the method for managing the connection rate of small mail handlers provided in an embodiment of the present invention;

[0030] Figure 6 A sixth flow chart of the method for managing the connection rate of small mail handlers provided in an embodiment of the present invention;

[0031] Figure 7 A seventh flow chart of the method for managing the connection rate of small mail handlers provided in an embodiment of the present invention;

[0032] Figure 8 A schematic diagram of the structure of a device for managing the connection rate of small mail carriers provided by an embodiment of the present invention;

[0033] Figure 9 Another structural diagram of the device for managing the connection rate of small mail carriers provided by an embodiment of the present invention;

[0034] Figure 10 A schematic diagram of the structure of a device for managing the connection rate of small mail carriers provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0035] The present invention provides a method, device, equipment, and storage medium for managing the connection rate of small parcel service providers. The present invention first collects the call connection data of small parcel service providers and analyzes it using multimodal data recognition technology to obtain comprehensive, multi-angle connection data. This step ensures the richness and accuracy of the data, providing a solid foundation for subsequent analysis. Next, this comprehensive connection data is statistically analyzed and analyzed in multiple dimensions based on analysis dimension rules to generate multiple scenario analysis data. This allows managers to comprehensively and meticulously grasp the data status and better understand the actual situation of the small parcel service provider's connection rate. Subsequently, an anomaly analysis model is used to conduct in-depth analysis of the scenario analysis data, accurately identifying anomalies, helping managers quickly identify potential problems. Finally, a correlation analysis method is used to calculate the correlation between the anomaly data and various feature data in the call connection data to determine the key factors affecting the connection rate. The scenario analysis data, anomaly data, and influencing factors are then visualized and output. This allows managers to intuitively view data from various dimensions and scenarios, easily obtain anomaly data and its potential influencing factors, and quickly formulate and implement improvement measures, effectively improving customer satisfaction.

[0036] The terms "first," "second," "third," "fourth," and the like (if any) in the description and claims of the present invention and in the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or precedence. It should be understood that the terms used in this manner are interchangeable 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 apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to these processes, methods, products, or apparatus.

[0037] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the method for managing the connection rate of small mail operators in the embodiment of the present invention includes:

[0038] S101 obtains the telephone connection data of the small parcel member and analyzes the telephone connection data using multimodal data recognition technology to obtain comprehensive connection data;

[0039] S102. Setting analysis dimension rules, performing statistics and analysis on the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data;

[0040] S103. Analyze the scene analysis data using an anomaly analysis model to obtain abnormal data;

[0041] S104. Use a correlation analysis method to calculate the correlation between the abnormal data and each feature data in the call connection data to obtain the influencing factors, associate the influencing factors with the abnormal data, and visualize the scenario analysis data, abnormal data and influencing factors and output them.

[0042] In this embodiment, the system's automatic triggering mechanism can be used to collect the phone connection data of the courier, including call volume, connection volume, unconnected volume, connection rate and other data. The scenarios involve full scenarios, active call testing, customer callbacks, etc., and multimodal data recognition technology is used to ensure the accuracy and completeness of the data. Multimodal data recognition technology is a technology that can process and analyze multiple forms of data (such as text, images, audio, video, etc.). By integrating data from multiple modes, it can more comprehensively and accurately understand the information and meaning behind the data. In this embodiment, multimodal data recognition technology analyzes the phone connection data. It can extract and integrate information from different angles and levels, thereby obtaining comprehensive connection data from multiple angles. The application of this technology ensures the richness and accuracy of the data, and provides a solid foundation for subsequent statistics, analysis and anomaly detection.

[0043] By setting analysis dimension rules and applying them to comprehensive connection data, we can generate multi-dimensional statistics and analysis data for multiple scenarios. This step aims to examine and understand connection data from different perspectives and scenarios, uncovering more detailed and targeted information. Multi-dimensional statistics and analysis allow managers to gain a comprehensive and detailed understanding of the connection rate of small-service personnel, gaining insight into problems from various scenarios, helping to identify potential patterns and problem areas, and providing richer and more specific evidence for subsequent management and decision-making.

