Hotel service defect prediction method, system and device and storage medium

Abstract

The invention provides a hotel service defect prediction method, system and device, and a storage medium. The method comprises the steps of obtaining historical hotspot events and hotel information; based on the historical hot events and the hotel information, constructing interaction features associated with the hotels and the hot events; based on the interaction features, training a preset composite model to obtain a service defect prediction model; acquiring a real-time hotspot event, and acquiring a corresponding defect prediction result according to the service defect prediction model; and corresponding pre-control prompts are provided for the corresponding possibly affected hotels according to the defect prediction result. Through the technical scheme provided by the invention, corresponding hotel defect prediction model training can be carried out according to the influence condition of the historical hot events on hotel service defects, and the real-time hot events collected in advance are analyzed according to the model obtained through training to obtain hotels possibly influenced by the real-time hot events; according to the method, targeted pre-control prompts and pre-control measures are given for service defect events which may influence hotels and may be caused based on real-time hot events, so that the service quality of the hotels and the hotel online management platform can be improved, the customer service satisfaction is improved, the operation and maintenance management cost of the platform is reduced, and the method has a promotable value.

Description

Background Art

[0002] With the rapid development of tourism and the popularization of the Internet, hotels often welcome a large number of visitors during hot events such as popular festivals, national examinations, concerts and marathons. Some hotels have poor service quality management, which leads to service defects such as price increases after customers make reservations and no rooms available upon arrival. This will affect the quality of customer service, resulting in a decline in customer experience, which in turn affects the hotel's online management platform and the hotel's reputation and economic benefits.

[0003] Service defects are a key indicator for measuring the quality of customer service, and hot events will have a significant impact on defects. Taking accurate preventive measures for hotels in advance can effectively reduce defects, so it is necessary to predict hotels with a high probability of service defects during hot events. Existing hotel service defect prediction methods mainly rely on manual experience and the impact of similar historical events, and the factors considered are not comprehensive, resulting in low coverage of hot event preventive hotels and low prediction accuracy, which affects the preventive effect. Summary of the invention

[0004] In view of the problems in the prior art, the purpose of the present disclosure is to provide a hotel service defect prediction method, system, device and storage medium. Specifically, the first aspect of the present disclosure provides a hotel service defect prediction method, which may include the following steps:

[0005] Get historical hot events and hotel information;

[0006] Based on historical hot events and hotel information, interactive features associated with hotels and hot events are constructed;

[0007] Based on the interactive features, the preset composite model is trained to obtain a service defect prediction model. The composite model includes a LightGBM model based on gradient boosting and a learning model based on the Wide&Deep architecture.

[0008] Obtain real-time hot events and obtain corresponding defect prediction results based on the service defect prediction model. Real-time hot events include hot events that are currently occurring and will occur within a preset event period in the future. Defect prediction results include the event heat, service defect impact category, and service defect impact level corresponding to the real-time hot events.

[0009] Provide corresponding preventive control tips to the hotels that may be affected based on the defect prediction results.

[0010] In a possible implementation of the first aspect above, the historical hot events include one or more combinations of historical event categories, historical event occurrence times, historical event durations, historical event impacts during the same period, and historical event impact ranges.

[0011] In a possible implementation of the first aspect above, the hotel information includes static configuration information and dynamic service information;

[0012] The static configuration information includes one or more combinations of hotel geographic location information, hotel room configuration information, hotel personnel configuration information, and hotel facility configuration information;

[0013] The dynamic service information includes one or more combinations of hotel historical visit information, hotel historical order information, hotel historical rating information, hotel historical customer complaint information, and hotel service defect information.

[0014] In a possible implementation of the first aspect, constructing an interactive feature associated with a hotel and a hot event based on historical hot events and hotel information includes the following steps:

[0015] Obtain a first feature set corresponding to historical hot events and a second feature set corresponding to hotel information;

[0016] Independently performing data cleaning and highly correlated feature selection on the first feature set and the second feature set, and uniformly numerically encoding the selected features to obtain the first numerical feature set and the second numerical feature set;

[0017] Based on a preset feature crossover method, a crossover operation is performed on at least part of the features in the first numerical feature set and the second numerical feature set to obtain interactive features, where the interactive features include service defects occurring in the hotel during hot events.

