Customer demand prediction method and system

Abstract

The invention provides a customer demand prediction method and system, and is suitable for the technical field of data processing, and the method comprises the steps: carrying out the feature extraction and analysis processing of a plurality of pieces of historical customer demand information, and generating a plurality of pieces of historical customer demand feature information; performing classification processing on the plurality of pieces of historical customer demand feature information to generate a plurality of pieces of historical customer demand group information; performing prediction calculation according to multiple pieces of historical customer demand group information, randomly generated customer group demand prediction weight information, randomly generated customer group demand prediction bias information, a preset customer group demand prediction weight adjustment step length and a preset customer group demand prediction bias adjustment step length; and obtaining multiple pieces of current customer demand information. According to the method, the prediction parameters are dynamically optimized for different customer groups, so that the customer demand prediction result effectively adapts to diversified computing resource constraints, the prediction accuracy and efficiency are improved, large-scale customer demands are accurately grasped, and optimization of service resource configuration is facilitated.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and in particular to a method and system for predicting customer demand. Background Art

[0002] As digitalization sweeps across industries, the field of customer demand forecasting continues to innovate. The continuous integration of cutting-edge technologies like big data and artificial intelligence is driving companies to actively explore ways to uncover customer needs and improve service quality.

[0003] Existing technologies typically rely on traditional data mining algorithms and basic machine learning models. For example, some companies use time series analysis to predict future demand trends based on the chronological order of historical customer demand data. For example, they analyze monthly product sales data from past years to estimate next month's sales. Other companies use simple linear regression models to establish a linear relationship between basic customer attributes (such as age and spending amount) and demand, thereby making preliminary quantitative forecasts of customer demand.

[0004] However, the time series analysis methods in existing technologies can often only capture short-term, regular changes in data. They are difficult to effectively respond to sudden and irregular factors in the market, such as policy adjustments and sudden social events that lead to drastic changes in customer demand. The prediction results often lag behind actual demand changes. The simple linear regression model oversimplifies the complex relationship between customer demand and influencing factors and cannot fully consider the diversity and dynamics of customer behavior. Summary of the Invention

[0005] In view of this, the embodiments of the present application provide a customer demand forecasting method and system, which aims to solve the problem of insufficient forecasting ability in the prior art when dealing with the complexity and variability of customer demand and the impact of sudden market factors.

[0006] A first aspect of an embodiment of the present application provides a customer demand forecasting method, comprising: Obtain multiple historical customer demand information; Extracting and analyzing the features of the plurality of historical customer demand information to generate a plurality of historical customer demand feature information; Classify the plurality of historical customer demand feature information to generate a plurality of historical customer demand group information; Based on the multiple historical customer demand group information, multiple randomly generated customer group demand prediction weight information, multiple randomly generated customer group demand prediction bias information, preset customer group demand prediction weight adjustment step and preset customer group demand prediction bias adjustment step, prediction calculation is performed to obtain multiple current customer demand information.

[0007] A second aspect of an embodiment of the present application provides a customer demand forecasting system, including: A historical customer demand information acquisition module is used to acquire multiple historical customer demand information; A historical customer demand feature information generation module is used to extract and analyze the features of the plurality of historical customer demand information to generate a plurality of historical customer demand feature information; A historical customer demand group information generation module is used to classify the plurality of historical customer demand feature information to generate a plurality of historical customer demand group information; The current customer demand information generation module is used to perform prediction calculations based on the multiple historical customer demand group information, multiple randomly generated customer group demand prediction weight information, multiple randomly generated customer group demand prediction bias information, preset customer group demand prediction weight adjustment step size, and preset customer group demand prediction bias adjustment step size to obtain multiple current customer demand information.

[0008] The third aspect of an embodiment of the present application provides a terminal device, which includes a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, it implements the steps of the customer demand forecasting method described in the first aspect above.

[0009] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium, comprising: storing a computer program, which, when executed by a processor, implements the steps of the customer demand forecasting method described in the first aspect above.

[0010] Compared with the prior art, the beneficial effects of the embodiments of the present application are as follows: the present application captures the key characteristics of customer demand from different granularities, groups customers with similar demand patterns into one category, deeply explores the unique demand patterns of different groups, and dynamically optimizes prediction parameters based on the characteristics of different customer groups, so that the calculated customer demand prediction results can not only adapt to diverse computing resource constraints, but also improve efficiency while ensuring prediction accuracy, providing strong support for enterprises to accurately grasp large-scale customer demand and optimize service resource allocation, thereby achieving flexible and efficient demand prediction and avoiding resource waste or untimely demand response. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0012] Figure 1 This is a schematic diagram of the implementation process of the customer demand forecasting method provided in Example 1 of the present application; Figure 2 This is a schematic diagram of the implementation process of the customer demand forecasting method provided in Example 2 of the present application; Figure 3 This is a schematic diagram of the implementation flow of the customer demand forecasting method provided in Example 3 of the present application; Figure 4 This is a schematic diagram of the implementation flow of the customer demand forecasting method provided in Example 4 of the present application; Figure 5 This is a schematic diagram of the implementation flow of the customer demand forecasting method provided in Example 5 of the present application; Figure 6 This is a schematic diagram of the implementation flow of the customer demand forecasting method provided in Example 6 of the present application; Figure 7 This is a schematic diagram of the implementation flow of the customer demand forecasting method provided in Example 7 of the present application; Figure 8 Schematic diagram of the structure of the customer demand forecasting system provided in an embodiment of the present application; Figure 9 It is a schematic diagram of the terminal device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0013] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0014] In order to illustrate the technical solution described in this application, specific embodiments are provided below.

[0015] Figure 1 The following is a flowchart of the customer demand forecasting method provided in Example 1 of the present application, which is described in detail as follows: Step S101: Acquire multiple historical customer demand information.

[0016] In this embodiment, historical customer demand information may be records of various needs generated by customers during their interactions with the enterprise over a period of time. These records include data reflecting customer demand characteristics, such as the content of customer demand for products or services, the time of demand, the frequency of demand, and the scale of demand. Historical customer demand information may be collected from the enterprise's business systems. For example, this information may include customer order information (including the type, quantity, and time of order) from the sales record system, customer consultation and repair request records from the customer service system, and customer browsing and click behavior data from interaction logs. This collected data is then pre-processed through cleaning and standardization to remove invalid information and unify the data format, forming a collection of historical customer demand information that can be used for subsequent feature extraction and analysis.

[0017] Step S102 : extracting and analyzing the features of the plurality of historical customer demand information to generate a plurality of historical customer demand feature information.

