An automated management system and method for device inventory based on data analysis

Abstract

The present invention relates to the field of automated inventory management, and provides an automated management system and method for equipment inventory based on data analysis. The method includes, in each inspection period of each inventory cycle, determining the equipment that is out of stock in the current cycle; predicting the data of the out-of-stock equipment based on the historical data of the equipment in the inventory; based on the predicted data of the out-of-stock equipment, using data analysis methods to match substitute equipment for the out-of-stock equipment in the potential inventory, and adding the successfully matched substitute equipment to the inventory; if there are out-of-stock equipment that have not been matched, determining the attribute distribution of the out-of-stock equipment that have not been matched based on the predicted data of the out-of-stock equipment that have not been matched, and searching for substitute equipment for the out-of-stock equipment that have not been matched in the potential inventory and adding it to the inventory; the present invention can improve the efficiency and response speed of inventory management, and reduce the operating costs and losses caused by out-of-stock.

Description

Technical Field

[0001] The present invention relates to the field of automated inventory management, and more specifically, to an automated management system and method for equipment inventory based on data analysis. Background Art

[0002] In traditional equipment inventory management, it usually relies on manual inspections and empirical judgments to determine which equipment in the inventory has a risk of out-of-stock. This method is not only inefficient, but also difficult to accurately predict the demand changes and attribute changes of equipment due to the lack of accurate data support, resulting in inflexible inventory management and difficulty in adapting to the needs of market changes. In addition, when equipment is out of stock, traditional methods often require manually searching for alternative equipment in the potential inventory, which is time-consuming and error-prone, increasing the operating costs and risks of enterprises.

[0003] Existing technologies often lack effective prediction means when dealing with equipment attribute changes, especially for those attributes that change over time and whose change laws are unknown. This results in difficulty in quickly finding suitable alternative equipment when equipment is out of stock, affecting the timely replenishment of inventory and the availability of equipment. At the same time, traditional inventory management systems often have problems of low computing efficiency and inaccurate data processing when dealing with large amounts of data, and it is difficult to meet the needs of modern enterprises for high-efficiency and high-precision inventory management.

[0004] In the process of implementing the embodiments of the present invention, the inventors found that there are at least the following problems or defects in the prior art: existing equipment inventory management methods lack effective prediction and matching mechanisms in the face of dynamic changes in equipment attributes, resulting in difficulty in quickly finding suitable alternative equipment when equipment is out of stock; in addition, existing systems have insufficient computing efficiency and accuracy when dealing with large amounts of equipment data, and it is difficult to meet the high requirements of modern enterprises for inventory management. Summary of the Invention

[0005] The present invention provides an automated management method for equipment inventory based on data analysis.

[0006] In a first aspect of the present invention, there is provided an automated management method for equipment inventory based on data analysis, including: determining the equipment out of stock in the current cycle at each inspection period of each inventory cycle; predicting the data of the out-of-stock equipment based on the historical data of the equipment in the inventory; based on the predicted data of the out-of-stock equipment, using data analysis methods to match alternative equipment for the out-of-stock equipment in the potential inventory, and adding the successfully matched alternative equipment to the inventory; if there are out-of-stock equipment that are not matched, determining the attribute distribution of the unmatched out-of-stock equipment based on the predicted data of the unmatched out-of-stock equipment, and searching for alternative equipment for the unmatched out-of-stock equipment in the potential inventory based on the attribute distribution of the unmatched out-of-stock equipment and adding it to the inventory.

[0007] Furthermore, predict the data of out-of-stock devices based on the historical data of devices in the inventory, including: directly predicting the first type of attributes of out-of-stock devices; for the second type of attributes of out-of-stock devices, use machine learning algorithms to match similar devices for out-of-stock devices from the inventory devices, and predict the second type of attributes of out-of-stock devices based on the values of the second type of attributes of similar devices; wherein, the first type of attributes are attributes that do not change with time or whose change patterns are known although they change with time, and the second type of attributes are attributes that change with time and whose change patterns are unknown.

[0008] Furthermore, use the following formula to predict the second type of attributes of out-of-stock devices based on the attribute values corresponding to similar devices: ; wherein, represents the predicted value of the i-th out-of-stock device on the k-th second type of attribute, represents the value of the m-th similar device on the k-th second type of attribute, and M represents the number of similar devices.

[0009] Furthermore, calculate the propensity score of each device being out of stock in the current period based on the following formula, and select the M devices with the propensity scores closest to that of the out-of-stock device from the inventory devices as similar devices: ; wherein, represents the propensity score of the i-th device being out of stock in the current period, which reflects the probability of the device being out of stock under given conditions, represents whether the i-th device is out of stock, represents that the i-th device is in an out-of-stock state, represents all the attribute values of the i-th device in the previous period, and P(·) represents probability.

[0010] Furthermore, use data analysis methods to match alternative devices for out-of-stock devices in the potential inventory, including: performing clustering analysis on each non-classified attribute of the inventory, and keeping the clustering of each classified attribute of the inventory unchanged; for each out-of-stock device, determine the cluster to which each attribute of the out-of-stock device belongs according to the predicted data of each attribute of the out-of-stock device; if there is a device in the potential inventory with the same cluster attribute distribution as the out-of-stock device, the matching is successful, and randomly select a device with the same cluster attribute distribution as the out-of-stock device as the alternative device matched to the out-of-stock device.

[0011] Furthermore, the method further includes: if the current time period is not the last time period of the current cycle, search for target new devices in the potential inventory based on the predicted data of the out-of-stock devices accumulated in the current cycle, and add the target new devices to the potential inventory.