[0018] In a possible implementation of the first aspect, in the process of acquiring real-time hot events and acquiring corresponding defect prediction results according to the service defect prediction model, the following steps are included:

[0019] Generate corresponding real-time interactive features based on real-time hot events;

[0020] According to the real-time interaction features, predictions are performed using the trained gradient-boosted LightGBM model and the Wide&Deep architecture-based learning model to obtain a first prediction value corresponding to the gradient-boosted LightGBM model and a second prediction value corresponding to the Wide&Deep architecture-based learning model;

[0021] Based on the stacking method, a linear weighted sum is performed on the first prediction value and the second prediction value, and the obtained result is used as the defect prediction result corresponding to the service defect prediction model.

[0022] In a possible implementation of the first aspect, the process of providing a corresponding pre-control prompt to a corresponding hotel according to the defect prediction result includes the following steps:

[0023] According to the event heat corresponding to the defect prediction result, obtain the possible affected hotels within the impact range of the real-time hot event;

[0024] According to the service defect impact category and service defect impact level corresponding to the defect prediction results, corresponding preventive control prompts are provided to hotels that may be affected.

[0025] In a possible implementation of the first aspect, the process of providing a corresponding pre-control prompt to a corresponding hotel according to the defect prediction result includes the following steps:

[0026] According to the pre-control prompts, voice calls are made to hotels that may be affected through the preset voice interaction system;

[0027] If the voice call reminder is not successfully delivered, continue to make voice call reminders to the hotels that may be affected at preset intervals, or manually deliver pre-control reminders to the hotels that may be affected; and

[0028] According to the service defect impact category and service defect impact level corresponding to the defect prediction results, defect compensation resources are prepared in advance for hotels that may be affected.

[0029] A second aspect of the present disclosure provides a hotel service defect prediction system, which is used to implement the hotel service defect prediction method provided by the first aspect. The service defect prediction system may specifically include:

[0030] Information acquisition unit, used to obtain historical hot events and hotel information;

[0031] A feature fusion unit is used to construct interactive features associated with hotels and hot events based on historical hot events and hotel information;

[0032] The model training unit is used to train the preset composite model based on the interactive features to obtain a service defect prediction model, where the composite model includes a LightGBM model based on gradient boosting and a learning model based on the Wide&Deep architecture;

[0033] A defect prediction unit is used to obtain real-time hot events and obtain corresponding defect prediction results according to a service defect prediction model, wherein the real-time hot events include hot events that are currently occurring and that will occur within a preset event segment in the future, and the defect prediction results include the event heat, service defect impact category, and service defect impact level corresponding to the real-time hot events;

[0034] The pre-control response unit is used to provide corresponding pre-control prompts to the corresponding hotels that may be affected according to the defect prediction results.

[0035] A third aspect of the present disclosure provides a hotel service defect prediction device, comprising:

[0036] processor;

[0037] a memory in which executable instructions for the processor are stored;

[0038] Among them, the processor is configured to execute the steps of the hotel service defect prediction method provided in the first aspect by executing executable instructions.

[0039] A fourth aspect of the present disclosure provides a computer-readable storage medium for storing a program, which, when executed, implements the steps of the hotel service defect prediction method provided in the first aspect.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] Through the technical solution provided by the present invention, it is possible to train a corresponding hotel defect prediction model according to the impact of historical hot events on hotel service defects, and analyze the real-time hot events collected in advance according to the trained model to obtain hotels that may be affected by the real-time hot events; for hotels that may be affected, targeted preventive prompts and preventive measures are given based on the service defect events that may be caused by the real-time hot events, which can improve the service quality of hotels and hotel online management platforms, reduce the operation and maintenance management costs of the platform while improving customer service satisfaction, and has promotional value. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Other features, objects and advantages of the present disclosure will become more apparent from a reading of the detailed description of non-limiting embodiments made with reference to the following accompanying drawings.

[0043] Figure 1 According to an embodiment of the present disclosure, a flow chart of a hotel service defect prediction method is provided.

[0044] Figure 2 According to an embodiment of the present disclosure, a flow chart of constructing interactive features associated with hotels and hot events is provided.

[0045] Figure 3 According to an embodiment of the present disclosure, a flow chart of obtaining corresponding defect prediction results according to a service defect prediction model is provided.

[0046] Figure 4 According to an embodiment of the present disclosure, a flow chart is provided for providing corresponding pre-control prompts to corresponding hotels according to defect prediction results.