[0018] In this embodiment, a plurality of historical customer demand information may be collected and sorted first. Such information covers the customer's past records in terms of demand content, time, frequency, scale, etc. to form an original data set. Then, these original historical customer demand information are preliminarily processed through a deep neural network to convert it into a high-dimensional feature vector containing rich information. This vector can extract various potential information from the original data, and then corresponding analyzers are set for different levels of feature granularity. These analyzers optimize sub-vectors of different lengths in the high-dimensional feature vector to ensure that sub-vectors of each length can effectively capture the key features of historical customer demand. In the optimization process, historical customer demand information can be combined with the original historical customer demand information to obtain the key features of historical customer demand. The method combines the label information of historical customer needs with the deviation of the analysis results corresponding to the sub-vectors at each level, and then integrates these deviations according to a certain importance. By continuously adjusting the network parameters, the sub-vectors at each level can accurately reflect the characteristics of historical customer needs. At the same time, in this process, information will naturally diffuse to all dimensions. Even the intermediate-length sub-vectors that have not been specifically optimized can maintain good accuracy through this diffusion. Therefore, for multiple input historical customer demand information, a high-dimensional vector containing features of different granularity can be generated through a single network processing. The sub-vectors at each level are valid historical customer demand feature information, thus obtaining multiple historical customer demand feature information.

[0019] Step S103 , classifying the plurality of historical customer demand feature information to generate a plurality of historical customer demand group information.

[0020] In this embodiment, representative intermediate granularity features from multiple historical customer demand feature information obtained through feature extraction may be first selected as the basis for classification. These features have captured the key commonalities and differences of customer needs through preliminary processing. Then, several group cores are initialized, each core represents a potential customer group, and the similarity between each historical customer demand feature information and each group core is calculated. The feature information is assigned to the most similar group, and then the core of each group is updated according to the assignment result to make it more consistent with the average level of all feature information in the group. The steps of feature assignment and core update are repeated until the changes in the group core are stable and no longer fluctuate significantly. After multiple iterative optimizations, the originally scattered multiple historical customer demand feature information are divided into several groups with similar internal features and significantly different external features. The set corresponding to each group is a historical customer demand group information, thereby generating multiple historical customer demand group information.

[0021] Step S104, performing prediction calculations based on the multiple historical customer demand group information, multiple randomly generated customer group demand prediction weight information, multiple randomly generated customer group demand prediction bias information, preset customer group demand prediction weight adjustment step, and preset customer group demand prediction bias adjustment step to obtain multiple current customer demand information.

[0022] In this embodiment, the preset customer group demand forecast weight adjustment step and the preset customer group demand forecast bias adjustment step can be manually preset. Based on the generated multiple historical customer demand group information, the randomly generated customer group demand forecast weight information and bias information can be matched for each group. These weights and biases are used to preliminarily construct the demand forecast relationship for each group. Then, using the actual demand data of the historical customer demand group, the deviation between the current forecast result and the actual result is calculated to evaluate the rationality of the weights and biases. Subsequently, according to the preset customer group demand forecast weight adjustment step and the preset customer group demand forecast bias adjustment step, the weights and biases with large deviations are adjusted: for groups with good forecast results, small-scale fine adjustments are made near their weights and biases to further optimize the forecast accuracy; for groups with poor forecast results, the adjustment range is expanded to explore new weight and bias combinations. Repeat the process of deviation evaluation and parameter adjustment, continuously iteratively optimize the weights and biases until the deviation between the prediction results and the actual historical data stabilizes at a low level, and finally apply the optimized weights and biases to the information of each historical customer demand group, and obtain the current customer demand information corresponding to each group through predictive calculation, thereby generating multiple current customer demand information. The process of predictive calculation using the finally optimized weights and biases can be to first call the corresponding optimized customer group demand prediction weight information and customer group demand prediction bias information for each historical customer demand group information, and then associate and integrate the historical customer demand feature information of the group with the corresponding weight information to capture the key influence relationship between features and demand, and then perform basic calibration on the integration results in combination with the bias information to initially obtain the basic value of the demand forecast for the group, and then refer to the changing pattern and feature distribution characteristics of the historical demand within the group to fine-tune the basic forecast value to ensure that the prediction result fits the actual demand pattern of the group. Finally, after the above integration, calibration and fine-tuning steps, corresponding current customer demand information is generated for each historical customer demand group, thereby obtaining multiple current customer demand information.

[0023] In this embodiment, the preset customer group demand forecast weight adjustment step size and the preset customer group demand forecast bias adjustment step size can be initially determined based on the characteristic dimension of the historical customer demand data and the number of groups. For example, if the dimension of the historical customer demand characteristic information is 2048 and is divided into 10 historical customer demand group information, the weight adjustment step size can be initially set between 0.01 and 0.1, and the bias adjustment step size can be set between 0.001 and 0.01. The step size can then be dynamically refined based on the degree of demand fluctuation within the group. For groups with large demand fluctuations (such as demand standard deviations exceeding 30% of the mean in historical data), the adjustment step size can be appropriately increased (such as the weight step size is set to 0.05 and the bias step size is set to 0.005) to accelerate adaptation to significant demand changes. For groups with stable demand (demand standard deviation is less than 10% of the mean), reduce the step size (such as setting the weight step size to 0.02 and the bias step size to 0.002). Finally, the step size can be calibrated through small-scale verification data, and some historical customer demand group information can be selected for pre-training. If it is found that the prediction deviation continues to decrease and the convergence is stable, maintain the current step size; if the deviation fluctuates repeatedly, reduce the step size by 10%; if the convergence is too slow, increase the step size by 10%, and finally determine the optimal adjustment step size combination suitable for different groups. For example, after calibration, the weight adjustment step size of a high volatility group is set to 0.06, and the bias adjustment step size is set to 0.006, and the ones of a low volatility group are set to 0.015 and 0.0015 respectively, so as to achieve accurate and efficient parameter optimization.

[0024] The customer demand prediction method provided in the embodiment of the present application captures the key characteristics of customer demand from different granularities, classifies customers with similar demand patterns, deeply explores the unique demand patterns of different groups, and dynamically optimizes prediction parameters according to the characteristics of different customer groups, so that the calculated customer demand prediction results can not only adapt to diverse computing resource constraints, but also improve efficiency while ensuring prediction accuracy, providing strong support for enterprises to accurately grasp large-scale customer demand and optimize service resource allocation, thereby achieving flexible and efficient demand prediction and avoiding resource waste or untimely demand response.

[0025] Figure 2 The flowchart of the customer demand forecasting method provided in the second embodiment of the present application is shown. The difference between the second embodiment and the first embodiment is that step S102 specifically includes: Step S201 : encoding and normalizing the plurality of historical customer demand information to generate a plurality of historical customer demand vectors.

[0026] In this embodiment, non-numeric data in historical customer demand information (such as demand content and product type) can be first converted into numerical codes, for example, encoding "hardware procurement" as 1 and "software service" as 2. Numerical data (such as demand quantity and consumption amount) are normalized to the range [0, 1] using the Min-Max method, for example, adjusting a customer's single demand quantity from 50 to 0.6 (the original range is 0-100). Subsequently, the encoded and normalized data are concatenated in chronological order or by feature category to generate a fixed-length historical customer demand vector, such as a 256-dimensional vector containing dimensions such as demand type, timestamp, quantity, and amount.

[0027] Step S202 : performing feature extraction calculation on the plurality of historical customer demand vectors according to a preset historical customer demand query feature extraction vector to obtain a plurality of historical customer demand query feature information.