[0044] Anomaly analysis models are used to conduct in-depth analysis of scenario analysis data to identify anomalies. Based on specific algorithms and rules, these models accurately filter out data points that do not conform to normal patterns. These anomalies may indicate problems or special circumstances during the call connection process. Anomaly analysis helps managers quickly and accurately identify anomalies in call connection data, promptly identifying potential problems and providing clear guidance for subsequent targeted problem resolution and optimizing call connection rates.

[0045] Finally, through correlation analysis, we can clearly identify the factors that influence abnormal connection rates, allowing managers to clearly understand the root causes of the problem. Visualization processing makes complex data intuitive and easy to understand. Managers can easily view data across various dimensions and scenarios, and conveniently obtain abnormal data and its influencing factors, allowing them to quickly formulate and implement improvement measures, effectively improving customer satisfaction.

[0046] By collecting the phone connection data of small parcel workers in real time, multi-dimensional analysis and intelligent management, statistical analysis is conducted according to different scenarios (such as active call testing and customer call back) to check the phone accessibility of small parcel workers. This can be used for management such as small parcel workers' connection rate assessment, thereby improving the phone connection rate of small parcel workers, reducing customer complaints, and improving the quality of terminal services.

[0047] See also Figure 2 The second embodiment of the method for managing the connection rate of small mail carriers in the embodiment of the present invention includes:

[0048] S201 obtains the telephone connection data of the small parcel member, the telephone connection data includes active call test data and customer call back data, extracts the call recording data in the telephone connection data;

[0049] S202 uses speech-to-text technology to obtain call text records based on call recording data;

[0050] S203. Evaluate customer sentiment and courier service based on call text records and call recording data to obtain customer sentiment scores and courier communication scores;

[0051] S204. Integrate the customer sentiment score and the courier communication score into the call connection data to obtain comprehensive call connection data.

[0052] In active call testing scenarios, if a customer inquires or complains about a package that "shows receipt" but has not yet been received, the system automatically triggers a call test to the salesperson who signed for the package. The system calls the courier number displayed on the official website to verify that their phone is accessible. In customer callback scenarios, the courier will call the customer when delivering the package, during pre-delivery calls, or in other situations. If the customer's call is not connected or in other circumstances, the customer needs to contact the salesperson again. In addition to the above scenarios, other types of connection data can also be included. Call recording data is a key reference factor in evaluating the service quality of courier services, and therefore requires in-depth analysis.

[0053] Using speech-to-text technology, call recordings are converted into text records. This converts audio information into text that is easier to store, retrieve, and analyze, making it easier to process the call content and quickly analyze large numbers of calls. It also facilitates integration with other text analysis tools to uncover more valuable information.

[0054] Based on the call transcripts and original call recordings, customer sentiment and courier service are evaluated, generating customer sentiment scores and courier communication scores, respectively. This combined evaluation of call transcripts and recordings provides a more comprehensive and accurate reflection of customer sentiment and courier service. The customer sentiment score helps understand customer satisfaction and emotional reactions during the call handling process, while the courier communication score provides a direct reflection of the courier's communication performance, providing a basis for subsequent personnel training and management.

[0055] Finally, the resulting customer sentiment scores and courier communication scores are integrated into the existing call connection data to create richer, comprehensive connection data. This incorporates call quality evaluation information into the basic connection data, providing more comprehensive data support for subsequent analysis and decision-making.

[0056] See also Figure 3 The third embodiment of the method for managing the connection rate of small mail carriers in the embodiment of the present invention includes:

[0057] S301 obtains call recording data, uses the speech recognition model to identify the language type in the call recording data, and obtains the language type data;

[0058] S302. According to the language data, using the corresponding speech-to-text technology, the call recording data is converted into a call text record;

[0059] S303. Use a deep learning-based correction model combined with language type data to correct the call text record.

[0060] In this embodiment, the speech recognition model can automatically and quickly identify the language type in the call recording by analyzing the characteristics of the speech signal and comparing it with the known language model, providing a basis for selecting appropriate technologies and models for subsequent speech-to-text processing, ensuring the accuracy and efficiency of transcription.