[0047] Figure 5 According to an embodiment of the present disclosure, a structural diagram of a hotel service defect prediction system is provided.

[0048] Figure 6 According to an embodiment of the present disclosure, a structural schematic diagram of a hotel service defect prediction device is provided.

[0049] Figure 7 According to an embodiment of the present disclosure, a schematic diagram of the structure of a computer-readable storage medium is provided. DETAILED DESCRIPTION

[0050] The following describes the implementation methods of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in the present application. The present application can also be implemented or applied through other different specific implementation methods, and the details in the present application can also be modified or changed in various ways according to different viewpoints and application systems without departing from the spirit of the present application. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0051] The following is a detailed description of the embodiments of the present application with reference to the accompanying drawings so that those skilled in the art can easily implement the present application. The present application can be embodied in a variety of different forms and is not limited to the embodiments described herein.

[0052] In the representations of the present application, the representations with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" etc. mean that the specific features, structures, materials or characteristics represented in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. Moreover, the represented specific features, structures, materials or characteristics may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples represented in the present application and the features of the different embodiments or examples, unless they contradict each other.

[0053] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the representation of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0054] In order to clearly describe the present application, components irrelevant to the description are omitted, and the same reference symbols are given to the same or similar components throughout the specification.

[0055] Throughout the specification, when a device is said to be "connected" to another device, this includes not only the case of "direct connection" but also the case of "indirect connection" by placing other elements therebetween. In addition, when a device is said to "include" a certain component, unless otherwise stated, it does not exclude other components, but means that other components may be included.

[0056] When a device is said to be "on" another device, it may be directly on the other device, but there may also be other devices between it. In contrast, when a device is said to be "directly" on another device, there are no other devices between it.

[0057] Although the terms first, second, etc. are used to represent various elements in some examples, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, the first interface and the second interface, etc. are represented. Moreover, as used in this article, the singular forms "one", "one" and "the" are intended to also include plural forms, unless there is an opposite indication in the context. It should be further understood that the terms "comprising" and "including" indicate the existence of features, steps, operations, elements, components, projects, kinds, and / or groups, but do not exclude the existence, occurrence or addition of one or more other features, steps, operations, elements, components, projects, kinds, and / or groups. The terms "or" and "and / or" used herein are interpreted as inclusive, or mean any one or any combination. Therefore, "A, B or C" or "A, B and / or C" means "any of the following: A; B; C; A and B; A and C; B and C; A, B and C". Only when the combination of elements, functions, steps or operations is inherently mutually exclusive in some way, will there be an exception to this definition.

[0058] The technical terms used herein are only used to refer to specific embodiments and are not intended to limit the present application. The singular form used herein also includes the plural form unless the sentence clearly indicates the contrary meaning. The meaning of "including" used in the specification is to specify specific characteristics, regions, integers, steps, operations, elements and / or components, and does not exclude the existence or addition of other characteristics, regions, integers, steps, operations, elements and / or components.

[0059] Although not defined differently, all terms, including technical and scientific terms used herein, have the same meaning as those generally understood by those skilled in the art to which this application belongs. Terms defined in commonly used dictionaries are additionally interpreted as having meanings consistent with the relevant technical literature and the contents of the current disclosure, and unless defined, shall not be overly interpreted as ideal or very formal meanings.

[0060] In some embodiments of the present disclosure, Figure 1 A flowchart of a hotel service defect prediction method is provided. Figure 1 As shown, the hotel service defect prediction method includes:

[0061] Step 101: Obtain historical hot events and hotel information. Among them, historical hot events may include one or more combinations of historical event categories, historical event occurrence time, historical event duration, historical concurrent event impact, and historical event impact range. Specifically, historical event categories may include large and small holidays, regional / national examinations, concerts / exhibitions, sports events, and other events that are likely to cause hotel service defects. The historical event occurrence time and historical event duration are used to associate hot events with hotel service defects in the same period, and the historical event impact range represents the popularity of historical hot events.

[0062] In the above step 101, the hotel information includes static configuration information and dynamic service information, wherein the static configuration information includes one or more combinations of hotel geographic location information, hotel room configuration information, hotel personnel configuration information, and hotel facility configuration information; the dynamic service information includes one or more combinations of hotel historical visit information, hotel historical order information, hotel historical rating information, hotel historical customer complaint information, and hotel service defect information. The static configuration information and dynamic service information can comprehensively cover the impact of historical hot spot information on the hotel.