[0028] In this embodiment, the preset historical customer demand query feature extraction vector can be manually preset or a randomly initialized 128-dimensional vector. Feature extraction is achieved by performing matrix multiplication with the historical customer demand vector. For example, a 256-dimensional historical customer demand vector can be multiplied by a 128-dimensional query extraction vector to obtain 128-dimensional historical customer demand query feature information, focusing on the core objective of customer needs (such as features related to "whether the need is urgent"). The query extraction vector is initially set to a value range of [-0.1, 0.1] and subsequently fine-tuned through training to adapt it to the demand forecasting task.

[0029] Step S203 : performing feature extraction calculation on the plurality of historical customer demand vectors according to the preset historical customer demand key feature extraction vector to obtain a plurality of historical customer demand key feature information.

[0030] In this embodiment, the preset historical customer demand key feature extraction vectors can be manually set and can be composed of three groups of 128-dimensional random vectors (corresponding to the three sub-features of demand time, type, and scale). Each group of vectors is multiplied by the sub-vector of the corresponding dimension in the historical customer demand vector. The multiplication results serve as the key feature information of the historical customer demand to extract key attribute features. For example, the time dimension sub-vector is multiplied by the first group of key vectors to obtain a key feature reflecting the periodicity of demand; the type dimension sub-vector is multiplied by the second group of key vectors to obtain a key feature that distinguishes demand categories. The initial range of the key extraction vectors is set to [-0.1, 0.1], and the weights of different groups of vectors are adjusted independently.

[0031] Step S204 : performing feature extraction calculation on the plurality of historical customer demand vectors according to a preset historical customer demand value feature extraction vector to obtain a plurality of historical customer demand value feature information.

[0032] In this embodiment, the preset historical customer demand value feature extraction vector can be manually preset and can be a 128-dimensional vector that matches the dimension of the key feature extraction vector. The numerical attributes of specific demand features are extracted from the historical customer demand vector through linear transformation, such as the fluctuation range of the demand quantity, the distribution characteristics of the service duration, etc. The initial construction of the value extraction vector refers to the feature importance of the historical data, and a higher initial weight (such as 0.2) is assigned to high-frequency demand features (such as "demand quantity"), and a lower initial weight (such as 0.05) is assigned to low-frequency features (such as "remarks information"). This can be achieved by multiplying the preset historical customer demand value feature extraction vector with multiple historical customer demand vectors, and the multiplication result is used as multiple historical customer demand value feature information.

[0033] Step S205 : performing interactive processing based on the plurality of historical customer demand query feature information and the plurality of historical customer demand key feature information to obtain a plurality of historical customer demand interactive information.

[0034] In this embodiment, the similarity between each historical customer demand query feature information and all historical customer demand key feature information is calculated, for example, by measuring the strength of association using a vector dot product. The dot product result of each historical customer demand query feature information and all historical customer demand key feature information can be normalized using a softmax function to obtain historical customer demand interaction information. The interaction results corresponding to the top five key features with the highest historical customer demand interaction information are concatenated to generate historical customer demand interaction information. This information captures the matching relationship between the query target and the key features, such as the strong interaction between the "urgent demand" query and the "time urgency" key feature.

[0035] Step S206 , performing fusion processing on the plurality of historical customer demand value feature information and the plurality of historical customer demand interaction information to generate a plurality of historical customer demand feature variables.

[0036] In this embodiment, the interaction results corresponding to the top five key features with the highest historical customer demand interaction information can be used as coefficients to perform a weighted summation on the corresponding historical customer demand value feature information. For example, a weighted fusion of the "quantity fluctuation" value feature with a weight of 0.3 and the "time distribution" value feature with a weight of 0.2 can be performed. After fusion, it can be converted into a 64-dimensional vector through a fully connected layer, namely the historical customer demand feature variable, integrating the numerical attributes of the key features and the interaction importance.

[0037] Step S207 : performing feature conversion and enhancement processing on the multiple historical customer demand vectors according to the multiple historical customer demand feature variables to generate multiple historical customer demand feature information.

[0038] In this embodiment, the historical customer demand feature variables and the original historical customer demand vector can be concatenated by dimension, the original information is preserved through residual connections, and a nonlinear transformation is introduced through the ReLU function to enhance the feature expression capability. The content calculated by the ReLU function is used as the historical customer demand feature information. For example, a 256-dimensional original vector and a 64-dimensional feature variable can be concatenated into a 320-dimensional vector. After transformation, historical customer demand feature information containing multi-scale features is generated, covering both original attributes and interactively enhanced features.

[0039] The customer demand prediction method provided in the embodiment of the present application further enhances the richness and pertinence of historical customer demand features through refined feature coding, multi-dimensional extraction and interactive fusion, provides a more accurate feature basis for subsequent group classification and demand prediction, enhances the ability to capture complex customer demand patterns while ensuring efficiency, and helps enterprises grasp the laws of customer demand more accurately.

[0040] Figure 3 The flowchart of the customer demand forecasting method provided in the third embodiment of the present application is shown. The difference between the third embodiment and the first embodiment is that the step S102 specifically includes: Step S301 : performing feature extraction processing on the plurality of historical customer demand information according to a preset historical customer demand occurrence frequency feature extraction vector to obtain a plurality of historical customer demand occurrence frequency feature information.

[0041] In this embodiment, the preset historical customer demand frequency feature extraction vector can be manually set. It can be based on frequency-related dimensions such as "weekly demand frequency," "demand interval days," and "quarterly demand peak frequency" in the historical data to initialize a 16-dimensional weight vector. The weight of the "weekly demand frequency" dimension, which has a significant impact on high-frequency demand, is set to 0.3, the weight of the "quarterly demand peak frequency" dimension, which reflects long-term regularities, is set to 0.25, and the weights of the remaining dimensions are assigned in descending order of importance (e.g., 0.15, 0.1, etc.). By performing a weighted calculation on the frequency-related data in the historical customer demand information (e.g., a customer's "weekly demand frequency is 2 times, demand interval is 3 days") and this vector, features that reflect the regularity of demand occurrence are extracted, and multiple historical customer demand frequency feature information is generated.

[0042] Step S302 : performing feature extraction processing on the plurality of historical customer demand information according to a preset historical customer demand spatiotemporal preference feature extraction vector to obtain a plurality of historical customer demand spatiotemporal preference feature information.

[0043] In this embodiment, the preset historical customer demand spatiotemporal preference feature extraction vector can be manually set. It can be based on the customer's time preference (such as "weekday / weekend" and "peak season / off-season") and spatial attributes (such as "service area" and "remote / on-site demand"). A 20-dimensional weight vector is initialized. In the time preference dimension, the weight of "peak season demand proportion" is set to 0.3, and the weight of "weekend demand frequency" is set to 0.2. In the spatial attribute dimension, the weight of "core area demand proportion" is set to 0.25, and the weights of the remaining dimensions are set to 0.05-0.1. For example, the information of a customer with "peak season demand proportion of 70% and core area demand proportion of 80%" is weighted extracted to generate features reflecting spatiotemporal preferences. Multiple historical customer demand spatiotemporal preference feature information is obtained, which can adapt the logic of multi-granularity features to capture information of different dimensions.