[0061] Next, based on the identified language data, the corresponding speech-to-text technology is selected to convert the call recording data into text. Different languages can have different speech-to-text models and algorithms. This step ensures the accuracy and professionalism of the transcription results. By using speech-to-text technology that matches the language, call recordings can be converted into text more accurately, improving transcription quality. It also supports processing in multiple languages and adapts to business needs in different language environments.

[0062] Finally, a deep learning-based correction model, combined with previously acquired language data, is used to correct the converted call transcripts. By learning and training from large amounts of text data, the deep learning correction model can identify and correct errors in the transcripts, optimize sentence structure, and more accurately align the transcripts with actual call content and language usage, providing a more reliable data foundation for subsequent transcript-based analysis and evaluation.

[0063] See also Figure 4 The fourth embodiment of the method for managing the connection rate of small mail carriers in the embodiment of the present invention includes:

[0064] S401. Obtain language texts from multiple regions and countries, use the language texts from multiple regions and countries as training sets, and train the Transformer model to obtain a correction model;

[0065] S402 uses a correction model to correct the call text record according to the language data to obtain the corrected text;

[0066] S403. Translate the corrected text into the official language version to obtain a translated text, and associate the translated text, the corrected text, and the call recording data.

[0067] The Transformer model is a deep learning-based architecture that, through training on large amounts of text data, can learn language features and patterns, enabling it to be used for tasks such as text correction and generation. By training on text from multiple regions and countries, the correction model adapts to the characteristics of different languages and dialects, improving its versatility and accuracy, and providing strong language understanding capabilities for subsequent call transcript correction. The resulting correction model, combined with language data, corrects call transcripts to produce more accurate and linguistically standardized corrections. Based on the input language data, the correction model specifically addresses grammatical errors, inappropriate word usage, and incoherent sentences in the transcripts. The generation of translated text allows call transcripts in different languages to be unified into a single language, facilitating cross-lingual communication and understanding, and enabling centralized management and analysis of multilingual call data. Linking translated text, corrected text, and call recording data creates a complete data chain, facilitating further insight into call data, helping to fully understand call situations and improve service quality.

[0068] See also Figure 5 The fifth embodiment of the method for managing the connection rate of small mail carriers in the embodiment of the present invention includes:

[0069] S501 uses blind source separation technology to separate the customer voice and small message staff voice in the call recording data to obtain customer call recordings and small message recordings;

[0070] S502. Using sentiment analysis technology, evaluate customer sentiment and courier service based on customer call recordings and courier call recordings to obtain the customer's first score and the courier's first score;

[0071] S503 extracts keywords from the call text record, evaluates customer sentiment and parcel service based on keywords, and obtains the second customer score and the second parcel score of the parcel staff;

[0072] S504. Summarize the customer's first score, the courier's first score, the customer's second score, and the courier's second score according to preset rules to obtain the customer's emotion score and the courier's communication score.

[0073] Blind source separation technology, even when the source is unknown, can separate different speech sources by analyzing the characteristics of the voice signal. This allows for subsequent analysis and processing of both the customer and the courier's speech, improving accuracy and targeting, preventing interference between the two voices, and enabling more precise assessments of customer sentiment and courier service. Furthermore, sentiment analysis technology can be used to evaluate the customer's emotional state and the courier's service performance based on recorded customer and courier calls, respectively, to generate a top-rated customer and courier rating. Sentiment analysis analyzes voice characteristics such as intonation, speech rate, and pauses, and, combined with methods such as sentiment dictionaries, determines the emotions and attitudes embedded in the speech. This provides intuitive sentiment scores, helping managers quickly understand the emotions and performance of both parties on the call.

[0074] Keywords are extracted from call transcripts, and then used to evaluate customer sentiment and courier service, resulting in a secondary customer and courier rating. Keywords reflect key information and themes from a call, and keyword analysis reveals the key content and emotional tendencies of the call. Text-based analysis, combined with the semantic information of keywords, provides a more comprehensive understanding of the call content, supplementing the evaluation of customer sentiment and courier service. This, in turn, complements the audio-based evaluation, enhancing the comprehensiveness and accuracy of the evaluation.