[0012] In the second aspect of the present invention, there is provided an automated management system for device inventory based on data analysis, including: a shortage device determination module for determining devices that are out of stock during the current period at each inspection period of each inventory cycle; a data prediction module for predicting the data of the shortage devices based on the historical data of the devices in the inventory; a substitute device matching module for matching substitute devices for the shortage devices in the potential inventory using a data analysis method based on the predicted data of the shortage devices, and adding the successfully matched substitute devices to the inventory; an attribute distribution matching module for, if there are shortage devices that have not been matched, determining the attribute distribution of the shortage devices that have not been matched based on the predicted data of the shortage devices that have not been matched, and searching for substitute devices for the shortage devices that have not been matched in the potential inventory based on the attribute distribution of the shortage devices that have not been matched and adding them to the inventory.

[0013] Further, predicting the data of the shortage devices based on the historical data of the devices in the inventory includes: directly predicting the first type of attributes of the shortage devices; for the second type of attributes of the shortage devices, using a machine learning algorithm to match similar devices for the shortage devices from the inventory devices, and predicting the second type of attributes of the shortage devices based on the values of the second type of attributes of the similar devices; wherein, the first type of attributes are attributes that do not change with time or whose change pattern is known although they change with time, and the second type of attributes are attributes that change with time and whose change pattern is unknown.

[0014] Further, the following formula is used to predict the second type of attributes of the shortage devices based on the attribute values corresponding to the similar devices: ; where represents the predicted value of the i-th shortage device on the k-th second type of attribute, represents the value of the m-th similar device on the k-th second type of attribute, and M represents the number of similar devices.

[0015] Further, the system further includes: a strategy adjustment module for, if the current period is not the last period of the current cycle, searching for target new devices in the potential inventory based on the predicted data of the shortage devices accumulated in the current cycle, and adding the target new devices to the potential inventory.

[0016] The above embodiments of the present invention have at least the following beneficial effects: The device inventory automated management method based on data analysis adopted by the present invention can improve the accuracy and efficiency of inventory management. By deeply analyzing the historical data of the devices in the inventory, this method can predict the future demand and attribute changes of out-of-stock devices, so as to make inventory adjustments and optimizations in advance. This method can reduce the operation interruptions and cost increases caused by out-of-stock, and at the same time can also improve the availability and flexibility of the devices. In addition, by using machine learning algorithms to match similar devices, the second type of attributes, that is, those attributes that change over time and whose change rules are unknown, can be predicted more accurately, so as to find more suitable and efficient substitutes for out-of-stock devices.

[0017] In addition, this method processes the non-classified attributes in the inventory through cluster analysis while keeping the clusters of classified attributes unchanged, which helps to quickly identify and match devices with similar attributes. This method can not only replenish out-of-stock devices in a timely manner within the current inventory cycle, but also predict and prepare potential new devices in the early stage of the cycle, so as to ensure the continuous optimization and update of the inventory. Through this automated and intelligent inventory management, enterprises can better respond to market changes, improve customer satisfaction, and ultimately achieve cost savings and profit growth. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By referring to the following detailed description with reference to the accompanying drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown by way of example and not by way of limitation, in which:

[0019] Figure 1 is a schematic flow chart of a device inventory automated management method based on data analysis provided by an embodiment of the present invention;

[0020] Figure 2 is a schematic structural diagram of a device inventory automated management system based on data analysis provided by an embodiment of the present invention;

[0021] Figure 3 schematically shows a schematic structural diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are given only to enable those skilled in the art to better understand and then implement the present invention, and do not limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present invention more thorough and complete, and to be able to convey the scope of the present invention fully to those skilled in the art.

[0023] The following referenceFigure 1 , Figure 1 is a schematic flowchart of a method for automated management of equipment inventory based on data analysis provided by an embodiment of the present invention. As Figure 1 shown, a method 100 for automated management of equipment inventory based on data analysis includes: at each inspection period of each inventory cycle, determining the equipment that is out of stock in the current cycle; predicting the data of the out-of-stock equipment based on the historical data of the equipment in the inventory; based on the predicted data of the out-of-stock equipment, using data analysis methods to match alternative equipment for the out-of-stock equipment in the potential inventory, and adding the successfully matched alternative equipment to the inventory; if there are out-of-stock equipment that are not matched, determining the attribute distribution of the unmatched out-of-stock equipment based on the predicted data of the unmatched out-of-stock equipment, and searching for alternative equipment for the unmatched out-of-stock equipment in the potential inventory based on the attribute distribution of the unmatched out-of-stock equipment and adding it to the inventory.

[0024] It should be noted that at each inspection period of each inventory cycle, determining the equipment that is out of stock in the current cycle means that the system needs to regularly check the inventory status to identify those equipment that are currently or expected to be in short supply soon.

[0025] Specifically, the out-of-stock equipment can be determined by comparing the inventory level with a predetermined safety inventory level. For example, if the current inventory level of a certain equipment is lower than the set safety inventory level, then the equipment is considered out of stock. The safety inventory level can be set according to factors such as historical sales data, seasonal changes, and supplier delivery times.

[0026] Preferably, the system can adopt real-time monitoring technologies, such as barcode scanning or RFID systems, to update the inventory data in real time. In addition, an early warning system can be set up to automatically issue an alarm when the inventory level approaches the safety inventory level, so as to take timely actions. For difficult-to-predict out-of-stock situations, advanced prediction algorithms, such as time series analysis or machine learning models, can be used to improve the accuracy of prediction.

[0027] It should be noted that predicting the data of the out-of-stock equipment based on the historical data of the equipment in the inventory involves using information such as historical sales data, usage frequency, and maintenance records to predict the future demand and usage of the equipment.

