IVR telephone traffic full-link analysis method and system

Abstract

The invention discloses an IVR telephone traffic full-link analysis method and system, and relates to the technical field of interactive voice response systems, and the method comprises the steps: collecting telephone traffic data, and carrying out the preprocessing; performing first user behavior pattern recognition on the collected data; and based on the user behavior analysis result, first service demand prediction is established, system resources are dynamically adjusted, and user experience and service efficiency are improved. According to the method, the telephone traffic data is collected by setting the data burying points, comprehensive and accurate capturing of interaction data between the IVR system and the users is achieved, high quality of the data is ensured through preprocessing, accurate division of user groups and prediction of service requirements are achieved through first user behavior pattern recognition, and the user experience is improved. By establishing the first service demand to predict and dynamically adjust the system resources, the user behavior and the system operation state can be monitored in real time, the demands of different user groups are met through personalized resource allocation, and the service efficiency, the flexibility and the user satisfaction are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of interactive voice response systems, and in particular to an IVR call full-link analysis method and system. Background Art

[0002] With the rapid development of information technology and the advancement of speech recognition technology, IVR interactive voice response systems have gradually become an important part of customer service, automatic information processing and system interfaces, providing users with convenient self-service. Traditional IVR systems mainly focus on voice navigation and simple interactions, such as selecting menus through buttons or voice commands. However, with the continuous growth of user needs and the advancement of technology, existing IVR systems have obvious deficiencies in user behavior analysis, service demand prediction and system performance optimization.

[0003] Most existing IVR systems use simple data recording and statistical methods, such as call duration, button selection, etc., which make it difficult to deeply analyze the voice content, button behavior, emotional changes and other data during the user's call, resulting in inaccurate user behavior identification and difficulty in accurately judging user intentions and needs, and thus unable to effectively predict users' future service needs, resulting in deviations in the company's resource allocation and service strategy formulation, and failure to meet user expectations. In addition, existing IVR systems find it difficult to monitor and analyze various data during calls in real time, and it is difficult to promptly discover and resolve performance bottlenecks or fault points, affecting system stability and reliability. Due to the above problems, the existing IVR system still needs to be improved in terms of user experience and service efficiency. Users may need to spend more time and energy to complete business processing, and companies may also lose customer trust due to their inability to respond and handle user needs in a timely manner.

[0004] Therefore, these existing technologies have failed to improve user experience and service efficiency as expected. Against the background of growing service demand, there is an urgent need for a new IVR call full-link analysis method to achieve more efficient collection and processing of user interaction data, thereby optimizing service configuration, realizing dynamic adjustment of resources, and improving user satisfaction and corporate profits. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is: the existing IVR call full-link analysis method is unable to fully capture user interaction behavior, user behavior identification is inaccurate, service demand prediction is inaccurate, and how to achieve deep insight into user behavior and real-time optimization.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: a full-link analysis method for IVR traffic, including collecting traffic data and performing preprocessing; performing first user behavior pattern recognition on the collected data; based on the user behavior analysis results, establishing a first service demand forecast and dynamically adjusting system resources to improve user experience and service efficiency.

[0008] As a preferred solution of the IVR traffic full-link analysis method described in the present invention, wherein: the collection of traffic data includes collecting traffic data by setting data embedding points.

[0009] As a preferred solution of the IVR traffic full-link analysis method described in the present invention, the collection of traffic data also includes real-time capture of interaction data between users and the IVR system.

[0010] As a preferred solution of the IVR traffic full-link analysis method described in the present invention, wherein: the first user behavior pattern recognition includes performing a traffic full-link analysis on the pre-processed data.

[0011] As a preferred solution of the IVR call full-link analysis method described in the present invention, wherein: the first user behavior pattern identification also includes identifying user behavior patterns and service demand trends.

[0012] As a preferred solution of the IVR call full-link analysis method described in the present invention, wherein: the establishment of the first service demand forecast includes establishing the first service demand forecast based on historical data and full-link analysis results, predicting future service volume and service type, and real-time monitoring of user behavior and system operation status.