[0044] Step S303 : performing feature extraction processing on the plurality of historical customer demand information according to a preset historical customer demand inter-correlation feature extraction vector to obtain a plurality of historical customer demand inter-correlation feature information.

[0045] In this embodiment, the preset historical customer demand intercorrelation feature extraction vector can be manually set. It can focus on the dependencies between requirements (e.g., "whether hardware purchases include installation services" or "the frequency of software upgrades and technical support") to initialize a 24-dimensional weight vector. The weight of the dimension corresponding to a strongly correlated requirement combination (e.g., "hardware purchase - installation services") can be set to 0.35, the weight of the dimension corresponding to a weakly correlated but potentially important requirement combination (e.g., "demand conversion rate within 7 days of consultation") can be set to 0.2, and the weights of the remaining correlated dimensions can be set between 0.05 and 0.15. By performing a weighted calculation on the correlated events recorded in historical customer demands, the synergistic patterns between requirements are extracted, generating multiple historical customer demand intercorrelation feature information.

[0046] Step S304 : performing matching processing on the plurality of historical customer demand occurrence frequency feature information and the plurality of historical customer demand spatiotemporal preference feature information to obtain a plurality of historical customer demand spatiotemporal preference frequency matching information.

[0047] In this embodiment, the degree of match between each historical customer demand frequency feature and the historical customer demand spatiotemporal preference feature can be calculated. For example, by comparing the overlap between "peak season high-frequency demand" and "peak season period in spatiotemporal preference," high weights are assigned to feature combinations with high matching (e.g., a weight of 0.8 for overlaps above 80%), and low weights are assigned to combinations with low matching (e.g., a weight of 0.2 for overlaps below 30%). The weighted matching results are then split by time period (daily, weekly, monthly), generating multiple historical customer demand spatiotemporal preference frequency matching information to capture patterns in "when and where demand frequently occurs." This can be achieved by nesting features to associate information of different granularities.

[0048] Step S305 , performing fusion processing based on the plurality of historical customer demand interrelated feature information and the plurality of historical customer demand spatiotemporal preference frequency matching information to generate a plurality of historical customer demand feature information.

[0049] In this embodiment, historical customer demand correlation feature information and spatiotemporal preference frequency matching information can be aligned by dimension. Key correlation features (such as "high-frequency hardware demand accompanied by installation services during peak season") are assigned a fusion weight of 0.4, core spatiotemporal frequency features (such as "core regional demand during quarterly peak periods") are assigned a weight of 0.3, and remaining weights are assigned to other features based on their importance. This weighted fusion integrates the correlation patterns and spatiotemporal frequency patterns of demand, and then, through multi-layer processing, retains multi-granular features (such as macro-level quarterly patterns and micro-level weekly fluctuations). Ultimately, multiple historical customer demand feature information is generated, achieving a comprehensive depiction of demand across the frequency, spatiotemporal, and correlation dimensions.

[0050] The customer demand prediction method provided in the embodiment of the present application focuses on the three core dimensions of demand frequency, spatiotemporal preference and correlation, and combines the precise feature extraction and multi-dimensional fusion of preset extraction vectors. The generated historical customer demand feature information can more comprehensively capture the dynamic laws and internal correlations of customer demand, and provide a feature basis that is more in line with actual scenarios for subsequent group classification and demand prediction, helping enterprises to achieve more accurate pattern mining and prediction decisions in large-scale customer demand analysis.

[0051] Figure 4 The flowchart of the customer demand forecasting method provided in the fourth embodiment of the present application is shown. The difference between the fourth embodiment and the first embodiment is that step S103 specifically includes: Step S401 : randomly extracting the plurality of historical customer demand feature information based on preset historical customer demand group quantity information to obtain a plurality of historical customer demand core feature information.

[0052] In this embodiment, the preset number of historical customer demand groups can be manually set, and can be based on the total number and diversity of historical customer demand feature information. If the total feature information is 1,000 and the demand patterns are significantly different, the preset number of groups can be 5 to 8; if the feature similarity is high, it can be preset to 3 to 5. For example, for a data set containing 1,200 feature information, the preset number of groups is 6. Feature information equal to the number of groups is then randomly extracted from multiple historical customer demand feature information as the initial hub, such as 6 randomly selected from 1,200, each representing the initial core of a group, to generate multiple historical customer demand hub feature information.

[0053] Step S402 : obtaining a plurality of historical customer demand peripheral feature information based on the plurality of historical customer demand feature information and the plurality of historical customer demand central feature information.

[0054] In this embodiment, the remaining historical customer demand feature information, excluding the extracted core features, can be defined as peripheral feature information. For example, out of 1200 feature information, 6 are core features, and the remaining 1194 are peripheral feature information, ensuring that each peripheral feature can be used for subsequent distance calculations with the core feature.

[0055] Step S403, calculate the Manhattan distance of the multiple historical customer demand central feature information and the multiple historical customer demand peripheral feature information to obtain multiple historical customer demand feature radial distance information; the historical customer demand feature radial distance information corresponds one-to-one to the historical customer demand central feature information and the historical customer demand peripheral feature information.

[0056] In this embodiment, the Manhattan distance can be calculated for each historical customer demand peripheral feature information and all the central feature information, that is, the absolute value of the difference is calculated dimension by dimension according to the feature dimension and the sum is calculated. For example, the absolute values ​​of the difference between a peripheral feature and the central feature in five dimensions such as "demand frequency" and "time and space preference" are 0.2, 0.3, 0.1, 0.4, and 0.2 respectively, and the radial distance after summing is 1.2. The distance results of each peripheral feature and each central feature are recorded separately to generate multiple historical customer demand feature radial distance information corresponding to each other, and quantify the degree of difference between the features.

[0057] Step S404 : taking the minimum value of the plurality of historical customer demand feature radial distance information corresponding to the plurality of historical customer demand surrounding feature information as the plurality of historical customer demand surrounding feature group confirmation distance information.

[0058] In this embodiment, for each historical customer demand peripheral feature information, the minimum value is selected from the radial distances between it and all the hub features as the basis for confirming that the peripheral feature belongs to a certain group. For example, if the distances between a peripheral feature and six hubs are 1.2, 0.8, 1.5, 0.9, 1.1, and 1.3, respectively, the minimum value of 0.8 is its group confirmation distance information, reflecting the strength of the association between the peripheral feature and the nearest hub.

[0059] Step S405 , classifying the plurality of historical customer demand surrounding feature information based on the plurality of historical customer demand surrounding feature group confirmation distance information and the historical customer demand core feature information, and generating a plurality of historical customer demand to-be-confirmed group information.