[0028] Specifically, the prediction model can include machine learning algorithms such as linear regression, decision trees, and neural networks. The parameter settings should be adjusted according to the characteristics of the historical data. For example, if the data shows that the equipment demand has obvious seasonality, then the prediction model should be able to capture this seasonal change.

[0029] More specifically, the prediction model may include a feature selection step to determine which historical data features are most important for predicting the demand for out-of-stock devices. In addition, techniques such as cross-validation can be employed to optimize the model parameters and ensure that the model has good generalization ability.

[0030] It should be noted that based on the prediction data of out-of-stock devices, a data analysis method is used to match alternative devices for out-of-stock devices in the potential inventory, which involves finding devices with similar functions to the out-of-stock devices in the existing inventory or supplier catalog.

[0031] Specifically, the matching of alternative devices can be achieved through methods such as attribute comparison and similarity calculation. For example, metrics such as cosine similarity and Euclidean distance can be used to evaluate the similarity between devices.

[0032] Preferably, the matching process may include a user feedback link that allows operators to review and adjust the recommended alternative devices. In addition, factors such as the availability of the devices, the cost-benefit ratio, and the reputation of the suppliers can also be considered to ensure that the selected alternative devices are not only technically feasible but also economically optimal.

[0033] In some embodiments, predicting the data of out-of-stock devices based on the historical data of devices in the inventory includes: directly predicting the first type of attributes of the out-of-stock devices; for the second type of attributes of the out-of-stock devices, using a machine learning algorithm to match similar devices for the out-of-stock devices from the inventory devices, and predicting the second type of attributes of the out-of-stock devices based on the second type of attribute values of the similar devices; wherein, the first type of attributes are attributes that do not change with time or whose change rules are known although they change with time, and the second type of attributes are attributes that change with time and whose change rules are unknown.

[0034] It should be noted that predicting the data of out-of-stock devices based on the historical data of devices in the inventory involves using information such as historical sales data, usage frequency, and maintenance records to predict the future demand and usage of devices.

[0035] Specifically, the prediction model may include machine learning algorithms such as linear regression, decision trees, and neural networks. The parameter settings should be adjusted according to the characteristics of the historical data. For example, if the data shows that the device demand has obvious seasonality, then the prediction model should be able to capture this seasonal change. The first type of attributes may include attributes that do not change with time such as the specifications and models of the devices, while the second type of attributes may include attributes that change with time such as the usage frequency and the number of repairs of the devices.

[0036] Preferably, the prediction model may include a feature selection step to determine which historical data features are most important for predicting the demand of out-of-stock devices. In addition, techniques such as cross-validation can be employed to optimize the model parameters and ensure that the model has good generalization ability.

[0037] It should be noted that for the second type of attributes of out-of-stock devices, a machine learning algorithm is used to match similar devices from in-stock devices for out-of-stock devices, and the second type of attributes of out-of-stock devices are predicted based on the values of the second type of attributes of the similar devices.

[0038] Specifically, similar devices can be matched through methods such as attribute comparison and similarity calculation. For example, metrics such as cosine similarity and Euclidean distance can be used to evaluate the similarity between devices. The selection of similar devices can be based on multiple attributes, such as usage frequency, number of repairs, service life, etc.

[0039] More specifically, the matching process may include a user feedback session that allows operators to review and adjust the recommended similar devices. In addition, factors such as device availability, cost-benefit ratio, and supplier reputation can also be considered to ensure that the selected similar devices are not only technically feasible but also economically optimal.

[0040] It should be noted that the following formula is used to predict the second type of attributes of out-of-stock devices based on the attribute values of similar devices, which involves the application of a mathematical model to quantify the attribute relationship between similar devices and out-of-stock devices.

[0041] Specifically, the parameters in the formula may include the number of similar devices, the values of each similar device on the second type of attributes, etc. For example, the weighted average method can be used to calculate the predicted value, where the weights can be determined based on the similarity between the similar devices and the out-of-stock devices.

[0042] Preferably, the prediction formula can be further refined to include a time decay factor to reflect the impact of attribute values changing over time. In addition, multiple prediction models can be compared to select the best prediction method.

[0043] In some embodiments, the following formula is used to predict the second type of attributes of out-of-stock devices based on the attribute values of similar devices: ; where represents the predicted value of the i-th out-of-stock device on the k-th second type of attribute, represents the value of the m-th similar device on the k-th second type of attribute, and M represents the number of similar devices.

[0044] It should be noted that predicting the second - type attributes of out - of - stock devices based on the attribute values of similar devices using the following formula means that we use a mathematical model to quantify and predict those attributes that change over time and whose change patterns are unknown.

[0045] Specifically, the parameters in the formula can include the number of similar devices, the values of each similar device on the second - type attributes, etc. For example, the weighted - average method can be used to calculate the predicted value, where the weights can be determined based on the similarity between the similar devices and the out - of - stock devices. The similarity here can be calculated in various ways, such as cosine similarity, Euclidean distance, etc.

[0046] Preferably, the prediction formula can be further refined to include a time - decay factor to reflect the impact of the change of attribute values over time. In addition, multiple prediction models can be compared to select the best prediction method.

[0047] It should be noted that in the formula, represents the predicted value of the \(i\) - th out - of - stock device on the \(k\) - th second - type attribute, while represents the value of the \(m\) - th similar device on the \(k\) - th second - type attribute, and \(M\) represents the number of similar devices. This formula is a weighted - average calculation method, where the attribute values of each similar device contribute to the predicted value.

[0048] Specifically, the contribution of each similar device to the predicted value can be weighted by its similarity. The similarity can be calculated based on multiple factors, such as the usage frequency of the device, maintenance records, historical performance, etc. The setting of the weights can be adjusted according to the actual situation to ensure the accuracy of the prediction results.