[0013] As a preferred solution of the IVR call full-link analysis method described in the present invention, the dynamic adjustment of system resources includes combining key performance indicators with user behavior, triggering set thresholds and judgment rules based on the first service demand prediction result; when the system detects that the user satisfaction is lower than the set threshold, the voice prompts and interaction processes of the IVR are automatically adjusted; when the user's call duration exceeds twice the average call duration, the user is judged as a long call user, and service resources are allocated to the user; when it is predicted that the service demand will increase by more than a set ratio, the resource expansion mechanism is triggered to increase service resources; when the system performance indicator is lower than the set threshold, the performance optimization mechanism is triggered to optimize and adjust the system.

[0014] Another object of the present invention is to provide an IVR call full-link analysis system, which can establish a first service demand prediction and dynamically adjust system resources based on the user behavior analysis results to improve user experience and service efficiency. It solves the problem that the current IVR call full-link analysis technology lacks in-depth analysis of user behavior data, cannot effectively predict users' future service needs, and cannot dynamically adjust system resources according to the prediction results, resulting in unreasonable resource allocation and low service efficiency.

[0015] As a preferred solution of the IVR call full-link analysis system described in the present invention, it includes: a data processing module, a user behavior analysis module, and a resource dynamic adjustment module; the data processing module is used to collect call data in real time, capture the interaction data between the user and the IVR system by setting data embedding points, and pre-process the collected data; the user behavior analysis module is used to perform call full-link analysis on the pre-processed data, identify user behavior patterns and service demand trends; the resource dynamic adjustment module is used to establish a first service demand forecast and dynamically adjust system resources based on the user behavior analysis results, so as to improve user experience and service efficiency.

[0016] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement a step of an IVR call full-link analysis method.

[0017] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of an IVR call full-link analysis method.

[0018] Beneficial effects of the present invention: The IVR traffic full-link analysis method provided by the present invention collects traffic data by setting data embedding points, realizes comprehensive and accurate capture of IVR system and user interaction data, ensures high data quality through preprocessing, improves the accuracy and efficiency of subsequent analysis, reduces analysis errors caused by data quality problems, and lays a solid data foundation for user behavior pattern recognition and service demand prediction. By deeply analyzing the preprocessed data, the user's behavior pattern and service demand trend are identified, revealing the deep-level law of user interaction, providing a basis for service optimization, and realizing accurate division of user groups and prediction of service needs through first user behavior pattern recognition, thereby facilitating the formulation of more personalized service strategies, and providing data support for resource allocation and service process improvement of the IVR system. By establishing a service demand prediction model and dynamically adjusting system resources, the IVR system can monitor user behavior and system operation status in real time, adjust resource configuration in time, realize on-demand allocation of IVR system resources, and improve resource utilization efficiency. Through real-time monitoring and automatic adjustment, the system's response speed and service quality are improved. Through personalized resource allocation, the needs of different user groups are met, and service efficiency, flexibility and user satisfaction are improved. The present invention achieves better results in terms of accuracy, flexibility and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0020] Figure 1 The present invention provides an overall flow chart of an IVR call full-link analysis method according to the first embodiment of the present invention.

[0021] Figure 2 A module schematic diagram of an IVR call full-link analysis system provided for the third embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0023] Example 1, reference Figure 1, is an embodiment of the present invention, and provides an IVR call full-link analysis method, comprising:

[0024] S1: Collect traffic data and perform pre-processing.

[0025] Furthermore, collecting call data includes collecting call data by setting data embedding points, wherein the set data embedding points include but are not limited to user incoming calls, button selections, voice recognition, information broadcasts, manual transfers, etc., and the collected call data types include but are not limited to call duration, number of calls, user behavior paths, voice recognition accuracy, user satisfaction, etc.

[0026] It should be noted that collecting traffic data also includes real-time capture of the interaction data between users and the IVR (Interactive Voice Response) system, and by using efficient data collection tools, the accuracy and real-time nature of the data can be ensured. The specific collection methods are as follows:

[0027] The IVR system collects voice data through voice acquisition technology and converts the user's voice input into text form, which usually involves steps such as voice signal acquisition, preprocessing, feature extraction and pattern matching. The preprocessing step may include noise reduction, filtering, endpoint detection, etc. to ensure the accuracy of subsequent processing, including but not limited to the use of advanced speech recognition algorithms and models, such as deep learning models (such as LSTM, Transformer, etc.), combined with a large amount of voice data for training and optimization to achieve high-precision speech recognition.