[0075] Finally, according to pre-set rules, the customer's first score, the courier's first score, the customer's second score, and the courier's second score are aggregated to produce a customer sentiment score and a courier communication score. Pre-set rules can range from simple averaging and weighted summation to more complex comprehensive evaluation models. Integrating the evaluation results from both voice and text takes into account multiple factors, resulting in a more comprehensive and objective customer sentiment score and courier communication score.

[0076] See also Figure 6 The sixth embodiment of the method for managing the connection rate of small mail carriers in the embodiment of the present invention includes:

[0077] S601. Set analysis dimension rules, the analysis dimensions include time period, region, call type;

[0078] S602. For the time period dimension, the call connection data is split and analyzed separately according to the time period, and the peak call time period is marked;

[0079] S603. For the regional dimension, separate data analysis is conducted at the group, region, business province, distribution, grid, and branch levels;

[0080] S604. For call types, analyze them separately according to the types of active calls and customer calls.

[0081] In this embodiment, the system supports separate data analysis by multiple dimensions such as group, region, business province, distribution, grid, and outlets. The call details are subdivided to display the time, name, mobile phone number, status, etc. of the small parcel staff, and generate a report on the proportion of reasons for non-connection. Specifically, the multi-dimensional scenarios are as follows: data is summarized by group, region, business province, distribution, and grid dimensions. The table summarizes each scenario by business province and displays data analysis for each scenario (such as full scenario, active call test, customer call, industrial mobile phone, proportion of reasons for active call test non-connection, etc.). The statistical data of each scenario can also be linked to the call details, showing the time, name, mobile phone number, status, etc. in sequence.

[0082] Specifically, within the time period dimension, call connection data is segmented into different time periods, and the data for each time period is analyzed separately. Time periods with unique characteristics, such as high call volume or low connection rates, are identified and labeled. This helps uncover patterns and characteristics in call connection data across different time periods, such as periods with concentrated customer inquiries or periods of high workload for small parcel handlers, providing a basis for developing flexible work strategies. Based on in-depth analysis of data from different time periods, managers can develop strategies that better suit actual operations, such as adjusting work hours and optimizing shift schedules, to improve overall operational efficiency and service quality.

[0083] See also Figure 7 The seventh embodiment of the method for managing the connection rate of small mail carriers in the embodiment of the present invention includes:

[0084] S701. Obtain historical scene analysis data and telephone connection data, preprocess the data and mark the outliers therein to obtain preprocessed data;

[0085] S702. Using preprocessed data to train the isolation forest model to obtain an anomaly analysis model;

[0086] S703. Associate the scenario analysis data with the call connection data to obtain associated data, and input the associated data into an anomaly analysis model to obtain anomaly data.

[0087] In this example, historical scenario analysis data and call connection data are first collected and preprocessed, including cleaning and standardization. Outliers are also identified and marked, ultimately generating preprocessed data. The goal of preprocessing is to improve data quality and usability. Preprocessing can remove noise, fill missing values, and correct erroneous data, making the data more accurate and reliable.

[0088] The preprocessed data is then used to train the Isolation Forest model, resulting in a model capable of detecting anomalies, known as the Anomaly Analysis Model. The Isolation Forest Model is an anomaly detection algorithm based on unsupervised learning. By learning the distribution characteristics of the data, it can identify anomalous data points that differ from normal data patterns. This automated anomaly detection method can automatically and quickly identify anomalous data from large amounts of data, reducing the workload and time required for manual screening.

[0089] The scenario analysis data is then correlated and integrated with the call connection data to form correlated data. This correlated data is then fed into a previously trained anomaly analysis model, which identifies anomalous data based on learned features and patterns. By correlating scenario analysis data with call connection data, comprehensive analysis from multiple dimensions is possible, improving the comprehensiveness and accuracy of anomaly detection.