[0049] More specifically, ensemble learning methods in machine learning, such as random forest or gradient - boosting tree, can be used to determine the weights of each similar device for the predicted value. In addition, the usage environment and usage conditions of the device can be considered to further improve the accuracy of the prediction.

[0050] In some embodiments, the tendency score of each device being out of stock in the current period is calculated based on the following formula, and \(M\) devices with the closest tendency scores to the out - of - stock device are selected from the in - stock devices as similar devices: ; where represents the tendency score of the \(i\) - th device being out of stock in the current period, which reflects the probability of the device being out of stock under given conditions, represents whether the \(i\) - th device is out of stock, represents that the \(i\) - th device is in an out - of - stock state, represents all the attribute values of the \(i\) - th device in the previous period, and \(P(\cdot)\) represents probability.

[0051] It should be noted that the propensity score for each device being out of stock in the current period is calculated based on the following formula, and M devices with the propensity scores closest to that of the out-of-stock device are selected from the in-stock devices as similar devices, which involves a quantitative assessment of the device's out-of-stock probability. This method helps to identify devices with similar usage patterns and out-of-stock risks, so as to find the best substitutes for out-of-stock devices.

[0052] Specifically, in the formula represents the propensity score of the i-th device being out of stock in the current period, represents all the attribute values of the i-th device in the previous period, and P(·) represents probability. The propensity score can be calculated through logistic regression, decision trees, or other classification algorithms, which can predict the likelihood of a device being out of stock based on the device's historical data.

[0053] Preferably, the calculation of the propensity score can include multiple factors, such as the usage frequency of the device, maintenance records, inventory levels, etc. In addition, the usage environment and usage conditions of the device can also be considered to further improve the accuracy of the prediction.

[0054] It should be noted that the formula for calculating the propensity score can be designed to reflect the level of the device's out-of-stock risk. The higher the score, the greater the likelihood of the device being out of stock in the current period.

[0055] Specifically, the propensity score can be calculated through the following steps: First, collect the historical data of the device, including sales records, usage frequency, maintenance records, etc.; then, use appropriate machine learning algorithms to train the model to predict the out-of-stock risk of the device; finally, calculate the propensity score of each device according to the output of the model.

[0056] More specifically, cross-validation can be used to optimize the model parameters to ensure that the model has good generalization ability. In addition, real-time data, such as the latest sales trends and market changes, can also be considered to dynamically adjust the calculation of the propensity score.

[0057] It should be noted that selecting M devices with the propensity scores closest to that of the out-of-stock device as similar devices helps to find devices that are similar to the target device in terms of usage pattern and out-of-stock risk.

[0058] Specifically, the selection of similar devices can be achieved by calculating the difference in propensity scores between devices. For example, the Euclidean distance or Manhattan distance can be used to measure the score difference between devices, and the M devices with the smallest score difference are selected.

[0059] Preferably, the selection process of similar devices may include a user review session, allowing management personnel to review and adjust the recommended similar devices. In addition, factors such as device availability, cost-benefit ratio, and supplier reputation can also be considered to ensure that the selected similar devices are not only technically feasible but also economically optimal.

[0060] In some embodiments, a data analysis method is used to match substitute devices for out-of-stock devices in the potential inventory, including: performing clustering analysis on each non-classified attribute of the inventory, while keeping the clustering of each classified attribute of the inventory unchanged; for each out-of-stock device, determining the cluster to which each attribute of the out-of-stock device belongs according to the predicted data of each attribute of the out-of-stock device; if there is a device in the potential inventory with the same distribution of cluster attributes as the out-of-stock device, the matching is successful, and a device with the same distribution of cluster attributes is randomly selected as the substitute device for the out-of-stock device.

[0061] Specifically, the data analysis method may include clustering analysis, classification analysis, or association rule learning, etc. Clustering analysis can help identify devices with similar attributes, while classification analysis can help predict whether a device is suitable as a substitute. The parameter settings of these methods need to be adjusted according to the specific attributes and historical data of the devices.

[0062] Preferably, the data analysis method can combine real-time data and prediction models to improve the accuracy of matching. For example, real-time inventory data and predicted device requirements can be used to dynamically adjust the parameters of clustering and classification.

[0063] It should be noted that performing clustering analysis on each non-classified attribute of the inventory while keeping the clustering of each classified attribute of the inventory unchanged means that when processing inventory data, different types of attributes need to be distinguished and appropriate methods need to be used for analysis. Non-classified attributes may include the usage frequency of the device, the number of repairs, etc., while classified attributes may include the type of the device, the brand, etc.

[0064] Specifically, clustering analysis can be performed by calculating the attribute similarity of the devices. For example, the K-means algorithm can be used to cluster the non-classified attributes, while the classified attributes can remain in their original classifications.

[0065] More specifically, more complex clustering methods such as hierarchical clustering or density clustering can be used to process high-dimensional data. In addition, the usage environment and usage conditions of the devices can also be considered to further improve the accuracy of clustering analysis.

[0066] It should be noted that for each out-of-stock device, determining the cluster to which each attribute of the out-of-stock device belongs according to the predicted data of each attribute of the out-of-stock device involves matching the attributes of the out-of-stock device with the attributes of the devices in the potential inventory.

[0067] Specifically, the matching can be performed by calculating the attribute similarity between the out-of-stock devices and each device in the potential inventory. For example, metrics such as cosine similarity or Euclidean distance can be used to evaluate the similarity between devices.

[0068] Preferably, the matching process may include a user feedback session, allowing the operator to review and adjust the recommended alternative devices. In addition, factors such as device availability, cost-benefit ratio, and supplier reputation can also be considered.

