Internet of Things equipment sales management system based on artificial intelligence

Abstract

The invention discloses an Internet of Things equipment sales management system based on artificial intelligence, and relates to the technical field of sales management. Comprising a monitoring center which is in signal connection with a data acquisition module, a data processing module, a data analysis module and a data feedback module. The data acquisition module is used for acquiring image data of basic information of equipment; the data processing module is used for carrying out data preprocessing on the collected image data; the data analysis module performs unsalable condition screening on data sets of N groups of equipment categories by adopting an isolated forest algorithm; and the data feedback module is based on feedback information of the equipment unsalable condition, and an enterprise adopts a targeted strategy to adjust the feedback information, so that the problem of inventory overstock is effectively solved, and the occupation of capital cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of sales management, and in particular to an Internet of Things equipment sales management system based on artificial intelligence. Background Art

[0002] With the expansion of enterprise scale and the enrichment of product lines, inventory management has become more and more complicated. Once the inventory management is improper or the old equipment in the inventory is not cleared in time, the equipment will be unsalable. When searching for specific unsalable equipment products among a large number of equipment products, it will not only take up a lot of search time, but also increase the capital occupation cost and maintenance cost of the enterprise.

[0003] In the past, companies have mostly used sales management systems to monitor the sales data of each device in real time, and based on the changing trends of sales data, used sales analysis tools to predict the sales of the device in the future, and thus inferred whether the device would be unsalable. However, sales data and inventory data in such management systems often lag behind and cannot reflect market changes in a timely manner. Moreover, if the data is inaccurate or the analysis model is flawed, these will lead to misjudgment of unsalable conditions, causing companies to miss the best time to deal with unsalable equipment when they discover that the equipment is unsalable.

[0004] Therefore, how to quickly find specific unsalable equipment among a large number of equipment products, solve the problem of equipment inventory backlog, and protect the interests of the enterprise is a problem we need to solve. To this end, an IoT equipment sales management system based on artificial intelligence is now provided. Summary of the invention

[0005] In order to solve the existing technical problems, the purpose of the present invention is to provide an Internet of Things equipment sales management system based on artificial intelligence.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an artificial intelligence-based IoT device sales management system, comprising a monitoring center, wherein the monitoring center is signal-connected with a data acquisition module, a data processing module and a data feedback module; The data acquisition module is used to obtain image data of basic information of IoT devices; The data processing module is used to perform data preprocessing on the collected image data; The data analysis module analyzes N groups of equipment category data using an isolation forest algorithm, isolates abnormal equipment data using segmentation points for equipment data points, and implements screening of unsalable equipment; The data feedback module is based on the feedback information of the unsalable equipment, and the enterprise adopts targeted strategies to make adjustments.

[0007] Preferably, the process of the data acquisition module for acquiring image data of basic information of the device includes: The image data of the basic information of the equipment includes a list of equipment and sample pictures of equipment sample pictures; The equipment list includes equipment model, unit, unit price, amount, monthly sales volume and inventory volume; The data acquisition module is selected by the enterprise as an image sensor that supports WiFi connection, and the enterprise downloads and installs a dedicated APP provided by the image sensor manufacturer in the mobile device; After opening the APP, follow the prompts to add the image sensor device, configure the WiFi connection of the image sensor in the APP, and ensure that the image sensor is connected to the same WiFi network as the mobile device; after a successful connection, the enterprise can receive the remotely transmitted image data in the APP and automatically store it in the database.

[0008] Preferably, the process of the data processing module for performing data preprocessing on the collected image data includes: The equipment list image data is identified using OCR technology; The sample images of the device models are classified for all device models through the target detection model; according to the classification results, the device categories, monthly sales volume and inventory data are collected together to establish N sets of data sets.