[0063] Step 102: Based on the historical hot events and hotel information, construct interactive features associated with hotels and hot events. The specific acquisition of interactive features will be described in detail later, and will not be repeated here.

[0064] Step 103: Based on the interactive features, the preset composite model is trained to obtain a service defect prediction model. The composite model includes a LightGBM model based on gradient boosting and a learning model based on Wide&Deep architecture. The LightGBM model based on gradient boosting and the learning model based on Wide&Deep architecture will be described in detail later and will not be described here.

[0065] Step 104: Obtain real-time hot events, and obtain corresponding defect prediction results according to the service defect prediction model. The real-time hot events include hot events that are currently occurring and will occur within a preset event period in the future, and the defect prediction results include the event heat, service defect impact category, and service defect impact level corresponding to the real-time hot events.

[0066] Step 105: Providing corresponding pre-control prompts to the corresponding hotels that may be affected according to the defect prediction results.

[0067] It is understandable that through the above steps 101 to 105, the service defect prediction model can be trained according to the impact of historical hot events on hotel service defects, and the real-time hot events collected in advance are analyzed according to the service defect prediction model to obtain in advance the service defect impact that may affect the hotel due to the real-time hot events and perform advance control. The specific implementation of the above steps 101 to 105 will be further explained in conjunction with the embodiments below:

[0068] In the above embodiment, specifically, in the specific implementation of step 101, historical hot events can be collected and acquired by automatically crawling and collecting hot event information on the network. At the same time, in order to avoid missing historical hot events, manual collection and historical booking abnormality warnings in the hotel information service system can be combined to comprehensively collect and supplement hot event information.

[0069] In the above embodiment, specifically, in the specific implementation of step 102, Figure 2 A schematic diagram of a process for constructing interactive features associated with hotels and hot events is shown, such as Figure 2 As shown, the following steps may be specifically included:

[0070] Step 201: Obtain a first feature set corresponding to historical hot events and a second feature set corresponding to hotel information.

[0071] Step 202: Data cleaning and highly correlated feature selection are performed independently on the first feature set and the second feature set, and the selected features are uniformly numerically encoded to obtain the first numerical feature set and the second numerical feature set. It can be understood that the first feature set corresponding to the historical hot events and the second feature set corresponding to the hotel information are raw data, including text features, numerical features, category features and other different feature types, which need to be cleaned and converted for data preprocessing to ensure the quality and consistency of the acquired features, and further adopt feature engineering technology to select and combine the features to extract interactive features with predictive capabilities.

[0072] In the above step 202, data cleaning may specifically include data processing methods such as filling missing values, processing outliers, and removing duplicate values ​​to ensure the integrity and accuracy of the acquired features corresponding to the original data; high correlation feature selection is used to extract features with high correlation with the target variable, and specifically a model-based method, such as decision trees, random forests and other models, is used to select the feature subset with the strongest correlation with the service defect event, thereby reducing the complexity of the service defect prediction model and improving its accuracy in the subsequent model training process; unified numerical encoding may specifically include using one-hot encoding to convert categorical features into numerical features for subsequent machine learning model processing. Those skilled in the art may select a suitable data processing method according to actual needs, which is not limited here.

[0073] Step 203: Based on the preset feature crossover method, crossover operations are performed on at least part of the features in the first numerical feature set and the second numerical feature set to obtain interactive features. The interactive features include the service defects that occurred in the hotel during hot events. Specifically, the preset feature crossover method in the above step 203 can capture the nonlinear relationship between features, thereby further improving the performance of the subsequent service defect prediction model; specifically, a variety of feature crossover methods can be used, such as Cartesian product crossover, polynomial crossover, etc., to generate new features with the strongest correlation with service defect events, more accurately grasp the impact of historical hot events on hotel services, and provide more useful information for subsequent model training.