[0069] In some embodiments, the method further includes: if the current period is not the last period of the current cycle, then based on the predicted data of the out-of-stock devices accumulated in the current cycle, search for target new devices in the potential inventory, and add the target new devices to the potential inventory.

[0070] It should be noted that if the current period is not the last period of the current cycle, then based on the predicted data of the out-of-stock devices accumulated in the current cycle, search for target new devices in the potential inventory, and add the target new devices to the potential inventory. This sentence means that the system not only focuses on the current out-of-stock situation but also has foresight, capable of predicting future demands and preparing in advance.

[0071] Specifically, the system can analyze the attributes and demand patterns of all out-of-stock devices in the current cycle and use a prediction model to predict the types and quantities of devices that may be needed in the future. These predicted data can guide procurement decisions to ensure timely replenishment of inventory.

[0072] Preferably, the system can combine market trend analysis and historical sales data to optimize the prediction model. In addition, the delivery time and cost of suppliers can also be considered to determine the best procurement timing and quantity.

[0073] It should be noted that the process of searching for target new devices involves the evaluation and screening of the potential inventory. This requires a comprehensive evaluation mechanism to determine which devices are most likely to become future out-of-stock devices.

[0074] Specifically, the potential out-of-stock risk of devices can be evaluated by analyzing data such as device usage frequency, repair records, and customer feedback. In addition, the life cycle and replacement cycle of devices can also be considered to predict future demands.

[0075] More specifically, machine learning algorithms such as random forest or support vector machine can be used to evaluate and rank potential new devices. These algorithms can predict the demand probability of devices based on historical data and current trends.

[0076] It should be noted that adding target new devices to the potential inventory is a dynamic process that needs to be continuously adjusted according to new data and market changes. The system can set up a dynamic update mechanism to update the device list in the potential inventory regularly or in real time based on new sales data and market information.

[0077] Preferably, the system can provide a user interface that allows managers to manually adjust the device list in the potential inventory to account for special circumstances or urgent needs that may not be predicted by the algorithm. In addition, an early warning system can be set up to automatically alert managers to make inventory adjustments when a high stock - out risk is predicted.

[0078] It should be understood that in the present invention, predicting the data of out - of - stock devices based on the historical data of devices in the inventory is a key link in the automation of inventory management. Those skilled in the art can understand that predicting the data of out - of - stock devices can be achieved through various means, and these means are all based on in - depth analysis of historical data, aiming to improve the accuracy and efficiency of inventory management and reduce operating costs and losses caused by stock - outs.

[0079] For the first - type attributes of out - of - stock devices (i.e., attributes that do not change with time or change regularly over time and whose patterns are known, such as device specifications, models, initial installation parameters, etc.), those skilled in the art can use a variety of conventional methods for prediction. For example, directly consult the product manual or technical documentation of the device to obtain fixed attribute information, which remains basically unchanged throughout the life cycle of the device and can be used as the basic data for prediction. In addition, for some attributes that change slowly but regularly over time, such as the rated service life of the device, it can be calculated and predicted based on the used time and established decay patterns. In actual operation, if the specifications and models of the device are consistent with the records in the inventory management system, the information can be directly called for subsequent analysis; for attributes such as service life, the remaining service life can be calculated based on the usage duration of the device and a preset aging model to predict its availability in future cycles.

[0080] For the second type of attributes of out-of-stock devices (i.e., attributes that change over time and whose patterns are unknown, such as the actual usage frequency of the device, the number of failures, the performance degradation, etc.), the present invention provides a method for predicting the second type of attributes of out-of-stock devices by matching similar devices from in-stock devices based on machine learning algorithms. However, those skilled in the art should be aware that, in addition to this method, other technical means can also be used to achieve the same purpose. For example, based on time series analysis methods, models can be built for the historical usage frequency data of the device, such as autoregressive moving average model (ARMA) or autoregressive integrated moving average model (ARIMA), etc. By analyzing features such as trends, seasonality, and periodicity in the historical data, the future usage frequency can be predicted. At the same time, for the number of failures, a probability distribution model based on the historical failure interval time, such as the Weibull distribution model, can be constructed to estimate the probability and time of the next failure, so as to predict the failure risk of out-of-stock devices in the future cycle.

[0081] When matching similar devices based on machine learning algorithms, the present invention adopts a method of calculating the difference in propensity scores between devices (such as Euclidean distance or Manhattan distance) to select the M devices with the propensity scores closest to those of the out-of-stock device as similar devices. Those skilled in the art can adopt other effective similarity measurement methods according to the actual situation. For example, a distance measurement method based on information entropy can be used to measure similarity by calculating the difference between the information entropies of device attributes; or a density-based clustering method can be adopted to regard devices with similar density distributions as similar devices. In addition, when selecting similar devices, in addition to considering attributes such as the usage frequency, repair times, and service life of the device, factors such as the operating environment and workload of the device can also be incorporated to more comprehensively evaluate the similarity between devices.

[0082] For predicting the second type of attributes of out-of-stock devices based on the attribute values of similar devices, those skilled in the art can understand that in practical applications, this formula can be improved and optimized according to specific requirements. For example, a weight factor can be introduced to assign different weights according to the similarity of similar devices and out-of-stock devices in key attributes, so that the prediction result is more in line with the actual situation. In addition, in addition to the weighted average method, other statistics such as the median and mode can also be used to represent the attribute values of similar devices, so as to predict the attributes of out-of-stock devices. At the same time, during the prediction process, expert experience or industry standards can be combined to correct and adjust the prediction result to improve the accuracy of the prediction.