[0009] Preferably, the process of using OCR technology to identify the image data of the device list in the data processing module includes: S1. Image enhancement: Using the histogram equalization method, the histogram of the original image is converted into an image with a probability density of 1 through the integral probability density function, so that the grayscale distribution of the image is more uniform, the details and contrast of the image are improved, and the text and data are more prominent and easy to identify; S2, OCR recognition: using the OCR model to recognize all text information in the list image, and preliminarily organizing the text data recognized by OCR in a table format for subsequent data processing; S3, data cleaning: further clean the formatted text data, remove the text data of the unit, unit price and amount in the equipment list, and remove the repeated parts in the remaining data information; S4. Create a data table: organize the cleaned data into a table, and fill all the data into the Excel table according to the format of three columns of information: equipment model, monthly sales volume and inventory volume.

[0010] Preferably, the target detection model includes an image input layer, a feature extraction layer, a weight distribution layer, a feature fusion layer and an image classification layer.

[0011] Preferably, the process of classifying the sample images of the device model by the target detection model includes: Image input layer: As the entry point of the model, this layer is responsible for receiving image data of device model samples; Feature extraction layer: This layer automatically extracts key feature information from the input image, usually the local shape, texture, and color distribution in the image that are critical for device classification; Weight assignment layer: Based on the image features output by the feature extraction layer, the corresponding weight value is calculated according to the image features of each sample image, and the calculated weight value is used to assign weights to the image features of each sample image; The feature fusion layer performs weighted addition of the image features of each sample image, and performs feature splicing on the weighted image features to generate a comprehensive and highly discriminative image feature, namely, a comprehensive image feature; Image classification layer: Based on the comprehensive image features, this layer inputs the features into the DNN network for classification, determines the device category to which each device sample image belongs, and divides them into N groups of data sets.

[0012] Preferably, the equipment unsalable situation in the data analysis module includes too little equipment sales and too much equipment inventory backlog.

[0013] Preferably, the process of the data analysis module using the isolation forest algorithm to screen the data sets of N groups of equipment categories for equipment unsalability includes: First, use N sets of data sets to pre-train N isolation forest models, and then use the trained isolation forest models to filter abnormal data; The training process of the isolation forest model includes: B1. Model construction: 100 base learners are used to build the isolation forest model, and various hyperparameter information of the model is configured; B2. Feature selection: Select equipment category, monthly sales volume and inventory as the three main features for model training; B3. Data standardization: normalize the selected features to form a standard data set; B4. Training the isolation forest model: Use the processed data set to train the isolation forest model. Use the isolation forest algorithm to train the standard data set to obtain the model results, and perform data mining on the screened data. The process of using the trained isolation forest model to screen out slow-selling devices includes: C1. Calculate the anomaly score: After the model training is completed, input the model, monthly sales volume and inventory of a certain device into the model, and then the anomaly score value of the device model can be directly calculated. The higher the value, the more serious the unsalable situation of the device model; C2. Set anomaly score threshold: Set a slow-selling threshold based on the actual situation to distinguish which device models are truly slow-selling products. When the anomaly score is greater than or equal to the threshold, it means that the device model has a slow-selling problem. When the anomaly score is less than the threshold, it means that the device model does not have a slow-selling problem and is in a normal state, and there is no need to adjust the strategy. C3. Repeat the above steps S1-S6 N times until all the unsalable equipment models in N equipment categories are screened out. Preferably, the process in which the enterprise adopts targeted strategies to make adjustments based on the feedback information of the slow-selling situation in the data feedback module includes: After the company collects feedback information on all slow-selling equipment models in N equipment categories, it quickly adjusts its strategy.

[0014] The present invention provides an Internet of Things equipment sales management system based on artificial intelligence, which has the following beneficial effects: image data collection of basic information of all equipment models, data preprocessing of the collected image data, wherein sample images of various equipment models are classified through a target detection model, and the preprocessed N groups of data sets are screened for unsalable products using an isolation forest algorithm. Based on feedback information, enterprises can quickly lock in specific equipment models in the equipment category to which they belong and adjust strategies, which greatly shortens the search time, effectively improves equipment sales efficiency and inventory backlog problems, improves work efficiency, and protects the interests of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of an IoT device sales management system based on artificial intelligence provided in an embodiment of the present application; Figure 2 A flow chart of a method for recognizing device list image data using OCR technology provided in an embodiment of the present application; Figure 3 A flow chart of a method for training an isolation forest model to screen for unsalable specific equipment provided in an embodiment of the present application. DETAILED DESCRIPTION