[0060] In this embodiment, each historical customer demand peripheral feature information may be assigned to the group containing the central feature corresponding to its confirmation distance. For example, the peripheral features with a minimum distance of 0.8 are assigned to the group corresponding to the central feature. After all peripheral features are assigned, a temporary group is formed with the central feature as the core and the peripheral features as members. Each group contains the central feature and the peripheral features to which it belongs, generating multiple historical customer demand group information to be confirmed. If a group contains less than 5% of the total number of peripheral features (e.g., a group contains only 20 out of 1,200), it can be marked as a group to be adjusted.

[0061] Step S406 : performing iterative classification processing on the plurality of historical customer demand to-be-confirmed group information to obtain a plurality of historical customer demand surrounding feature information.

[0062] In this embodiment, the average characteristic value of all characteristic information in each group to be confirmed can be recalculated and used as a new central feature; steps S403 to S405 are repeated to recalculate the distance and distribute the peripheral features according to the new central feature until the change in the central feature of the group in two iterations is less than the preset threshold (such as the average difference in each dimension is less than 0.05). For example, after the initial central feature is iterated, the "demand frequency" dimension is adjusted from 0.6 to 0.58, and the "time and space preference" dimension is adjusted from 0.7 to 0.69. The iteration is stopped when the change is stable. Each group finally formed contains stable central features and attribution features, and generates multiple historical customer demand group information to ensure that the features within the group are similar and the differences between groups are significant.

[0063] The customer demand forecasting method provided in the embodiment of the present application, by presetting a reasonable number of groups, precise classification based on distance and iterative optimization, generates historical customer demand group information that can more accurately reflect the demand patterns of different customer groups, provide a more targeted group basis for subsequent forecasts, and improve the rationality of group division while ensuring classification efficiency, helping enterprises to accurately explore the demand patterns of different groups.

[0064] Figure 5 The flowchart of the customer demand forecasting method provided in the fifth embodiment of the present application is shown. The difference between the fifth embodiment and the fourth embodiment is that the step S406 specifically includes: Step S501 : calculating the mean of the plurality of historical customer demand to be confirmed groups information to obtain the plurality of historical customer demand to be confirmed groups central information.

[0065] In this embodiment, the average value of all characteristic information contained in each historical customer demand pending confirmation group is calculated by dimension. For example, the "demand frequency" dimension of a certain group contains characteristic values ​​of 0.6, 0.7, and 0.5, with an average of 0.6; the "time and space preference" dimension contains characteristic values ​​of 0.8, 0.7, and 0.9, with an average of 0.8. The average values ​​of each dimension are combined to form a new hub for the group, generating multiple hub information for historical customer demand pending confirmation groups. The new hub must fit the overall distribution of characteristics within the group.

[0066] Step S502, determine whether the central information of the plurality of historical customer demands to be confirmed groups is the same as the central characteristic information of the plurality of historical customer demands; if so, proceed to step S503; if not, proceed to step S504.

[0067] In this embodiment, the preset judgment threshold is that the sum of the absolute values ​​of the feature differences in each dimension is ≤ 0.1. For example, if the difference in "demand frequency" between the new hub and the original hub is 0.02 and the difference in "time and space preference" is 0.03, and the total difference is 0.05≤0.1, then the judgment is the same; if the total difference is greater than 0.1 (such as the sum of the differences in each dimension between the original hub and the new hub is 0.15), then the judgment is different. This threshold is set based on the reasonable range of feature fluctuations in historical data to ensure that the iteration is terminated after the group hub is stable, avoiding premature termination of classification due to small fluctuations.

[0068] Step S503 : generating a plurality of historical customer demand group information based on the plurality of historical customer demand to-be-confirmed group information.

[0069] In this embodiment, when the center of the group to be confirmed is the same as the original center, it means that the group core has stabilized. At this time, the information of each group to be confirmed is directly determined as the final group, which contains the stable center characteristics and the peripheral characteristic information of the belonging. For example, after the center of the six groups to be confirmed is determined to be stable, the core dimension characteristics of each group, such as "demand frequency" and "time and space preference", no longer change significantly, and multiple historical customer demand group information is generated to ensure high similarity of characteristics within the group and obvious differences between groups.

[0070] Step S504 , using the plurality of historical customer demand to-be-confirmed group hub information as a plurality of historical customer demand hub feature information, and returning to step S402 .

[0071] In this embodiment, if the center of the group to be confirmed is different from the original center, it means that the core of the group still needs to be optimized. The newly calculated center information of the group to be confirmed replaces the original center feature information. For example, the "demand frequency" center with an average of 0.6 is used to replace the original center's 0.5 as the initial center for a new round of iteration. Then, the distance between the surrounding features and the new center is recalculated and the group is assigned until the center is stable.

[0072] The customer demand forecasting method provided in the embodiment of the present application realizes dynamic optimization of the group center and precise control of iteration termination through a clear mean calculation and threshold judgment mechanism, avoids blind iteration or premature convergence in the classification process, and generates historical customer demand group information with higher stability and stronger representative features, providing more reliable customer group information for subsequent demand forecasting, helping enterprises to more accurately explore the demand patterns of different groups.

[0073] Figure 6 The flowchart of the customer demand forecasting method provided in the sixth embodiment of the present application is shown. The difference between the sixth embodiment and the first embodiment is that the step S104 specifically includes: Step S601: Generate multiple customer group demand forecast parameter group information based on the multiple randomly generated customer group demand forecast weight information and the multiple randomly generated customer group demand forecast bias information; the customer group demand forecast parameter group information includes customer group demand forecast weight information and customer group demand forecast bias information.

[0074] In this embodiment, each randomly generated customer group demand prediction weight information and its corresponding randomly generated customer group demand prediction bias information can be combined to form a parameter group. For example, random weights (such as 0.2, 0.3, 0.1, etc.) and biases (such as 0.05, 0.03, 0.02, etc.) are generated for the six historical customer demand groups respectively. The weight and bias of each group correspond one-to-one, and six customer group demand prediction parameter group information are generated to ensure that each group has an independent prediction parameter combination.

[0075] Step S602 : calculating multiple customer group demand prediction accuracy representation information based on the multiple historical customer demand group information and multiple customer group demand prediction parameter group information.

[0076] In this embodiment, the demand forecast parameter group information for each customer group can be applied to the corresponding historical customer demand group information. The group characteristics are weighted and integrated using the weights in the parameter group. A forecast result is obtained by combining bias calibration. The forecast result is then compared with the actual historical demand data for the group, and the degree of deviation (such as mean absolute error) is calculated as an accuracy indicator. For example, if the average error between the forecast result and the actual demand for a group is 5%, the accuracy indicator is 95% (1-error rate). Multiple customer group demand forecast accuracy indicators are generated to quantify the forecast effect of the parameter group.

[0077] Step S603 , determining whether the maximum value of the plurality of customer group demand prediction accuracy representation information is greater than or equal to a preset customer group demand prediction accuracy representation threshold; if so, proceeding to step S604 ; if not, proceeding to step S605 .