[0083] In actual inventory management scenarios, there are differences in equipment types, usage environments, and management requirements among different enterprises. Therefore, those skilled in the art can make adaptive adjustments to the prediction method according to specific circumstances. For example, for some enterprises with high requirements for equipment real-time performance, the frequency of real-time data collection and analysis can be increased to update the prediction model in a timely manner; for enterprises with a wide variety of equipment, a hierarchical prediction method can be adopted. First, group the equipment according to equipment categories or functions, and then formulate personalized prediction strategies for each group of equipment. In addition, with the development of technology and the accumulation of new data, those skilled in the art can continuously optimize the prediction model, introduce new algorithms and technologies, such as recurrent neural networks (RNN) or long short-term memory networks (LSTM) in deep learning algorithms, to better process time series data and improve the prediction ability for out-of-stock equipment data.

[0084] The above-mentioned embodiments of the present invention have the following beneficial effects: The automated equipment inventory management method based on data analysis of the present invention can systematically handle the problem of equipment out-of-stock. This method can automatically determine out-of-stock equipment during the inspection period of each inventory cycle, and predict the attribute changes of these equipment through historical data analysis, so as to prepare alternative equipment in advance. This method can reduce production interruptions caused by equipment out-of-stock, and at the same time improve the scientific nature of inventory management and the accuracy of prediction through data analysis.

[0085] Furthermore, this method matches similar equipment through machine learning algorithms and predicts the attributes of out-of-stock equipment based on the attributes of similar equipment, which can more accurately find suitable substitutes for out-of-stock equipment. In addition, this method also includes finding equipment in the potential inventory that has the same attribute distribution as the out-of-stock equipment, which can improve the matching efficiency of alternative equipment and the utilization rate of inventory. Through these measures, enterprises can more effectively manage inventory, reduce inventory backlogs, lower costs, and improve the response speed to market changes.

[0086] As Figure 2 shown, an automated equipment inventory management system 200 based on data analysis in some embodiments, the system 200 includes: an out-of-stock equipment determination module 201, configured to determine the equipment out-of-stock in the current cycle during each inspection period of each inventory cycle; a data prediction module 202, configured to predict the data of out-of-stock equipment based on the historical data of the equipment in the inventory; an alternative equipment matching module 203, configured to match alternative equipment for out-of-stock equipment in the potential inventory by using a data analysis method based on the predicted data of the out-of-stock equipment, and add the successfully matched alternative equipment to the inventory; an attribute distribution matching module 204, configured to, if there are un-matched out-of-stock equipment, determine the attribute distribution of the un-matched out-of-stock equipment based on the predicted data of the un-matched out-of-stock equipment, and find alternative equipment for the un-matched out-of-stock equipment in the potential inventory based on the attribute distribution of the un-matched out-of-stock equipment and add it to the inventory.

[0087] It can be understood that the various modules recorded in the device inventory automated management system 200 based on data analysis correspond to the respective steps in the method for automated management of device inventory based on data analysis described in the reference. Figure 1 Thus, the operations, features, and beneficial effects described above for the method for automated management of device inventory based on data analysis are equally applicable to the device inventory automated management system 200 based on data analysis and the modules included therein, and will not be elaborated herein.

[0088] In some embodiments, predicting data of out-of-stock devices based on historical data of devices in the inventory includes: directly predicting the first type of attributes of the out-of-stock devices; for the second type of attributes of the out-of-stock devices, using a machine learning algorithm to match similar devices for the out-of-stock devices from the inventory devices, and predicting the second type of attributes of the out-of-stock devices based on the second type of attribute values of the similar devices; wherein, the first type of attributes are attributes that do not change with time or whose change pattern is known although they change with time, and the second type of attributes are attributes that change with time and whose change pattern is unknown.

[0089] It should be noted that predicting data of out-of-stock devices based on historical data of devices in the inventory includes directly predicting the first type of attributes of the out-of-stock devices; for the second type of attributes of the out-of-stock devices, using a machine learning algorithm to match similar devices for the out-of-stock devices from the inventory devices, and predicting the second type of attributes of the out-of-stock devices based on the second type of attribute values of the similar devices.

[0090] Specifically, the first type of attributes may include attributes such as the specifications, models, service lives, etc. of the devices that do not change with time or whose change pattern is known. The second type of attributes may include attributes such as usage frequency, failure rate, etc. that change with time and whose change pattern is unknown. Direct prediction can use methods such as time series analysis, linear regression, etc., while indirect prediction requires first finding similar devices through methods such as clustering, classification, etc., and then predicting based on the attributes of these devices.

[0091] Preferably, the prediction model can combine multiple machine learning algorithms, such as support vector machines, random forests, neural networks, etc., to improve the accuracy and robustness of the prediction. In addition, ensemble learning methods, such as model fusion or stacking, can also be used to synthesize the prediction results of different models.

[0092] It should be noted that the following formula is used to predict the second type of attributes of out-of-stock devices based on the attribute values corresponding to similar devices. This formula is a mathematical expression used to quantify the attribute relationship between similar devices and out-of-stock devices and make predictions accordingly.

[0093] Specifically, the parameters in the formula need to be set according to the actual data. For example, if the weighted average method is used, the weights can be determined based on the similarity between similar devices and out-of-stock devices. Methods such as cosine similarity and Euclidean distance can be used to calculate the similarity.

[0094] More specifically, an adaptive weight adjustment mechanism can be used to dynamically adjust the weights according to the accuracy of historical predictions, so as to improve the adaptability and accuracy of the prediction model. In addition, a regularization term can be introduced to avoid overfitting and improve the generalization ability of the model.

[0095] It should be noted that this method also includes continuous optimization and adjustment of the prediction model to adapt to changing data and environmental conditions. Through continuous learning and optimization, the system can improve its ability to predict the attributes of out-of-stock devices, thereby more effectively managing inventory.