[0016] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0017] like Figure 1 As shown, an IoT device sales management system based on artificial intelligence includes a monitoring center, characterized in that the monitoring center signal is connected to a data acquisition module, a data processing module and a data feedback module; The data acquisition module is used to obtain image data of basic information of the device; The data processing module is used to perform data preprocessing on the collected image data; The data analysis module uses an isolation forest algorithm to screen the data sets of N groups of equipment categories for equipment unsalability; The data feedback module is based on the feedback information of the unsalable equipment, and the enterprise adopts targeted strategies to make adjustments; It should be further explained that, in a specific implementation process, the process of the data acquisition module for acquiring image data of basic information of the device includes: The image data of the basic information of the equipment includes a list of equipment and sample pictures of equipment models; The equipment list includes equipment model, unit, unit price, amount, monthly sales volume and inventory volume; The data acquisition module is that the enterprise selects an image sensor that supports WiFi connection, downloads and installs a dedicated application provided by the image sensor manufacturer, namely an APP, in the mobile device; after opening the APP, follow the prompts to add the image sensor device, configure the WiFi connection of the image sensor in the APP, and ensure that the image sensor is connected to the same WiFi network as the mobile device; after the connection is successful, the enterprise can receive the remotely transmitted image data in the APP and automatically store it in the database; It should be further explained that, in the specific implementation process, the process of the data processing module for performing data preprocessing on the collected image data includes: The equipment list image data is identified using OCR technology; The sample images of the device models are used to classify all device models through the target detection model; according to the classification results, the device categories, monthly sales volume and inventory volume data are collected together to establish N data sets.

[0018] It should be further explained that, in the specific implementation process, the process of using OCR technology to recognize the image data of the device list in the data processing module includes: S11 Image Enhancement: Using the histogram equalization method, the histogram of the original image is converted into an image with a probability density of 1 through the integral probability density function, so that the grayscale distribution of the image is more uniform, the details and contrast of the image are improved, and the text and data are more prominent and easy to identify; S12 OCR recognition: Use the OCR model to recognize all text information in the list image, and preliminarily organize the text data recognized by OCR in a table format for subsequent data processing; S13 data cleaning: further cleaning the formatted text data, removing the text data of the unit, unit price and amount in the equipment list, and removing the repeated parts in the remaining data information; S14 creates a data table: organize the cleaned data into a table, and fill all the data into the Excel table according to the format of three columns of information: equipment model, monthly sales volume and inventory volume.