[0090] The above describes the method for managing the connection rate of small parcel workers in the embodiment of the present invention. The following describes the device for managing the connection rate of small parcel workers in the embodiment of the present invention. Figure 8 In one embodiment of the present invention, a device for managing the connection rate of small mail carriers includes:

[0091] The multimodal analysis module 10 is used to obtain the call connection data of the courier and analyze the call connection data using multimodal data recognition technology to obtain comprehensive connection data;

[0092] A statistical analysis module 20 is configured to set analysis dimension rules and perform statistics and analysis on the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data;

[0093] Anomaly analysis module 30, configured to analyze the scene analysis data using an anomaly analysis model to obtain anomaly data;

[0094] The correlation analysis module 40 is used to calculate the correlation between the abnormal data and each feature data in the call connection data using a correlation analysis method to obtain the influencing factors, associate the influencing factors with the abnormal data, and visualize the scenario analysis data, abnormal data and influencing factors and output them.

[0095] See also Figure 9 In one embodiment of the present invention, a device for managing the connection rate of small mail carriers includes:

[0096] The multimodal analysis module 10 is used to obtain the call connection data of the courier and analyze the call connection data using multimodal data recognition technology to obtain comprehensive connection data;

[0097] A statistical analysis module 20 is configured to set analysis dimension rules and perform statistics and analysis on the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data;

[0098] Anomaly analysis module 30, configured to analyze the scene analysis data using an anomaly analysis model to obtain anomaly data;

[0099] A correlation analysis module 40 is configured to calculate the correlation between the abnormal data and each feature data in the call connection data using a correlation analysis method to obtain influencing factors, associate the influencing factors with the abnormal data, and visualize and output the scenario analysis data, abnormal data, and influencing factors;

[0100] In this embodiment, the multimodal analysis module 10 includes:

[0101] Extraction submodule 11, used to obtain the courier's phone connection data, the phone connection data includes active call test data and customer call back data, and extract call recording data from the phone connection data;

[0102] The conversion submodule 12 is used to obtain a call text record based on the call recording data using speech-to-text technology;

[0103] The scoring submodule 13 is used to evaluate the customer's emotions and the courier's service based on the call text records and call recording data, and obtain the customer's emotion score and the courier's communication score;

[0104] An integration submodule 14 is used to integrate the customer sentiment score and the courier communication score into the call connection data to obtain comprehensive call connection data;

[0105] In this embodiment, the conversion submodule 12 includes:

[0106] The recognition unit 121 is used to obtain call recording data and identify the language type in the call recording data using a speech recognition model to obtain language type data;

[0107] The conversion unit 122 is used to convert the call recording data into a call text record using a corresponding speech-to-text technology according to the language type data;

[0108] a correction unit 123 for correcting the call text record using a correction model based on deep learning and in combination with the language type data;

[0109] In this embodiment, the correction unit 123 includes:

[0110] Constructing subunit 1231 for obtaining language texts from multiple regions and countries, and training a Transformer model using the language texts from multiple regions and countries as a training set to obtain a correction model;

[0111] The correction subunit 1232 is configured to correct the call text record according to the language type data using a correction model to obtain a corrected text;

[0112] The translation subunit 1233 is used to translate the corrected text into the official language version, obtain the translated text, and associate the translated text, the corrected text, and the call recording data;

[0113] In this embodiment, the scoring submodule 13 includes:

[0114] The separation unit 131 is used to separate the customer voice and the courier voice in the call recording data using blind source separation technology to obtain the customer call recording and the courier call recording;

[0115] The first evaluation unit 132 is configured to use sentiment analysis technology to evaluate the customer's emotions and the courier's service based on the customer's call recording and the courier's call recording, respectively, to obtain the customer's first score and the courier's first score;

[0116] The second evaluation unit 133 is configured to extract keywords from the call text record, evaluate the customer's emotions and the courier's service based on the keywords, and obtain a second customer rating and a second courier rating;

[0117] A summary calculation unit 134 is used to summarize the customer's first score, the courier's first score, the customer's second score, and the courier's second score according to a preset rule to obtain a customer emotion score and a courier's communication score;

[0118] In this embodiment, the statistical analysis module 20 includes:

[0119] A setting unit 21 is used to set analysis dimension rules, where the analysis dimensions include time period, region, and call type;

[0120] The first analysis unit 22 is used to split and analyze the call connection data according to the time period dimension, and mark the time period with peak call volume;

[0121] The second analysis unit 23 is used to perform separate data analysis on the regional dimension at the levels of group, region, business province, distribution, grid and outlet;

[0122] The third analysis unit 24 is used to analyze the call type according to the active call test and the customer call type;

[0123] In this embodiment, the abnormality analysis module 30 includes:

[0124] A preprocessing unit 31 is used to obtain historical scene analysis data and call connection data, preprocess the data and mark abnormal values therein to obtain preprocessed data;

[0125] A training unit 32 is used to train the isolation forest model using the preprocessed data to obtain an anomaly analysis model;

[0126] The correlation unit 33 is used to correlate the scene analysis data with the call connection data to obtain correlation data, and input the correlation data into the abnormality analysis model to obtain abnormal data.

