Item Package Quantity Prediction Model for Retail Data Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Electronic retail stores face challenges in accurately displaying item package quantities due to unclear or inaccurate information in database fields, which hampers customer search and purchase decisions.
Innovation Solution
A method is developed to extract item package quantities from unstructured text in attribute fields using a prediction model that classifies the correctness of quantities from multiple fields, and refines these quantities within a classification node by removing outliers, constructing an inverted index for filtering, and displaying relevant quantities to customers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If item information is retrieved from database fields, then item details are displayed on electronic retail store page, but the information may be unclear or inaccurate
Solution Approach 1:
The patent introduces an intermediary prediction model that acts as a mediator between the unstructured text in database fields and the final displayed quantity information. This model processes multiple attribute fields (item_name, item_bullet_point, item_pkg_qty, number_of_items) and uses machine learning to predict the most accurate quantity, thereby resolving the contradiction between information availability and information accuracy.
Solution Approach 2:
The system changes the parameter of quantity representation from direct database field values to predicted values based on multiple processed attributes. By transforming the quantity information through a prediction model that considers multiple parameters (item name, bullet points, package quantity field, number of items field), the system improves accuracy while managing the complexity of extracting precise information from unstructured text.
2Measurement precision
If multiple attribute fields are used to extract quantity information, then accuracy improves, but processing complexity increases
Solution Approach 1:
The patent segments the quantity extraction task into multiple independent attribute field processing steps. Each attribute field (item_name, item_bullet_point, item_pkg_qty, number_of_items) is processed separately to extract candidate quantity values, then these segmented results are fed into the prediction model. This segmentation reduces the overall complexity by breaking down the complex extraction task into manageable parts.
Solution Approach 2:
The prediction model performs self-service by automatically selecting the most reliable quantity information from multiple attribute fields without requiring manual intervention. The model uses training data to learn which fields are most reliable for different product types and automatically makes decisions about which quantity values to trust, thereby managing system complexity through automation.
3Loss of information
If prediction model classifies quantity correctness from multiple fields, then information accuracy improves, but processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-processing and extracting candidate quantity values from all relevant attribute fields before the prediction model makes its classification. The system prepares the data in advance by identifying potential quantity values in item_name, item_bullet_point, item_pkg_qty, and number_of_items fields, so that when the prediction model needs to classify correctness, the work is already partially done, reducing actual processing time.
Solution Approach 2:
The system changes the processing approach from sequential field-by-field validation to parallel extraction followed by centralized prediction. Multiple attribute fields are processed in parallel to extract candidate quantities, then a single prediction model classification step determines the correct value. This parameter change in processing strategy reduces overall time compared to sequential verification of each field.
Data Source
AI summary
A technology is described for determining an item package quantity. An example method may include extracting a first item package quantity from a first attribute field for a product and extracting a second item package quantity from a second attribute field for the product. The first item package quantity and the second item package quantity may then be classified using a prediction model that determines a likely correctness of the first item package quantity and the second item package quantity. An item package quantity may then be selected from the first item package quantity and second item package quantity having a classification that indicates that the item package quantity is likely correct.