[0019] Furthermore, the target detection model includes an image input layer, a feature extraction layer, a weight distribution layer, a feature fusion layer, and an image classification layer; It should be further explained that, in the specific implementation process, the process of classifying the sample images of the device model through the target detection model includes: Image input layer: As the entry point of the model, this layer is responsible for receiving image data of device model samples; Feature extraction layer: This layer automatically extracts key feature information from the input image, usually the local shape, texture, and color distribution in the image that are critical for device classification; Weight assignment layer: Based on the image features output by the feature extraction layer, the corresponding weight value is calculated according to the image features of each sample image, and the calculated weight value is used to assign weights to the image features of each sample image; The feature fusion layer performs weighted addition of the image features of each sample image, and performs feature splicing on the weighted image features to generate a comprehensive and highly discriminative image feature, namely, a comprehensive image feature; Image classification layer: Based on comprehensive image features, this layer inputs the features into the DNN network for classification, determines the device category to which each device sample image belongs, and divides it into N groups of data sets; Furthermore, the equipment unsalable situation in the data analysis module includes too little equipment sales and too much equipment inventory backlog; Furthermore, the process of the data analysis module using the isolation forest algorithm to screen the data sets of N groups of equipment categories for equipment unsalability includes: First, use N sets of data sets to pre-train N isolation forest models, and then use the trained isolation forest models to filter abnormal data; The training process of the isolation forest model includes: S21 model construction: 100 base learners are used to build the isolation forest model, and various hyperparameter information of the model is configured; S22 Feature selection: Equipment category, monthly sales volume and inventory volume are selected as the three main features for model training; S23 Data standardization: normalize the selected features to form a standard data set; S24 training isolation forest model: use the processed data set to train the isolation forest model, use the isolation forest algorithm to train the standard data set to obtain the model result, and perform data mining on the screened data; The process of using the trained isolation forest model to screen out slow-selling devices includes: S25 Calculate anomaly score: After the model training is completed, input the model, monthly sales volume and inventory of a certain device into the model, and then the anomaly score value of the device model can be directly calculated. The higher the value, the more serious the unsalable situation of the device model; S26 sets an abnormal score threshold: according to the actual situation, a slow-selling threshold is set to distinguish which equipment models are truly slow-selling products; if the abnormal score is greater than or equal to the threshold, it means that the equipment model has a slow-selling problem; if the abnormal score is less than the threshold, it means that the equipment model does not have a slow-selling problem and is in a normal state, and no strategy adjustment is required; Repeat the above steps S1-S6 N times until the unsalable equipment models in N equipment categories are screened out; It should be further explained that, in the specific implementation process, the feedback information of the slow-selling equipment in the data feedback module, the method for the enterprise to adopt targeted strategies to make adjustments includes: After the company collects feedback information on all slow-selling equipment models in N equipment categories, it can check the causes and quickly adjust the strategy to solve the slow-selling problem.

[0020] For example: If equipment is unsalable due to a decline in market demand or a change in consumer preferences, the company can adjust the functions of the equipment products in a timely manner according to market demand to meet the actual needs of consumers; if the equipment is priced unreasonably or too high to discourage consumers, the company can adopt a sales strategy to highlight the advantages of the equipment model, use discounts and promotions to attract consumers, stimulate buyers' desire to buy, and effectively improve sales efficiency; if the equipment has quality problems such as unstable performance and high failure rate, which seriously affect consumers' willingness to buy, the company must take measures to improve product quality, increase R&D investment, strengthen equipment quality control and testing, and ensure stable and reliable product performance.

[0021] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An IoT device sales management system based on artificial intelligence, including a monitoring center, characterized in that: The monitoring center is signal-connected with a data acquisition module, a data processing module and a data feedback module; The data acquisition module is used to obtain image data of basic information of IoT devices; The data processing module is used to perform data preprocessing on the collected image data; The data analysis module analyzes N groups of equipment category data using an isolation forest algorithm, isolates abnormal equipment data using segmentation points for equipment data points, and implements screening of unsalable equipment; The data feedback module is based on the feedback information of the unsalable equipment, and the enterprise adopts targeted strategies to make adjustments.

2. According to the artificial intelligence-based Internet of Things equipment sales management system according to claim 1, it is characterized in that: The process of the data acquisition module for acquiring image data of basic information of the device includes: The image data of the basic information of the equipment includes a list of equipment and sample pictures of equipment sample pictures; The equipment list includes equipment model, unit, unit price, amount, monthly sales volume and inventory volume; The data acquisition module is selected by the enterprise as an image sensor that supports WiFi connection, and the enterprise downloads and installs a dedicated APP provided by the image sensor manufacturer in the mobile device; After opening the APP, follow the prompts to add the image sensor device, configure the WiFi connection of the image sensor in the APP, and ensure that the image sensor is connected to the same WiFi network as the mobile device; after a successful connection, the enterprise can receive the remotely transmitted image data in the APP and automatically store it in the database.

3. According to claim 1, an artificial intelligence-based IoT device management system is characterized in that: The process of the data processing module for performing data preprocessing on the collected image data includes: The equipment list image data is identified using OCR technology; The sample images of the device models are classified for all device models through the target detection model; according to the classification results, the device categories, monthly sales volume and inventory data are collected together to establish N sets of data sets.