[0074] In the above embodiment, specifically, in the specific implementation of step 103, the composite model includes a LightGBM model based on gradient boosting and a learning model based on the Wide&Deep architecture. The specific training process of the above model will be described in detail below:

[0075] In the LightGBM algorithm, the embodiment provided by the present disclosure iteratively constructs multiple simple regression trees, each of which fits the error of the model's last prediction, and combines them into a more powerful model to improve prediction accuracy. The iterative process of the gradient boosting method is as follows:

[0076] Input: T = {(x1,y1),(x2,y2),…,(x n ,y n )}

[0077] Input: Gradient boosted tree f M (x)

[0078] Step S1: Initialization: In the initial stage of gradient boosting, a simple prediction value f0(x) (such as the average of all sample labels) is given as the initial prediction.

[0079] Step S2: Calculate the residual: For each sample, calculate the residual r between its true value and the current model prediction value mi =y i -f m-1 (x),i=1,2,…,N.

[0080] Step S3: Train a new model: Use the residual as the target value to train a new regression tree model T(x; Θ m ); this new model will attempt to fit the current residuals to correct the prediction errors of the previous model.

[0081] Step S4: Update prediction: Add the prediction value of the new model to the previous prediction value as the updated prediction value f m (x) = f m-1 (x)+T(x;Θ m ).

[0082] Step D5: Iteration: Repeat the above process until the maximum number of iterations is met or the residual is small enough to obtain a boosted tree.

[0083]

[0084] In the specific training process of the above-mentioned boosted tree, the histogram algorithm can be used to bin the eigenvalues, and the continuous eigenvalues ​​can be divided into multiple discrete intervals. For each feature, the LightGBM model will count the sum of its gradients and the number of samples in each bin, and construct the histogram of the feature, thereby reducing the memory required to store the eigenvalues. By traversing the histogram to calculate the gain of the split point, there is no need to access the original data corresponding to the hotel information and historical hot spot information, thereby effectively accelerating the training process. At the same time, gradient-based one-side sampling (GOSS) is used to reduce the number of samples in the training process according to sampling, and the samples can be sorted according to the absolute value of the gradient of the samples, retaining the samples that contribute more to the model training, while avoiding overfitting of the samples with smaller gradients. In addition, the Leaf-wise growth strategy is also used in the training process to select the leaf node with the largest gain for splitting, which can find the optimal tree structure more quickly, and limit the maximum depth of the tree and the maximum number of leaf nodes to prevent overfitting during the model training process.

[0085] In the deep table data learning network based on the Wide&Deep architecture, the linear network and the deep network can be integrated to fully utilize the advantages of the two models: the linear model is good at capturing the linear relationship and low-order interaction between features by explicitly defining feature crosses, while the deep network can automatically learn the high-order interaction and complex nonlinear relationship of features, and use the attention mechanism in the deep network to selectively focus on the important parts of the input features. Specifically, the Wide part is a linear model, which usually uses logistic regression as the basis, and can be expressed as:

[0086] p(y=1∣x)=σ(Wwide·x+bwide);

[0087] Where p(y=1|x) represents the probability that the target variable y is a positive class when a given input x is given. In this scenario, x is the constructed feature and y is the probability that a defect occurs in the hotel. σ(·) represents the Logistic sigmoid function, which maps the linear output to the probability interval (0,1). Wdeep and bdeep represent the weight vector and bias term of the linear model.

[0088] In the above embodiment, the Deep part is a deep neural network that can automatically learn the nonlinear representation and high-order interaction of features, which can be simplified as follows:

[0089] hdeep=DNN(x;Wdeep,bdeep);

[0090] Where hdeep is the feature representation of the deep network output; DNN(·) is a deep neural network with multiple hidden layers; Wdeep, bdeep represent the weights and bias parameters of the deep network. The Wide&Deep model optimizes performance by jointly training the Wide and Deep parts. The final prediction is the weighted sum of the outputs of the two parts:

[0091] p(y=1∣x)=σ(α·(Wwide·x+bwide)+β·hdeep).

[0092] It can be understood that by training and compounding the LightGBM model based on gradient boosting and the learning model based on the Wide&Deep architecture, the prediction accuracy of the service defect prediction model can be effectively improved, the generalization processing capability of the service defect prediction model for feature data can be enhanced, and the explanatory power of the model decision process can be provided to a certain extent. The service defect prediction model provided by the present disclosure has good adaptability and can cope with the distribution and type of feature data corresponding to different hot events and hotel information.

[0093] In the above embodiment, specifically, in the specific implementation of step 104, Figure 3 A schematic diagram of a process for obtaining corresponding defect prediction results according to a service defect prediction model is shown. Figure 3 Specifically, the following steps may be included:

[0094] Step 301: Generate corresponding real-time interactive features according to real-time hot events.