[0096] Specifically, the model can be continuously updated through online learning or incremental learning methods to reflect the latest data trends. In addition, model evaluation and parameter adjustment can be performed regularly to ensure the accuracy of the prediction results.

[0097] Preferably, the system can provide a feedback mechanism that allows users to evaluate and provide feedback on the prediction results, and these feedbacks can be used for further training and optimization of the model. In addition, expert knowledge, such as the experience judgment of maintenance personnel, can be introduced to assist the prediction and decision-making process of the model.

[0098] In some embodiments, the following formula is used to predict the second type of attribute of the out-of-stock device based on the attribute values corresponding to the similar devices: ; where represents the predicted value of the i-th out-of-stock device on the k-th second type of attribute, represents the value of the m-th similar device on the k-th second type of attribute, and M represents the number of similar devices.

[0099] It should be noted that the following formula is used to predict the second type of attribute of the out-of-stock device based on the attribute values corresponding to the similar devices, which involves a specific mathematical model for predicting the attributes of out-of-stock devices based on the known attributes of similar devices. This method can improve the accuracy of the prediction because it utilizes the actual usage data of the existing devices in the inventory.

[0100] Specifically, in the formula represents the predicted value of the i-th out-of-stock device on the k-th second type of attribute, while represents the value of the m-th similar device on the k-th second type of attribute, and M represents the number of similar devices. This formula is a weighted average calculation method, where the attribute values of each similar device contribute to the predicted value.

[0101] Preferably, the prediction formula can be further refined to include a time decay factor to reflect the impact of the change of attribute values over time. In addition, multiple prediction models can be adopted for comparison to select the best prediction method.

[0102] It should be noted that the denominator part in the formula, that is, the expression under the summation symbol, represents the weighted summation of the k-th attribute values of all similar devices. This weight can be determined based on the similarity between the similar device and the out-of-stock device. The higher the similarity, the greater the weight.

[0103] Specifically, multiple methods can be used to calculate the similarity, such as cosine similarity, Euclidean distance, etc. The weight can also be adjusted according to other factors such as the usage frequency of the device, maintenance records, historical performance, etc.

[0104] More specifically, ensemble learning methods in machine learning, such as random forest or gradient boosting tree, can be adopted to determine the weight of each similar device on the predicted value. In addition, the usage environment and usage conditions of the device can be considered to further improve the accuracy of the prediction.

[0105] Specifically, the system can set up a dynamic update mechanism to update the prediction model regularly or in real time according to new sales data and market information. The system can provide a user interface that allows managers to manually adjust the parameters of the prediction model to consider those special situations or emergency needs that may not be predicted by the algorithm. In addition, an early warning system can be set up to automatically remind managers to adjust the inventory when a high out-of-stock risk is predicted.

[0106] In some embodiments, the system further includes: a policy adjustment module, configured to, if the current time period is not the last time period of the current cycle, search for target new devices in the potential inventory based on the predicted data of the out-of-stock devices accumulated in the current cycle, and add the target new devices to the potential inventory.

[0107] It should be noted that the system further includes a policy adjustment module, configured to, if the current time period is not the last time period of the current cycle, search for target new devices in the potential inventory based on the predicted data of the out-of-stock devices accumulated in the current cycle, and add the target new devices to the potential inventory.

[0108] Specifically, the policy adjustment module can be a software component that uses historical data and real-time data to predict future device requirements. This module can analyze the usage situation and out-of-stock records of all devices in the current cycle, so as to predict which devices may become out-of-stock devices in the future.

[0109] Preferably, the strategy adjustment module can integrate advanced data analysis tools and prediction algorithms, such as time series analysis, machine learning models, etc., to improve the accuracy of prediction. In addition, market trends, seasonal changes, and other external factors can also be considered to optimize the prediction results.

[0110] It should be noted that the working principle of the strategy adjustment module is to identify potential new equipment requirements based on prediction data. This involves in-depth analysis of equipment usage patterns and prediction of future market changes.

[0111] Specifically, the module can set a series of thresholds and parameters, such as prediction confidence, importance level of equipment, etc., to determine which equipment should be regarded as target new equipment. These parameters can be set according to historical data and business rules.

[0112] More specifically, the strategy adjustment module can adopt an adaptive learning algorithm to continuously adjust the prediction model based on past prediction results and actual inventory situations. In addition, a user intervention mechanism can also be set up to allow managers to manually adjust the prediction results according to business needs.

[0113] Specifically, the system can set up a dynamic update mechanism to update the list of equipment in the potential inventory regularly or in real time according to new sales data and market information. The system can provide a user interface to allow managers to manually adjust the list of equipment in the potential inventory to consider special situations or emergency needs that may not be predicted by the algorithm. In addition, an early warning system can also be set up to automatically remind managers to adjust the inventory when a high out-of-stock risk is predicted.

[0114] Next, refer to Figure 3 , which shows a schematic structural diagram of the structure 300 of an electronic device suitable for implementing some embodiments of the present invention. The electronic devices in some embodiments of the present invention may include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 3 The terminal devices shown are only examples and should not impose any limitations on the functions and usage scopes of the embodiments of the present invention.

[0115] As Figure 3As shown, the electronic device 300 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 301, which may perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage device 308 into the random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other through a bus 304. The input / output (I / O) interface 305 is also connected to the bus 304.

[0116] Generally, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 3 the electronic device 300 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. More or fewer devices may be alternatively implemented or had. Figure 3 Each block shown in may represent a device or, as needed, multiple devices.