[0028] The IVR system collects key input data through key input collection technology. The IVR system usually provides a key menu for users to choose from. Users select the required services or information through key input. The system captures the key input data in real time and associates it with the user's call records for subsequent analysis. The specific collection method includes detecting the user's key input signal and capturing the user's key selection in real time. It usually involves the steps of collecting, decoding and storing key signals, including but not limited to the use of hardware-based signal detection technology, such as key circuit detection, signal amplification and decoding, to ensure the accuracy and real-time nature of key input.

[0029] The IVR system collects call duration data through call duration collection technology. The IVR system records the start time and end time of the user's call. The call duration data can be obtained by calculating the difference between the start time and the end time. The call duration data can be used to analyze indicators such as the duration distribution of user calls and the average call duration, thereby evaluating the service efficiency of the IVR system. The specific collection method includes recording the start time and end time of the user's call and calculating the call duration, which usually involves steps such as obtaining and calculating timestamps, including but not limited to using timestamp recording technology based on the system clock to ensure the accuracy and reliability of call duration.

[0030] The IVR system collects hang-up reason data through hang-up reason collection technology. When the user hangs up the phone, the IVR system will record the reason for hanging up. The reasons for hanging up may include the user actively hanging up after completing the business process, the user choosing to hang up after the system prompts, the user hanging up due to dissatisfaction or the service cannot meet the needs, etc. The hang-up reason data is of great significance for analyzing user satisfaction and optimizing service processes. The specific collection method includes detecting the user's hang-up signal, combining the system's internal logical judgment to determine the reason for hanging up, usually involving signal detection, logical judgment and storage steps, including but not limited to using technology based on signal detection and logical judgment, combined with user behavior patterns and system internal rules to achieve accurate judgment of the reason for hanging up.

[0031] It should also be noted that preprocessing includes cleaning, denoising, and formatting the collected raw data to ensure data quality. The purpose of data cleaning is to remove irrelevant information, duplicate data, erroneous data, etc. in the raw data to retain data that is useful for subsequent analysis. The specific preprocessing steps are as follows:

[0032] (1) Removal of irrelevant information: Identify and delete irrelevant information by writing scripts or using data cleaning tools (such as Excel, Python's Pandas library, etc.), using regular expression matching, conditional filtering, etc.

[0033] (2) Duplicate data removal: use data comparison algorithms to identify and delete duplicate data records, using hash algorithms, unique key constraints, etc.

[0034] (3) Error data correction: Identify and correct erroneous data based on business rules or data quality checking tools, using data verification algorithms, rule engines, etc.

[0035] (4) Signal denoising: For the collected signal data (such as audio, images, etc.), signal processing technology is used to denoise the data, such as filters (such as low-pass filters, high-pass filters, band-pass filters, etc.), Fourier transform, wavelet transform, etc.

[0036] (5) Text denoising: For text data, remove irrelevant punctuation, stop words, special characters, etc., regular expression matching, and natural language processing technology (such as word segmentation, part-of-speech tagging, etc.).

[0037] The purpose of data denoising is to reduce the noise in the original data and improve the accuracy and reliability of the data.

[0038] S2: Perform a first user behavior pattern recognition on the collected data.

[0039] Furthermore, the first user behavior pattern recognition includes performing a full-link traffic analysis on the pre-processed data to identify user behavior patterns and service demand trends.

[0040] It should be noted that the first user behavior pattern recognition may be performed by a variety of methods, such as cluster analysis and association rule mining or time series prediction and machine learning classification.

[0041] As a preferred embodiment, in this embodiment, the first user behavior pattern recognition adopts cluster analysis and association rule mining, and its principle steps are as follows:

[0042] First, cluster analysis is used to divide users into different behavior types according to their behavior data, such as high-frequency users, low-frequency users and long-call users. Specifically, the K-means algorithm is used with the K value set to 3. Based on the users' call duration (for example, users with an average call duration of more than 5 minutes are classified as long-call users) and call frequency (for example, users with more than 10 calls per month are classified as high-frequency users), groups with similar behavior patterns are identified. By analyzing groups with similar behavior patterns, it is found that, for example, long-call users are more inclined to make service calls during specific time periods (such as 6 p.m. to 9 p.m.), which suggests that operators can allocate more resources during these time periods to improve service efficiency and response speed.