[0078] In this embodiment, the preset customer group demand forecast accuracy threshold can be manually set and can be set based on business needs. For example, it can be set to 90% for the core customer group and 85% for the general group, and the comprehensive threshold is set to 88%. If the maximum value of the accuracy characterization information is 92% ≥ 88%, the current parameter group is judged to meet the standard; if the maximum value is 85% < 88%, it is judged to be unsatisfactory. This threshold refers to the average accuracy of historical forecast tasks to ensure that the forecast results meet the actual application requirements and avoid low accuracy affecting decision-making.

[0079] Step S604 , performing prediction calculation based on the customer group demand prediction parameter group information corresponding to the maximum value of the plurality of customer group demand prediction accuracy representation information and a plurality of historical customer demand group information to obtain a plurality of current customer demand information.

[0080] In this embodiment, when the maximum accuracy reaches the target, the parameter set corresponding to that maximum value is selected as the optimal parameter set and applied to all historical customer demand group information. By integrating group characteristics with optimal weights, combining bias calibration, and fine-tuning based on historical group demand patterns, a current demand forecast for each group is generated, with examples such as "current demand for high-frequency peak season groups is expected to increase by 10%" and "demand for stable regional groups remains stable." This generates multiple current customer demand information sets, ensuring that forecasts are generated based on optimal parameters.

[0081] Step S605 , performing optimization calculation based on the multiple customer group demand forecast accuracy characterization information, multiple customer group demand forecast parameter group information, preset customer group demand forecast weight adjustment step, and preset customer group demand forecast bias adjustment step to generate multiple customer group demand forecast parameter group information to be optimized.

[0082] In this embodiment, for parameter groups whose accuracy does not meet the requirements, the parameters are adjusted according to the accuracy characterization information: the parameter groups with lower accuracy (such as accuracy of 80%) have their adjustment range expanded by preset steps (weight step of 0.05, bias step of 0.005), and the parameter groups with accuracy close to the threshold (such as accuracy of 86%) are fine-tuned by small steps (weight step of 0.02, bias step of 0.002). For example, a certain weight is increased from 0.2 to 0.25, and the bias is fine-tuned from 0.05 to 0.045, thereby generating information on parameter groups to be optimized for demand forecasting of multiple customer groups.

[0083] Step S606 , using the plurality of customer group demand forecast parameter group information to be optimized as the plurality of customer group demand forecast parameter group information, and returning to step S602 .

[0084] In this embodiment, if the accuracy does not meet the target, the optimized parameter set to be optimized replaces the original parameter set, and the iteration is repeated until the maximum accuracy is met. For example, after two iterations, the parameter set accuracy increases from 85% to 90%. Once the threshold requirement is met, the iteration is stopped to ensure that a high-quality parameter set is obtained through continuous optimization.

[0085] The customer demand forecasting method provided in the embodiment of the present application ensures that the forecasting parameters always adapt to the group demand characteristics through a closed-loop mechanism of parameter group generation, accuracy evaluation and dynamic optimization, and generates reliable current customer demand information on the premise of meeting the accuracy threshold. It not only improves the targetedness of parameter optimization, but also ensures the practicality of the forecast results, helping enterprises to achieve efficient resource allocation and rapid response to demand based on accurate forecasting.

[0086] Figure 7 The flowchart of the customer demand forecasting method provided in the seventh embodiment of the present application is shown. The difference between the seventh embodiment and the sixth embodiment is that the step S605 specifically includes: Step S701 : taking the customer group demand prediction parameter group information corresponding to the maximum value of the plurality of customer group demand prediction accuracy representation information as the customer group demand prediction core parameter group information.

[0087] In this embodiment, the maximum value can be selected from the demand forecast accuracy representation information of multiple customer groups. For example, the maximum accuracy value is 92%. The parameter group corresponding to this maximum value (such as a weight of 0.3 and a bias of 0.04) is determined as the core parameter group and used as the optimization benchmark. The core parameter group must have the current optimal prediction effect and provide a reference for subsequent peripheral parameter optimization.

[0088] Step S702 : obtaining a plurality of customer group demand prediction peripheral parameter group information based on the plurality of customer group demand prediction parameter group information and the customer group demand prediction core parameter group information.

[0089] In this embodiment, the remaining parameter groups, excluding the core parameter group, can be defined as peripheral parameter groups. For example, if there are six parameter groups in total, one of which is the core parameter group, then the remaining five are peripheral parameter groups. The peripheral parameter groups are compared with the core parameter group and their accuracy information is lower than that of the core parameter group (e.g., 85%, 88%, etc.), providing candidates for parameter optimization.

[0090] Step S703, based on the customer group demand prediction core parameter group information, the preset customer group demand prediction weight adjustment step and the preset customer group demand prediction bias adjustment step, the plurality of customer group demand prediction peripheral parameter group information is optimized to obtain a plurality of customer group demand prediction peripheral parameter group information to be optimized.

[0091] In this embodiment, the preset customer group demand forecast weight adjustment step size and bias adjustment step size can be set based on the gap between the core parameter group accuracy and the peripheral parameter group: if the gap between the peripheral parameter group accuracy and the core is large (e.g., the gap is ≥5%), a larger step size (weight step size 0.05, bias step size 0.005) is used; if the gap is small (e.g., the gap is <3%), a smaller step size (weight step size 0.02, bias step size 0.002) is used. For example, if the accuracy of a peripheral parameter group is 85%, which is 7% away from the core accuracy of 92%, the weight is adjusted from 0.2 to 0.25 and the bias is adjusted from 0.03 to 0.035 according to the larger step size to generate the peripheral parameter group to be optimized, so that the peripheral parameters are closer to the core parameters.

[0092] Step S704 : generating a plurality of customer group demand prediction parameter group information to be optimized based on the customer group demand prediction core parameter group information and the customer group demand prediction peripheral parameter group information to be optimized.

[0093] In this embodiment, the core parameter group can be combined with the optimized peripheral parameter groups to form a new set of parameter groups to be optimized. For example, the core parameter group (92% accuracy) is retained and the five optimized peripheral parameter groups (with improved accuracy of 88%-90%) are incorporated to generate six parameter groups to be optimized. The parameter groups to be optimized should cover the stability advantages of the core parameters and the optimization potential of the peripheral parameters, providing higher-quality candidate parameters for the next round of accuracy evaluation.

[0094] The customer demand forecasting method provided in the embodiment of the present application uses a hierarchical optimization mechanism to determine the optimization benchmark through the core parameter group and the targeted adjustment of the peripheral parameter group. It not only retains the advantages of the current optimal parameters, but also promotes the convergence of peripheral parameters to high precision through dynamic step size, reduces the cost of blind iteration, and makes parameter optimization more targeted and efficient. The generated parameter group to be optimized can reach the accuracy threshold more quickly, providing strong support for the final generation of reliable current customer demand information, and helping enterprises improve the accuracy and efficiency of demand forecasting.

[0095] Corresponding to the method of the above embodiment, Figure 8 A structural block diagram of a customer demand forecasting system provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown. Figure 8 The exemplary customer demand forecasting system may be an execution subject of the customer demand forecasting method provided in the aforementioned first embodiment.