[0095] Step 302: According to the real-time interaction features, predictions are performed using the trained gradient boosting-based LightGBM model and the Wide&Deep architecture-based learning model to obtain a first prediction value corresponding to the gradient boosting-based LightGBM model and a second prediction value corresponding to the Wide&Deep architecture-based learning model.

[0096] Step 303: Perform a linear weighted sum of the first prediction value and the second prediction value based on the stacking method, and use the obtained result as the defect prediction result corresponding to the service defect prediction model.

[0097] In the above embodiment, specifically, in the specific implementation of step 105, Figure 4 A schematic diagram of a process for providing corresponding pre-control prompts to corresponding hotels according to defect prediction results is shown, such as Figure 4 As shown, the following steps may be specifically included:

[0098] Step 401: According to the event heat corresponding to the defect prediction result, the possible affected hotels within the influence range of the real-time hot event are obtained.

[0099] Step 402: Provide corresponding pre-control prompts to the hotels that may be affected according to the service defect impact category and service defect impact level corresponding to the defect prediction result. It can be understood that by pre-training the possible affected hotels and event type classification of real-time hot events, the pre-control standards in the automated process in the hotel online management platform are called to determine the pre-control gestures, so as to achieve the hierarchical and layered pre-control of the hotels that may be affected in advance, thereby improving the service quality of the hotel during the hot event period and improving customer service satisfaction.

[0100] Furthermore, in the above embodiment, the pre-control prompt can be used to make a voice call prompt to the possible affected hotel through a preset voice interaction system (Interactive Voice Response, IVR); specifically, the text-to-speech technology can be used to convert the pre-control prompt content into voice content, and initiate a voice call to the possible affected hotel; when the front desk of the possible affected hotel answers the voice call, the voice interaction system will automatically play the pre-control prompt content, wherein the pre-control prompt content includes information on upcoming hot events, and requires the possible affected hotel to prepare dynamic adjustment plans and plans for room rates and room status in a timely manner, so as to avoid service defects such as hotel price increases and full rooms after customers place orders.

[0101] In the case that the voice call prompt is not successfully delivered, that is, no one at the hotel front desk answers the voice call, the voice interaction system can record the unsuccessful delivery and trigger the intervention of the manual team to try to contact the hotel again, and communicate the reminder content with the hotel manually to ensure that the pre-control information can be delivered to the hotel that may be affected in time. At the same time, according to the service defect impact category and service defect impact level corresponding to the defect prediction results, defect compensation resources can be prepared in advance for the hotel that may be affected. In the event that real-time hot events are about to occur and continue, the operation and maintenance team of the hotel online management platform pays attention to the defect dashboard that may affect the hotel in real time, and promptly handles possible abnormal situations such as service defects or order backlogs, and promptly and reasonably mobilizes defect compensation resources, such as increasing service personnel, optimizing service processes, adjusting resource allocation, etc., to achieve timely response to service defect issues, effectively ensuring that customer service satisfaction does not fluctuate significantly due to real-time hot events.

[0102] In some embodiments of the present disclosure, a corresponding evaluation and adjustment mechanism is also provided to regularly evaluate the prediction accuracy of the hotel service prediction model provided in the aforementioned embodiments and the predictive effect of the targeted preventive control prompts, and adjust and optimize the model and the preventive control prompt strategy based on the evaluation results, thereby ensuring the timeliness and sustainability of the hotel service defect prediction method proposed in the present disclosure.

[0103] In some embodiments of the present disclosure, Figure 5 FIG. 1 shows a schematic diagram of the structure of a hotel service defect prediction system, which is used to implement the hotel service defect prediction method provided in the above embodiment. Figure 5 As shown, the service defect prediction system may specifically include:

[0104] The information acquisition unit 501 is used to acquire historical hot events and hotel information.

[0105] The feature fusion unit 502 is used to construct interactive features associated with hotels and hot events based on historical hot events and hotel information.

[0106] The model training unit 503 is used to train the preset composite model based on the interactive features to obtain a service defect prediction model, wherein the composite model includes a LightGBM model based on gradient boosting and a learning model based on the Wide&Deep architecture.