[0043] At the same time, after cluster analysis, association rule mining technology is further used to analyze user behavior to identify service demand trends. First, the Apriori algorithm is used to conduct an in-depth analysis of the user's IVR selection and call history, and the support threshold is set to 0.05 and the confidence threshold is set to 0.7 to identify frequent item sets. The analysis results may show that 30% of users choose button 1 to query the bill after calling the customer service hotline, and after completing the query, 70% of users will choose button 2 to get more preferential information. This finding suggests that the operation team should put relevant options in the front in the IVR menu design to reduce the time users spend searching in the menu, thereby improving the user experience. In addition, with the help of association coefficient analysis, the operation team can Identify the growth trend of specific service requests. For example, if you find the rule that "when users choose to consult technical support, 63% of users will choose to apply for fault repair", after comparing with historical data, if the number of technical consultation calls increases by 150% in a certain week, then you can predict the increase in subsequent repair service requests. When the operation team identifies such a trend, they can coordinate technical support personnel and spare parts resources in advance to cope with the upcoming peak of service requests and ensure the timeliness and effectiveness of the service. Through association rule mining, the relationship between the identified user behavior patterns and service needs provides a basis for the adjustment of operation strategies, thereby achieving more accurate resource allocation and service optimization, and ultimately improving user satisfaction and corporate benefits.

[0044] In another feasible embodiment, the first user behavior pattern recognition adopts time series prediction and machine learning classification, and its principle steps are as follows:

[0045] First, a time series prediction model is used to analyze historical user behavior data. For example, the ARIMA model is used to model user call records and identify seasonal and cyclical changes in call demand. By analyzing historical data, it is found that the number of user calls in a specific time period (such as the last week of each month) usually increases by 20%. This pattern can effectively provide a reference for resource allocation and service preparation. Based on this observation, the operation team can dynamically adjust resources at the end of each month in advance to meet expected user needs and avoid increased user waiting time due to insufficient resources during peak periods.

[0046] At the same time, combined with machine learning classification technology, such as decision tree algorithm, users can be finely classified according to their historical call duration, call frequency, and selected IVR options. During the classification process, the model may identify that a certain type of user has a probability of up to 75% to consult technical support in the future. Based on this prediction, the operation team can formulate corresponding service strategies, such as proactively pushing relevant service information or optimizing voice processes before users call the hotline to meet user expectations and needs. In addition, the system can coordinate technical support and customer service resources in advance based on the prediction results to ensure service quality during high-demand periods for users. By combining time series prediction with machine learning classification, it can effectively identify the relationship between user behavior patterns and demand changes, providing strong support for the optimization of operation strategies and resource allocation, thereby improving user satisfaction and overall business operation efficiency.

[0047] S3: Based on the results of user behavior analysis, establish the first service demand forecast and dynamically adjust system resources to improve user experience and service efficiency.

[0048] Furthermore, establishing the first service demand forecast includes establishing the first service demand forecast based on historical data and full-link analysis results, predicting future service volume and service type, and real-time monitoring of user behavior and system operation status. The first service demand forecast can be a variety of methods, such as dynamic optimization algorithms based on machine learning or demand forecasting and dynamic configuration based on time series analysis.

[0049] As a preferred embodiment, in this embodiment, the first service demand prediction adopts a dynamic optimization algorithm based on machine learning. The dynamic optimization algorithm based on machine learning uses machine learning technology to deeply mine the analysis results and prediction results, and identifies the key performance indicators (KPIs) and service demand trends in the IVR system. Then, based on these indicators and trends, the algorithm automatically adjusts the resource allocation and service strategy of the IVR system to optimize service quality and efficiency. The principle steps are as follows:

[0050] (1) Data input: Input the first user behavior pattern recognition analysis results and prediction result data, including user behavior patterns, service demand trends, system performance indicators, etc.

[0051] (2) Feature extraction: Extract key features from the input data, such as user call duration, call frequency, question type, service satisfaction, etc., and use machine learning technology to perform feature selection and dimensionality reduction to reduce redundant information and improve algorithm efficiency.