[0117] Furthermore, the storage medium of the embodiment of the present application stores program instructions capable of implementing all the above methods. Among them, the program instructions may be stored in the above storage medium in the form of a software product, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the methods described in various embodiments of the present application. And the foregoing storage medium includes: various media capable of storing program codes such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc, or a terminal device such as a computer, a server, a mobile phone, a tablet, etc.

[0118] The above description is only some preferred embodiments of the present invention and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present invention is not limited to the technical solution formed by the specific combination of the above technical features.

Claims

1. A method for automated equipment inventory management based on data analysis, characterized in that: The following steps are involved: During each inspection period of each inventory cycle, identify the equipment that is out of stock during the current cycle; Forecasting of out-of-stock equipment based on historical data of equipment in inventory; Based on the forecast data of out-of-stock equipment, data analysis methods are used to match substitute equipment for out-of-stock equipment in potential inventory, and the successfully matched substitute equipment is added to the inventory; If there is an unmatched out-of-stock device, the attribute distribution of the unmatched out-of-stock device is determined based on the forecast data of the unmatched out-of-stock device, and a replacement device of the unmatched out-of-stock device is searched in the potential inventory based on the attribute distribution of the unmatched out-of-stock device to be added to the inventory; Forecasting of out-of-stock equipment data based on historical data of equipment in inventory, including: For the first type of attributes of out-of-stock equipment, forecasts are made directly; For the second type of attribute of the out-of-stock device, a machine learning algorithm is used to match similar devices to the out-of-stock device from the inventory devices, and the second type of attribute of the out-of-stock device is predicted based on the second type of attribute values ​​of the similar devices; The first type of attribute is an attribute that does not change with time or changes with time but the change pattern is known, and the second type of attribute is an attribute that changes with time but the change pattern is unknown; The following formula is used to predict the second type of attribute of out-of-stock equipment based on the corresponding attribute values ​​of similar equipment: in, represents the predicted value of the kth second-type attribute for the i-th out-of-stock equipment, represents the value of the mth similar device on the kth second type attribute, and M represents the number of similar devices; Use data analysis methods to match out-of-stock equipment with replacement equipment in potential inventory, including: Cluster analysis is performed for each non-categorical attribute of the inventory, and the clustering remains unchanged for each categorical attribute of the inventory; For each out-of-stock device, determine the cluster to which each attribute of the out-of-stock device belongs according to the prediction data of each attribute of the out-of-stock device; If there is a device in the potential inventory with the same clustering attribute distribution as the out-of-stock device, the match is successful, and a device with the same clustering attribute distribution is randomly selected as the replacement device for the out-of-stock device.

2. The method for automated equipment inventory management based on data analysis according to claim 1, characterized in that: The out-of-stock propensity score of each device in the current cycle is calculated based on the following formula, and M devices with the closest propensity scores to the out-of-stock devices are selected from the inventory devices as similar devices: TendencyScore i =P(OutStock i =1|AttributeValues i,t-1 ) Among them, TendencyScore i Indicates the propensity score of the i-th device being out of stock in the current cycle, reflecting the probability of the device being out of stock under given conditions. i Indicates whether the i-th device is out of stock, OutStock i =1 means the i-th device is out of stock, AttributeValues i,t-1 represents all attribute values ​​of the ith device in the previous cycle, and P(·) represents the probability.

3. The method for automated equipment inventory management based on data analysis according to claim 1, further comprising: If the current time period is not the last time period of the current cycle, the target new equipment is searched in the potential inventory based on the forecast data of the out-of-stock equipment accumulated in the current cycle, and the target new equipment is added to the potential inventory.

4. An automated equipment inventory management system based on data analysis, the system implementing the method according to claim 1, characterized in that: Includes the following modules: An out-of-stock equipment determination module is used to determine the out-of-stock equipment in the current cycle in each inspection period of each inventory cycle; A data forecasting module is used to forecast the data of out-of-stock equipment based on the historical data of equipment in inventory; The substitute equipment matching module is used to match substitute equipment for out-of-stock equipment in potential inventory based on the forecast data of out-of-stock equipment by using data analysis methods, and add the successfully matched substitute equipment to the inventory; The attribute distribution matching module is used to determine the attribute distribution of the unmatched out-of-stock equipment based on the prediction data of the unmatched out-of-stock equipment if there is an unmatched out-of-stock equipment, and to search for a replacement device for the unmatched out-of-stock equipment in the potential inventory based on the attribute distribution of the unmatched out-of-stock equipment to add to the inventory.

5. The equipment inventory automation management system based on data analysis according to claim 4 is characterized in that: Forecasting of out-of-stock equipment data based on historical data of equipment in inventory, including: For the first type of attributes of out-of-stock equipment, forecasts are made directly; For the second type of attribute of the out-of-stock device, a machine learning algorithm is used to match similar devices to the out-of-stock device from the inventory devices, and the second type of attribute of the out-of-stock device is predicted based on the second type of attribute values ​​of the similar devices; The first type of attribute is an attribute that does not change with time or changes with time but the change pattern is known, and the second type of attribute is an attribute that changes with time but the change pattern is unknown.

6. The equipment inventory automation management system based on data analysis according to claim 5 is characterized in that: The following formula is used to predict the second type of attribute of out-of-stock equipment based on the corresponding attribute values ​​of similar equipment: in, represents the predicted value of the kth second-type attribute for the i-th out-of-stock equipment, represents the value of the mth similar device on the kth second type attribute, and M represents the number of similar devices.

7. The equipment inventory automation management system based on data analysis according to claim 6, wherein the system further comprises: The strategy adjustment module is used to search for target new equipment in the potential inventory based on the forecast data of out-of-stock equipment accumulated in the current cycle if the current period is not the last period of the current cycle, and add the target new equipment to the potential inventory.

Images