[0127] The parcel handler connection rate management device provided by the present invention integrates multi-source data such as parcel handlers' call records, work order status and geographic location through multimodal data recognition technology to construct a multi-dimensional data portrait to ensure the comprehensiveness and accuracy of the analysis; based on preset multi-dimensional analysis rules such as region, time period, customer type, etc., it automatically generates refined scenario insight reports to help managers have a penetrating grasp of service performance in different business scenarios; combines machine learning models to intelligently detect abnormal fluctuations, quickly lock abnormal nodes that deviate from the threshold, and use attribution analysis algorithms to explore the correlation patterns between features such as call frequency, response time, customer tags, etc., to accurately locate the root cause of the problem; finally, through a dynamic visual dashboard, it presents multi-dimensional data comparison, abnormal warning heat map and influencing factor weight distribution, so that managers can see through service shortcomings in real time, optimize resource allocation in a targeted manner (such as adjusting manpower deployment during peak hours), and achieve a closed-loop improvement in service quality.

[0128] The above is a detailed description of the small parcel service connection rate management device in the embodiment of the present invention from the perspective of modular functional entities. The following is a detailed description of the small parcel service connection rate management device in the embodiment of the present invention from the perspective of hardware processing.

[0129] Figure 10A schematic structural diagram of a parcel service connection rate management device provided in an embodiment of the present invention, wherein the parcel service connection rate management device 900 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 910 (for example, one or more processors) and a memory 920, and one or more storage media 930 (for example, one or more mass storage devices) storing application programs 933 or data 932. Among them, the memory 920 and the storage medium 930 may be temporary storage or permanent storage. The program stored in the storage medium 930 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the parcel service connection rate management device 900. Furthermore, the processor 910 may be configured to communicate with the storage medium 930, and execute a series of instruction operations in the storage medium 930 on the parcel service connection rate management device 900 to implement the steps of the parcel service connection rate management method provided in the above-mentioned method embodiments.

[0130] The small mailman connection rate management device 900 may also include one or more power supplies 940, one or more wired or wireless network interfaces 950, one or more input and output interfaces 960, and / or one or more operating systems 931, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 10 The structure of the device for managing the connection rate of small parcel carriers shown does not constitute a limitation on the device for managing the connection rate of small parcel carriers, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0131] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the method for managing the connection rate of small message operators.

[0132] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working process of the above-described equipment or device can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0133] 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 and includes several instructions for enabling a computer device (which can be a personal computer, server, or 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), magnetic disk or optical disk, etc., various media that can store program code.

[0134] It is understandable that those skilled in the art can make equivalent substitutions or changes based on the technical solution and inventive concept of the present invention, and all these changes or substitutions should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for managing the connection rate of small-sized mail carriers, characterized in that: The steps include: Obtain the call connection data of small courier and analyze it using multimodal data recognition technology to obtain comprehensive call connection data; Setting analysis dimension rules, and performing statistics and analysis on the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data; Analyze the scenario analysis data using an anomaly analysis model to obtain abnormal data; A correlation analysis method is used to calculate the correlation between the abnormal data and each feature data in the call connection data to obtain the influencing factors, and the influencing factors are associated with the abnormal data. The scenario analysis data, abnormal data and influencing factors are visualized and output.

2. The method for managing the connection rate of small parcel handlers according to claim 1, characterized in that: The method of obtaining the call connection data of the courier and analyzing the call connection data using multimodal data recognition technology to obtain comprehensive connection data specifically includes: Obtain the courier's call connection data, which includes active call test data and customer call-back data, and extract the call recording data from the call connection data; Use speech-to-text technology to obtain call text records based on call recording data; Evaluate customer sentiment and courier service based on call transcripts and recordings to obtain customer sentiment scores and courier communication scores; Integrate customer sentiment scores and courier communication scores into call connection data to obtain comprehensive call connection data.