[0052] (3) Use historical data and machine learning algorithms (such as decision trees, random forests, neural networks, etc.) to train prediction models. The goal of the model is to predict future service demand and system performance indicators based on current input features.

[0053] (4) Prediction and optimization: Use the trained model to predict service demand and system performance indicators in the future.

[0054] (5) Dynamically adjust the resource allocation and service strategy of the IVR system based on the prediction results. For example, when it is predicted that the demand for a certain type of service will increase, the corresponding service resources can be increased; when it is predicted that the system performance will decline, performance optimization can be carried out in advance.

[0055] Specifically, the dynamic optimization algorithm optimizes the resource allocation of the IVR system by deeply mining user behavior data and service demand trends. First, key features such as user call duration, call frequency, problem type and satisfaction are input from historical user behavior data. In the feature extraction process, machine learning algorithms (such as PCA) are used for feature selection and dimensionality reduction to extract the most significant indicators, reduce redundant information, and facilitate subsequent analysis. Then, in the model training stage, historical data sets are used to train decision trees or random forest models to predict future service needs. For example, the model may indicate that within a certain time period (such as 17:00 to 19:00 on Friday afternoon), users' demand for technical support consultation will increase by 30%. Based on this prediction, the operation team can dynamically adjust the resource allocation of the IVR system according to the service demand prediction results output by the model to achieve service optimization. For example, when the system predicts that the demand for technical support will increase, more customer service personnel can be arranged in advance to answer the calls, and the service process can be optimized to ensure the response speed and quality of the service during peak hours, thereby reducing user waiting time and improving user satisfaction.

[0056] In another feasible embodiment, the first service demand forecast adopts demand forecast and dynamic configuration based on time series analysis, and its principle steps are as follows:

[0057] By using time series analysis methods to analyze trends and seasonality of historical call records to predict future service needs, and using the ARIMA model to model monthly call volume, we can identify a trend that the call volume in the last week of each month usually increases by 20% compared with the same period last year. The operation team should use this information to plan resources in advance. For example, if it is predicted that there will be 1,500 call requests in the last week of the next month, the operation team can allocate customer service resources in advance and optimize the scheduling plan to meet the expected peak service demand. In addition, by combining machine learning classification algorithms (such as support vector machines) to segment users, the model analysis shows that the probability of a certain type of user (such as users who make more than 15 calls per month) consulting technical support services is 70%. Based on this in-depth analysis, the system can automatically push relevant service information before the user dials the hotline, and optimize the IVR menu to facilitate users to quickly obtain the required services. Through effective prediction and dynamic adjustment, the service response efficiency is improved, the user's waiting time is reduced, and the user experience is improved. The quantitative analysis in the implementation process is closely integrated with the scenario application, so that the prediction results are directly reflected in the real-time optimization of resource allocation and service strategies to achieve the best service effect.

[0058] It should be noted that real-time monitoring of user behavior and system operating status includes providing an intuitive visual interface to display analysis results and prediction results, and supporting real-time monitoring and decision-making support for managers. Managers can intuitively see the analysis results and prediction results through the visual interface, so as to make more accurate decisions. The operating status of the IVR system can be monitored in real time through dashboards and other means, such as call volume, average waiting time, satisfaction and other indicators. By comparing historical data, the operating trends and cyclical changes of the IVR system can be analyzed to provide a basis for future resource allocation and service strategies. According to the prediction results, future call volume and service needs can be predicted to prepare for early adjustment of resource allocation and service strategies.

[0059] It should also be noted that dynamic adjustment of system resources includes combining key performance indicators with user behavior, triggering set thresholds and judgment rules based on the first service demand prediction results; automatically adjusting the IVR voice prompts and interaction processes when the system detects that user satisfaction is lower than the set threshold; when the user's call duration exceeds twice the average call duration, the user is judged as a long call user, and additional service resources are allocated to the user; when it is predicted that the service demand will increase by more than a set proportion, the resource expansion mechanism is triggered to increase service resources; when the system performance indicator is lower than the set threshold, the performance optimization mechanism is triggered to optimize and adjust the system.