[0107] The defect prediction unit 504 is used to obtain real-time hot events and obtain corresponding defect prediction results according to the service defect prediction model. The real-time hot events include hot events that are currently occurring and will occur within a preset event period in the future, and the defect prediction results include the event heat, service defect impact category, and service defect impact level corresponding to the real-time hot events.

[0108] The pre-control response unit 505 is used to provide corresponding pre-control prompts to the corresponding hotels that may be affected according to the defect prediction results.

[0109] It can be understood that the functions implemented by the information acquisition unit 501 to the pre-control response unit 505 in the above functional modules correspond one-to-one to the actions performed by the above steps 101 to 105, and will not be described in detail here.

[0110] Those skilled in the art will appreciate that various aspects of the present disclosure may be implemented as systems, methods or program products. Therefore, various aspects of the present disclosure may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software implementations, which may be collectively referred to herein as "circuits", "modules" or "platforms".

[0111] Specifically, Figure 6 According to an embodiment of the present disclosure, a schematic diagram of the structure of a hotel service defect prediction device is shown. Figure 6 The electronic device 600 according to this embodiment of the present disclosure is described. Figure 6 The electronic device 600 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0112] like Figure 6 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0113] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 performs the steps according to various exemplary embodiments of the present disclosure. For example, the processing unit 610 can perform the following steps: Figure 1 The relevant steps of the hotel service defect prediction method shown in .

[0114] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory (RAM) 6201 and / or a cache storage unit 6202 , and may further include a read-only storage unit (ROM) 6203 .

[0115] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0116] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0117] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0118] The present disclosure also provides a computer-readable storage medium for storing a program, and the steps of the hotel service defect prediction method are implemented when the program is executed. In some possible implementations, various aspects of the present disclosure can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present disclosure described in the hotel service defect prediction method section of this specification.

[0119] Specifically, Figure 7 According to an embodiment of the present disclosure, a schematic diagram of the structure of a computer-readable storage medium is shown. Figure 7 As shown, a program product 800 for implementing the above-mentioned user call summary method according to an embodiment of the present disclosure is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto, and in this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0120] The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0121] Computer readable storage media may include data signals propagated in baseband or as part of a carrier wave, in which readable program codes are carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.

[0122] The program code for executing the specific implementation operations of the user call summary method provided by the aforementioned embodiment of the present disclosure can be written in any combination of one or more programming languages, and the programming language includes an object-oriented programming language, such as Java, C++, etc., and also includes a conventional procedural programming language, such as "C" language or a similar programming language. The program code can be executed entirely on the user computing device, partially on the user device, as an independent software package, partially on the user computing device and partially on the remote computing device, or completely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).

[0123] To sum up, through the technical solution provided by the present invention, it is possible to train the corresponding hotel defect prediction model according to the impact of historical hot events on hotel service defects, and analyze the real-time hot events collected in advance according to the trained model to obtain hotels that may be affected by the real-time hot events; for hotels that may be affected, targeted preventive prompts and preventive measures are given based on the service defect events that may be caused by real-time hot events, which can improve the service quality of hotels and hotel online management platforms, reduce the operation and maintenance management costs of the platform while improving customer service satisfaction, and has promotional value.

[0124] The above content is a further detailed description of the present disclosure in combination with specific preferred implementation modes, and it cannot be determined that the specific implementation of the present disclosure is limited to these descriptions. For ordinary technicians in the technical field to which the present disclosure belongs, several simple deductions or substitutions can be made without departing from the concept of the present disclosure, which should be regarded as falling within the scope of protection of the present disclosure.

Claims

1. A hotel service defect prediction method, It is characterized in that The steps include: Get historical hot events and hotel information; Based on the historical hot events and the hotel information, constructing interactive features associated with hotels and hot events; Based on the interaction features, a preset composite model is trained to obtain a service defect prediction model, wherein the composite model includes a LightGBM model based on gradient boosting and a learning model based on Wide&Deep architecture; Acquire real-time hot events, and acquire corresponding defect prediction results according to the service defect prediction model, wherein the real-time hot events include hot events that are currently occurring and that will occur within a preset event period in the future, and the defect prediction results include event heat, service defect impact category, and service defect impact level corresponding to the real-time hot events; According to the defect prediction results, corresponding preventive control prompts are provided to the corresponding hotels that may be affected.