[0096] Reference Figure 8 , the customer demand forecasting system includes: A historical customer demand information acquisition module 810 is used to acquire a plurality of historical customer demand information; A historical customer demand feature information generating module 820 is configured to extract features from and analyze the plurality of historical customer demand information to generate a plurality of historical customer demand feature information; A historical customer demand group information generating module 830 is configured to classify the plurality of historical customer demand feature information to generate a plurality of historical customer demand group information; The current customer demand information generation module 840 is used to perform prediction calculations based on the multiple historical customer demand group information, multiple randomly generated customer group demand prediction weight information, multiple randomly generated customer group demand prediction bias information, preset customer group demand prediction weight adjustment step and preset customer group demand prediction bias adjustment step to obtain multiple current customer demand information.

[0097] The process of each module realizing its own function in the customer demand forecasting system provided in the embodiment of the present application can be specifically referred to the aforementioned Figure 1 The description of the first embodiment is omitted here.

[0098] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0099] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, integers, steps, operations, elements and / or components, but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0100] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0101] As used in this specification and the appended claims, the term "if" can be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0102] In addition, in the description of the present specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish descriptions and should not be understood as indicating or implying relative importance. It should also be understood that although the terms "first", "second", etc. are used in the text to describe various elements in some embodiments of the present application, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first table can be named a second table, and similarly, a second table can be named a first table without departing from the scope of the various described embodiments. Both the first table and the second table are tables, but they are not the same table.

[0103] References to "one embodiment" or "some embodiments" in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in one or more embodiments of the present application. Thus, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in other embodiments" appearing in various places in this specification do not necessarily refer to the same embodiment, but rather mean "one or more but not all embodiments," unless otherwise specifically emphasized. The terms "including," "comprising," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0104] The customer demand forecasting method provided in the embodiments of the present application can be applied to terminal devices such as mobile phones, tablet computers, wearable devices, vehicle-mounted devices, augmented reality (AR) / virtual reality (VR) devices, laptop computers, ultra-mobile personal computers (UMPCs), netbooks, and personal digital assistants (PDAs). The embodiments of the present application do not impose any restrictions on the specific types of terminal devices.

[0105] For example, the terminal device can be a station (STAION, ST) in a WLAN, a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA) device, a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a vehicle networking terminal, a computer, a laptop computer, a handheld communication device, a handheld computing device, a satellite wireless device, a wireless modem card, a TV set-top box (STB), customer premise equipment (CPE) and / or other devices for communicating on a wireless system and a next-generation communication system, such as a mobile terminal in a 5G network or a mobile terminal in a future evolved Public Land Mobile Network (PLMN) network.

[0106] As an example and not a limitation, when the terminal device is a wearable device, the wearable device can also be a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear, such as glasses, gloves, watches, clothing and shoes. A wearable device is a portable device that is worn directly on the body or integrated into the user's clothes or accessories. Wearable devices are not only hardware devices, but also achieve powerful functions through software support, data interaction, and cloud interaction. Broadly speaking, wearable smart devices include those that are full-featured, large in size, and can achieve complete or partial functions without relying on smartphones, such as smart watches or smart glasses, as well as those that only focus on a certain type of application function and need to be used in conjunction with other devices such as smartphones, such as various smart bracelets and smart jewelry for vital sign monitoring.

[0107] Figure 9 This is a schematic diagram of the structure of a terminal device provided by an embodiment of the present application. Figure 9 As shown, the terminal device 9 of this embodiment includes: at least one processor 90 ( Figure 9 Only one is shown), a memory 91, wherein the memory 91 stores a computer program 92 that can be run on the processor 90. When the processor 90 executes the computer program 92, the steps in the above-mentioned customer demand forecasting method embodiments are implemented, such as Figure 1 Alternatively, when the processor 90 executes the computer program 92, the functions of the modules / units in the above-mentioned system embodiments are realized, for example, Figure 8Functions of modules 810 to 840 are shown.

[0108] The terminal device 9 can be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal device can include, but is not limited to, a processor 90 and a memory 91. It can be understood by those skilled in the art that Figure 9 It is only an example of the terminal device 9 and does not constitute a limitation on the terminal device 9. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the terminal device may also include an input and sending device, a network access device, a bus, etc.

[0109] The processor 90 may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0110] In some embodiments, the memory 91 may be an internal storage unit of the terminal device 9, such as a hard drive or memory of the terminal device 9. The memory 91 may also be an external storage device of the terminal device 9, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. Furthermore, the memory 91 may include both an internal storage unit of the terminal device 9 and an external storage device. The memory 91 is used to store an operating system, application programs, a boot loader, data, and other programs, such as the program code of the computer program. The memory 91 may also be used to temporarily store data that has been sent or is about to be sent.

[0111] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0112] An embodiment of the present application also provides a terminal device, which includes at least one memory, at least one processor, and a computer program stored in the at least one memory and executable on the at least one processor. When the processor executes the computer program, the terminal device implements the steps of any of the above-mentioned method embodiments.

[0113] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned various method embodiments can be implemented.

[0114] An embodiment of the present application provides a computer program product. When the computer program product is run on a terminal device, the terminal device can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0115] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application can implement all or part of the processes in the above-mentioned embodiment method by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium can include: any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium.

[0116] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0117] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0118] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0119] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A customer demand forecasting method, characterized in that: include: Obtain multiple historical customer demand information; Extracting and analyzing the features of the plurality of historical customer demand information to generate a plurality of historical customer demand feature information; Classify the plurality of historical customer demand feature information to generate a plurality of historical customer demand group information; Based on the multiple historical customer demand group information, multiple randomly generated customer group demand prediction weight information, multiple randomly generated customer group demand prediction bias information, preset customer group demand prediction weight adjustment step and preset customer group demand prediction bias adjustment step, prediction calculation is performed to obtain multiple current customer demand information.

2. The customer demand forecasting method according to claim 1, wherein: The step of extracting and analyzing the features of the plurality of historical customer demand information to generate a plurality of historical customer demand feature information specifically includes: Encoding and normalizing the plurality of historical customer demand information to generate a plurality of historical customer demand vectors; According to a preset historical customer demand query feature extraction vector, feature extraction calculation is performed on the multiple historical customer demand vectors to obtain multiple historical customer demand query feature information; Performing feature extraction calculation on the plurality of historical customer demand vectors according to a preset historical customer demand key feature extraction vector to obtain a plurality of historical customer demand key feature information; Performing feature extraction calculation on the plurality of historical customer demand vectors according to a preset historical customer demand value feature extraction vector to obtain a plurality of historical customer demand value feature information; Performing interactive processing based on the plurality of historical customer demand query feature information and the plurality of historical customer demand key feature information to obtain a plurality of historical customer demand interactive information; Performing fusion processing based on the plurality of historical customer demand value feature information and the plurality of historical customer demand interaction information to generate a plurality of historical customer demand feature variables; According to the multiple historical customer demand feature variables, feature conversion and enhancement processing are performed on the multiple historical customer demand vectors to generate multiple historical customer demand feature information.