[0060] It should also be noted that in the process of dynamically adjusting system resources, based on the first service demand prediction results and real-time monitoring of user behavior and system operation status, a variety of personalized customization rules can be implemented to finely manage resource allocation, thereby improving user experience and service efficiency, including but not limited to, first, for user satisfaction monitoring, set specific threshold rules, for example, the full score of the satisfaction score is 5 points. When the satisfaction score is lower than 3.0, the system should immediately optimize the IVR voice prompts, simplify the user operation process, and reduce the user's selection time among multiple options. At the same time, such events are automatically recorded in the background for subsequent analysis. Secondly, for call duration monitoring, the system can set rules. When the user's call time exceeds twice the average time (such as longer than 12 minutes), it will automatically identify it as a long call user and allocate additional resources. For example, by automatically transferring to a higher-level technical support team, we can ensure that user problems are resolved quickly. In addition, when service demand is expected to increase by more than 25% (for example, during promotional activities or holidays), the system will automatically trigger the resource expansion mechanism to deploy more customer service personnel as needed, such as increasing service personnel by 30% based on historical data, and arranging pre-job training in advance to ensure that service quality is not affected. Finally, when system performance indicators (such as response time, success rate) are lower than the set threshold, the system will immediately trigger the performance optimization mechanism and perform various adjustments, such as optimizing the configuration of background services, improving database connection efficiency, or implementing load balancing strategies to improve response speed and system stability. Such customized solutions not only improve response speed, but also can accurately meet the needs of different user groups and achieve efficient utilization and optimal configuration of resources.

[0061] Embodiment 2 is an embodiment of the present invention, which provides an IVR call full-link analysis method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0062] First, we selected an IVR system in a certain area for the experiment and set up data points to capture user call data, including user incoming calls, button selections, voice recognition, information broadcasts, and manual transfers. We selected at least one week of data for collection to ensure the adequacy and representativeness of the data volume. We configured efficient data collection tools to ensure real-time and accuracy. We set up a voice recognition system based on a deep learning model and implemented signal denoising. We used low-pass filters and Fourier transform techniques to remove unnecessary background noise. At the same time, we retrieved historical call records to collect users’ call duration, number of calls, behavior paths, voice recognition accuracy, and satisfaction scores, with satisfaction scores ranging from 1 to 5 points. We used these data to form a preliminary data set.

[0063] Next, use Python's Pandas library to preprocess the data and clean the collected data, including removing irrelevant information, cleaning duplicate data, correcting erroneous data, etc., to ensure the accuracy of the data. Remove irrelevant characters through regular expression matching, use the hash algorithm to remove duplicates, and correct erroneous data through the rule engine. At the same time, for audio data, perform signal denoising through wavelet transform to ensure the accuracy of subsequent analysis.

[0064] Next, when analyzing the preprocessed data, the K-means clustering algorithm is used to divide users into different behavior types, such as high-frequency users, low-frequency users, and long-call users. The K value is set to 3. By calculating the call duration and frequency, similar behavior groups are identified and analyzed. At the same time, the Apriori algorithm is used to mine association rules to identify the connection between user behavior patterns and service needs. For example, 30% of users choose to obtain discount information after completing the bill query.

[0065] Finally, based on the results of user behavior analysis, a machine learning dynamic optimization algorithm was used to establish a service demand prediction model. The key input features included user call duration, call frequency, problem type, and satisfaction. The trained decision tree model was used to predict future service needs and dynamically adjust resource allocation based on the prediction results. For example, the model predicted that between 17:00 and 19:00 on Friday night, user demand for technical support would increase by 30%, so the operations team prepared to increase customer service staff and adjust service priorities.

[0066] Refer to Table 1 for a comparative analysis of the experimental data.

[0067] Table 1 Experimental data record table

[0068]

[0069] Through the analysis of the data in Table 1, it can be seen that the performance of each type of user in different dimensions fully reflects the innovation and advantages of the invention content. For example, the call duration of high-frequency users is significantly longer than that of low-frequency users, and their voice recognition accuracy and satisfaction scores are as high as 92% and 4.2 points. Such data show that for high-frequency users, the adaptability and efficiency of the IVR system are better than the service for low-frequency users. By identifying and optimizing the service process of this user group, user satisfaction and customer loyalty can be greatly improved. In the process of comparing different user behaviors, although long-call users have higher call duration and satisfaction, the hang-up reason rate is 5%, indicating that users naturally hang up after completing the business. Compared with the hang-up reason rate of 10% for high-frequency users, high-frequency users obviously have more dissatisfaction in the service process, which suggests that special attention should be paid to high-frequency user groups when designing more friendly IVR processes.