3. The method for managing the connection rate of small parcel handlers according to claim 2, characterized in that: The speech-to-text technology is used to obtain call text records based on call recording data, specifically including: Obtain call recording data, use a speech recognition model to identify the language type in the call recording data, and obtain language type data; Based on the language data, corresponding speech-to-text technology is used to convert the call recording data into call text records; A deep learning-based correction model is used to correct call transcripts in combination with language data.

4. The method for managing the connection rate of small parcel handlers according to claim 3, characterized in that: The correction model based on deep learning is used to correct the call text records, specifically including: Obtain language texts from multiple regions and countries, use these language texts from multiple regions and countries as training sets, and train the Transformer model to obtain a correction model. Using a correction model, the call transcript is corrected according to the language data to obtain a corrected text; The corrected text is translated into an official language version to obtain a translated text, and the translated text, the corrected text, and the call recording data are associated.

5. The method for managing the connection rate of small parcel handlers according to claim 2, characterized in that: The customer sentiment and courier service are evaluated based on the call text records and call recording data to obtain the customer sentiment score and courier communication score, specifically including: Adopt blind source separation technology to separate the customer voice and the courier voice in the call recording data to obtain the customer call recording and the courier call recording; Using sentiment analysis technology, we evaluate customer sentiment and courier service based on recordings of both customer and courier calls, obtaining the top customer and courier ratings. Extract keywords from the call transcript, evaluate the customer's sentiment and the courier's service based on the keywords, and obtain the customer's second rating and the courier's second rating; The customer's first score, the courier's first score, the customer's second score, and the courier's second score are summarized according to preset rules to obtain the customer's emotion score and the courier's communication score.

6. The method for managing the connection rate of small parcel handlers according to claim 1, characterized in that: The setting of analysis dimension rules and performing statistics and analysis on the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data specifically includes: Set analysis dimension rules, including time period, region, and call type; For the time period dimension, the call connection data is split and analyzed separately according to the time period, and the time period with peak call volume is marked; For the regional dimension, separate data analysis is conducted at the group, region, business province, distribution, grid and outlet levels; For call types, analysis is conducted separately by active call test and customer call types.

7. The method for managing the connection rate of small parcel handlers according to claim 1, characterized in that: The use of the abnormal analysis model to analyze the scene analysis data to obtain abnormal data specifically includes: Obtain historical scenario analysis data and call connection data, preprocess the data and mark outliers to obtain preprocessed data; The isolation forest model is trained using preprocessed data to obtain an anomaly analysis model; The scenario analysis data is associated with the call connection data to obtain associated data, and the associated data is input into the anomaly analysis model to obtain anomaly data.

8. A device for managing the connection rate of small parcel handlers, characterized in that: include: The multimodal analysis module is used to obtain the call connection data of the small courier and analyze the call connection data using multimodal data recognition technology to obtain comprehensive connection data; A statistical analysis module is used to set analysis dimension rules and perform statistics and analysis on the comprehensive connection data according to the analysis dimension rules to obtain multiple scenario analysis data; Anomaly analysis module, used to analyze scene analysis data using an anomaly analysis model to obtain abnormal data; The correlation analysis module is used to calculate the correlation between abnormal data and each feature data in the call connection data using a correlation analysis method to obtain influencing factors, associate the influencing factors with the abnormal data, and visualize the scenario analysis data, abnormal data and influencing factors before output.

9. A device for managing the connection rate of small parcel handlers, characterized in that: comprising a memory and at least one processor, wherein the memory has computer-readable instructions stored therein; The at least one processor calls the computer-readable instructions in the memory to execute the various steps of the method for managing the connection rate of small mail operators as described in any one of claims 1-7.

10. A computer-readable storage medium having computer-readable instructions stored thereon, characterized in that: When the computer-readable instructions are executed by a processor, the various steps of the method for managing the connection rate of small mail carriers as described in any one of claims 1 to 7 are implemented.

Images