2. The hotel service defect prediction method according to claim 1, It is characterized in that The historical hot events include one or more combinations of historical event categories, historical event occurrence times, historical event durations, historical event impacts during the same period, and historical event impact ranges.

3. The hotel service defect prediction method according to claim 1, It is characterized in that The hotel information includes static configuration information and dynamic service information; The static configuration information includes one or more combinations of hotel geographic location information, hotel room configuration information, hotel personnel configuration information, and hotel facility configuration information; The dynamic service information includes one or more combinations of hotel historical visit information, hotel historical order information, hotel historical rating information, hotel historical customer complaint information, and hotel service defect information.

4. The hotel service defect prediction method according to claim 1, It is characterized in that The step of constructing the interactive features associated with the hotel and the hot events based on the historical hot events and the hotel information includes the following steps: Acquire a first feature set corresponding to the historical hot event and a second feature set corresponding to the hotel information; Independently performing data cleaning and highly correlated feature selection on the first feature set and the second feature set, and uniformly numerically encoding the selected features to obtain a first numerical feature set and a second numerical feature set; Based on a preset feature crossover method, a crossover operation is performed on at least part of the features in the first numerical feature set and the second numerical feature set to obtain the interactive features, where the interactive features include service defects occurring in the hotel during hot events.

5. The hotel service defect prediction method according to claim 1, It is characterized in that The process of obtaining the real-time hot events and obtaining the corresponding defect prediction results according to the service defect prediction model includes the following steps: Generate corresponding real-time interactive features according to the real-time hot events; According to the real-time interaction feature, the gradient-boosted LightGBM model and the Wide&Deep architecture-based learning model obtained through training are respectively used to perform predictions to obtain a first prediction value corresponding to the gradient-boosted LightGBM model and a second prediction value corresponding to the Wide&Deep architecture-based learning model; Based on the stacking method, a linear weighted sum is performed on the first prediction value and the second prediction value, and the obtained result is used as the defect prediction result corresponding to the service defect prediction model.

6. The hotel service defect prediction method according to claim 1, It is characterized in that The process of providing corresponding pre-control prompts to the corresponding hotel according to the defect prediction results includes the following steps: According to the event heat corresponding to the defect prediction result, the potentially affected hotels within the impact range of the real-time hot event are obtained; According to the service defect impact category and the service defect impact level corresponding to the defect prediction result, the corresponding pre-control prompt is provided to the potentially affected hotel.

7. The hotel service defect prediction method according to claim 1 or claim 6, It is characterized in that The process of providing corresponding pre-control prompts to the corresponding hotel according to the defect prediction results includes the following steps: According to the pre-control prompt, a voice call prompt is made to the hotels that may be affected through a preset voice interaction system; In the case where the voice call prompt is not successfully delivered, continuously making the voice call prompt to the potentially affected hotel at a preset interval, or manually delivering the pre-control prompt to the potentially affected hotel; as well as According to the service defect impact category and service defect impact level corresponding to the defect prediction result, defect compensation resources are prepared in advance for the potentially affected hotels.

8. A hotel service defect prediction system, used to implement the hotel service defect prediction method according to any one of claims 1 to 7, It is characterized in that include: Information acquisition unit, used to obtain historical hot events and hotel information; A feature fusion unit, used to construct an interactive feature associated with the hotel and the hot event based on the historical hot event and the hotel information; A model training unit, used for training a preset composite model based on the interaction features to obtain a service defect prediction model, wherein the composite model includes a LightGBM model based on gradient boosting and a learning model based on Wide&Deep architecture; A defect prediction unit, used to obtain real-time hot events, and obtain corresponding defect prediction results according to the service defect prediction model, wherein the real-time hot events include hot events that are currently occurring and that will occur within a preset event segment in the future, and the defect prediction results include event heat, service defect impact category, and service defect impact level corresponding to the real-time hot events; The pre-control response unit is used to provide corresponding pre-control prompts to the corresponding hotels that may be affected according to the defect prediction results.

9. A hotel service defect prediction device, It is characterized in that include: processor; a memory storing executable instructions of the processor; Wherein, the processor is configured to execute the steps of the hotel service defect prediction method described in any one of claims 1 to claim 7 by executing the executable instructions.

10. A computer-readable storage medium for storing a program, It is characterized in that When the program is executed by a processor, the steps of the hotel service defect prediction method described in any one of claims 1 to 7 are implemented.

Images