[0070] Furthermore, the high call volume and satisfaction scores of users with increased technical support needs on Friday nights verified the necessity and effectiveness of implementing a resource allocation strategy based on data analysis and behavior pattern recognition. Compared with the existing technology, the dynamic optimization algorithm can not only accurately predict service needs, but also adjust resource allocation in advance during peak service periods, thereby reducing user waiting time and improving service experience. It can be seen that the present invention is creative in improving user experience and service efficiency.

[0071] Example 3, reference Figure 2 , as an embodiment of the present invention, provides an IVR call full-link analysis system, including a data processing module, a user behavior analysis module, and a resource dynamic adjustment module.

[0072] The data processing module is used to collect call data in real time. By setting data embedding points, it captures the interaction data between users and the IVR system and pre-processes the collected data. The user behavior analysis module is used to perform full-link call analysis on the pre-processed data to identify user behavior patterns and service demand trends. The resource dynamic adjustment module is used to establish the first service demand forecast and dynamically adjust system resources based on the user behavior analysis results, thereby improving user experience and service efficiency.

[0073] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0074] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0075] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0076] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc. It should be noted that the above embodiments are only used to illustrate the technical solution of the present invention and are not limited. Although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

[0077] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for analyzing the entire link of IVR traffic, characterized in that: include: Collect traffic data and pre-process it; Performing first user behavior pattern recognition on the collected data; Based on the results of user behavior analysis, the first service demand forecast is established and system resources are dynamically adjusted to improve user experience and service efficiency.

2. The IVR traffic full-link analysis method according to claim 1, characterized in that: The collecting of traffic data includes collecting traffic data by setting data embedding points.

3. The IVR traffic full-link analysis method according to claim 2, characterized in that: The collecting of traffic data also includes capturing the interaction data between the user and the IVR system in real time.

4. The IVR traffic full-link analysis method according to claim 3, characterized in that: The first user behavior pattern recognition includes performing full-link traffic analysis on the preprocessed data.

5. The IVR traffic full-link analysis method according to claim 4, characterized in that: The first user behavior pattern identification also includes identifying user behavior patterns and service demand trends.

6. The IVR traffic full-link analysis method according to claim 5, characterized in that: The establishing of the first service demand forecast includes establishing the first service demand forecast based on historical data and full-link analysis results, predicting future service volume and service type, and monitoring user behavior and system operation status in real time.

7. The IVR traffic full-link analysis method according to claim 6, characterized in that: The dynamically adjusting system resources includes combining key performance indicators with user behaviors, triggering setting thresholds and determination rules based on the first service demand prediction result; When the system detects that user satisfaction is lower than the set threshold, it automatically adjusts the IVR voice prompts and interaction process; When a user's call duration exceeds twice the average call duration, the user is identified as a long-call user and additional service resources are allocated to the user; When it is predicted that the service demand will increase beyond the set ratio, the resource expansion mechanism is triggered to increase service resources; When the system performance index is lower than the set threshold, the performance optimization mechanism is triggered to optimize and adjust the system.

8. A system using the IVR traffic full-link analysis method according to any one of claims 1 to 7, characterized in that: Including data processing module, user behavior analysis module, and resource dynamic adjustment module; The data processing module is used to collect traffic data in real time, capture the interaction data between users and the IVR system by setting data embedding points, and pre-process the collected data; The user behavior analysis module is used to perform full-link traffic analysis on the pre-processed data to identify user behavior patterns and service demand trends; The resource dynamic adjustment module is used to establish a first service demand forecast and dynamically adjust system resources based on the user behavior analysis results to improve user experience and service efficiency.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the IVR traffic full-link analysis method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the IVR traffic full-link analysis method described in any one of claims 1 to 7 are implemented.

Images