4. The artificial intelligence-based IoT device management system according to claim 3 is characterized in that: The process of using OCR technology to recognize the image data of the equipment list in the data processing module includes: S1. Image enhancement: Using the histogram equalization method, the histogram of the original image is converted into an image with a probability density of 1 through the integral probability density function, so that the grayscale distribution of the image is more uniform, the details and contrast of the image are improved, and the text and data are more prominent and easy to identify; S2, OCR recognition: using the OCR model to recognize all text information in the list image, and preliminarily organizing the text data recognized by OCR in a table format for subsequent data processing; S3, data cleaning: further clean the formatted text data, remove the text data of the unit, unit price and amount in the equipment list, and remove the repeated parts in the remaining data information; S4. Create a data table: organize the cleaned data into a table, and fill all the data into the Excel table according to the format of three columns of information: equipment model, monthly sales volume and inventory volume.

5. According to the artificial intelligence-based Internet of Things equipment sales management system of claim 3, it is characterized in that: The target detection model includes an image input layer, a feature extraction layer, a weight allocation layer, a feature fusion layer and an image classification layer.

6. The IoT device sales management system based on artificial intelligence according to claim 3 is characterized in that: The process of classifying sample images of the device model through the object detection model includes: Image input layer: As the entry point of the model, this layer is responsible for receiving image data of device model samples; Feature extraction layer: This layer automatically extracts key feature information from the input image, usually the local shape, texture, and color distribution in the image that are critical for device classification; Weight assignment layer: Based on the image features output by the feature extraction layer, the corresponding weight value is calculated according to the image features of each sample image, and the calculated weight value is used to assign weights to the image features of each sample image; The feature fusion layer performs weighted addition of the image features of each sample image, and performs feature splicing on the weighted image features to generate a comprehensive and highly discriminative image feature, namely, a comprehensive image feature; Image classification layer: Based on the comprehensive image features, this layer inputs the features into the DNN network for classification, determines the device category to which each device sample image belongs, and divides them into N groups of data sets.

7. The IoT device sales management system based on artificial intelligence according to claim 1 is characterized in that: The equipment unsalable situation in the data analysis module includes too little equipment sales and too much equipment inventory backlog.

8. The IoT device sales management system based on artificial intelligence according to claim 1 is characterized in that: The process of the data analysis module using the isolation forest algorithm to screen the data sets of N groups of equipment categories for equipment unsalability includes: First, use N sets of data sets to pre-train N isolation forest models, and then use the trained isolation forest models to filter abnormal data; The training process of the isolation forest model includes: B1. Model construction: 100 base learners are used to build the isolation forest model, and various hyperparameter information of the model is configured; B2. Feature selection: Select equipment category, monthly sales volume and inventory as the three main features for model training; B3. Data standardization: normalize the selected features to form a standard data set; B4. Training the isolation forest model: Use the processed data set to train the isolation forest model. Use the isolation forest algorithm to train the standard data set to obtain the model results, and perform data mining on the screened data. The process of using the trained isolation forest model to screen out slow-selling devices includes: C1. Calculate the anomaly score: After the model training is completed, input the model, monthly sales volume and inventory of a certain device into the model, and then the anomaly score value of the device model can be directly calculated. The higher the value, the more serious the unsalable situation of the device model; C2. Set anomaly score threshold: Set a slow-selling threshold based on the actual situation to distinguish which device models are truly slow-selling products. When the anomaly score is greater than or equal to the threshold, it means that the device model has a slow-selling problem. When the anomaly score is less than the threshold, it means that the device model does not have a slow-selling problem and is in a normal state, and there is no need to adjust the strategy. C3. Repeat the above steps S1-S6 N times until the unsalable equipment models in the N equipment categories are screened out.

9. The IoT device sales management system based on artificial intelligence according to claim 1 is characterized in that: The process of the enterprise taking targeted strategies to adjust the feedback information based on the unsalable situation in the data feedback module includes: After the company collects feedback information on all slow-selling equipment models in N equipment categories, it quickly adjusts its strategy.

Images