3. The customer demand forecasting method according to claim 1, wherein: The step of extracting and analyzing the features of the plurality of historical customer demand information to generate a plurality of historical customer demand feature information specifically includes: Performing feature extraction processing on the plurality of historical customer demand information according to a preset historical customer demand occurrence frequency feature extraction vector to obtain a plurality of historical customer demand occurrence frequency feature information; Performing feature extraction processing on the plurality of historical customer demand information according to a preset historical customer demand spatiotemporal preference feature extraction vector to obtain a plurality of historical customer demand spatiotemporal preference feature information; Performing feature extraction processing on the plurality of historical customer demand information according to a preset historical customer demand inter-correlation feature extraction vector to obtain a plurality of historical customer demand inter-correlation feature information; Matching the plurality of historical customer demand frequency feature information with the plurality of historical customer demand spatiotemporal preference feature information to obtain a plurality of historical customer demand spatiotemporal preference frequency matching information; A fusion process is performed based on the mutual correlation feature information of the multiple historical customer demands and the spatiotemporal preference frequency matching information of the multiple historical customer demands to generate multiple historical customer demand feature information.

4. The customer demand forecasting method according to claim 1, wherein: The step of classifying the plurality of historical customer demand feature information to generate a plurality of historical customer demand group information specifically includes: According to the preset number of historical customer demand groups, the plurality of historical customer demand feature information is randomly sampled to obtain a plurality of historical customer demand central feature information; Obtaining a plurality of historical customer demand peripheral feature information based on the plurality of historical customer demand feature information and the plurality of historical customer demand central feature information; Calculating the Manhattan distances of the plurality of historical customer demand central feature information and the plurality of historical customer demand peripheral feature information to obtain a plurality of historical customer demand feature radial distance information; wherein the historical customer demand feature radial distance information corresponds one-to-one to the historical customer demand central feature information and the historical customer demand peripheral feature information; The minimum value of the plurality of historical customer demand feature radial distance information corresponding to the plurality of historical customer demand surrounding feature information is used as the plurality of historical customer demand surrounding feature group confirmation distance information; Classify the plurality of historical customer demand surrounding feature information based on the plurality of historical customer demand surrounding feature group confirmation distance information and the historical customer demand central feature information to generate a plurality of historical customer demand to-be-confirmed group information; Iterative classification processing is performed on the plurality of historical customer demand to-be-confirmed group information to obtain a plurality of historical customer demand peripheral feature information.

5. The customer demand forecasting method according to claim 4, wherein: The step of iteratively classifying the plurality of historical customer demand to-be-confirmed group information to obtain the plurality of historical customer demand surrounding feature information specifically includes: Calculating the mean of the plurality of historical customer demand to be confirmed groups information to obtain the plurality of historical customer demand to be confirmed groups central information; Determining whether the plurality of historical customer demand to-be-confirmed group central information is the same as the plurality of historical customer demand central feature information; If yes, generating multiple historical customer demand group information based on the multiple historical customer demand to be confirmed group information; If not, the central information of the multiple historical customer demands to be confirmed groups is used as the central feature information of multiple historical customer demands, and the process returns to the step of obtaining the peripheral feature information of multiple historical customer demands based on the multiple historical customer demands feature information and the central feature information of multiple historical customer demands.

6. The customer demand forecasting method according to claim 1, wherein: The step of performing prediction calculations based on the multiple historical customer demand group information, the multiple randomly generated customer group demand prediction weight information, the multiple randomly generated customer group demand prediction bias information, the preset customer group demand prediction weight adjustment step, and the preset customer group demand prediction bias adjustment step to obtain the multiple current customer demand information specifically includes: Generate multiple sets of customer group demand forecast parameter information based on the multiple randomly generated customer group demand forecast weight information and the multiple randomly generated customer group demand forecast bias information; the customer group demand forecast parameter group information includes the customer group demand forecast weight information and the customer group demand forecast bias information; Calculating multiple customer group demand prediction accuracy representation information based on the multiple historical customer demand group information and multiple customer group demand prediction parameter group information; Determining whether a maximum value of the plurality of customer group demand prediction accuracy representation information is greater than or equal to a preset customer group demand prediction accuracy representation threshold; If so, performing prediction calculation based on the customer group demand prediction parameter group information corresponding to the maximum value of the plurality of customer group demand prediction accuracy representation information and the plurality of historical customer demand group information to obtain a plurality of current customer demand information; If not, performing optimization calculation based on the multiple customer group demand forecast accuracy characterization information, multiple customer group demand forecast parameter group information, preset customer group demand forecast weight adjustment step, and preset customer group demand forecast bias adjustment step to generate multiple customer group demand forecast parameter group information to be optimized; The multiple customer group demand forecast parameter group information to be optimized is used as multiple customer group demand forecast parameter group information, and the step of calculating multiple customer group demand forecast accuracy characterization information based on the multiple historical customer demand group information and multiple customer group demand forecast parameter group information is returned.

7. The customer demand forecasting method according to claim 6, wherein: The step of performing optimization calculation based on the plurality of customer group demand forecast accuracy characterization information, the plurality of customer group demand forecast parameter group information, the preset customer group demand forecast weight adjustment step, and the preset customer group demand forecast bias adjustment step to generate the plurality of customer group demand forecast parameter group information to be optimized specifically includes: The customer group demand forecast parameter group information corresponding to the maximum value of the plurality of customer group demand forecast accuracy representation information is used as the customer group demand forecast core parameter group information; Obtaining multiple customer group demand forecast peripheral parameter group information based on the multiple customer group demand forecast parameter group information and the customer group demand forecast core parameter group information; Optimizing the plurality of customer group demand forecast peripheral parameter group information based on the customer group demand forecast core parameter group information, the preset customer group demand forecast weight adjustment step, and the preset customer group demand forecast bias adjustment step to obtain a plurality of customer group demand forecast peripheral parameter group information to be optimized; Based on the customer group demand prediction core parameter group information and the customer group demand prediction peripheral parameter group information to be optimized, multiple customer group demand prediction parameter group information to be optimized are generated.

8. A customer demand forecasting system, characterized in that: include: A historical customer demand information acquisition module is used to acquire multiple historical customer demand information; A historical customer demand feature information generation module is used to extract and analyze the features of the plurality of historical customer demand information to generate a plurality of historical customer demand feature information; A historical customer demand group information generation module is used to classify the plurality of historical customer demand feature information to generate a plurality of historical customer demand group information; The current customer demand information generation module is used to perform prediction calculations based on the multiple historical customer demand group information, multiple randomly generated customer group demand prediction weight information, multiple randomly generated customer group demand prediction bias information, preset customer group demand prediction weight adjustment step size, and preset customer group demand prediction bias adjustment step size to obtain multiple current customer demand information.

9. A terminal device, characterized in that: The terminal device includes a memory and a processor, wherein the memory stores a computer program that can be run on the processor, and when the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images