Attribute Data Interval Classification with Missing Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for classifying website members based on attribute data intervals face inaccuracies when user attributes are missing, leading to low accuracy in member classification due to the inability to accurately determine attribute data intervals.
Innovation Solution
The method involves setting a preset missing value for missing attribute data, merging it into initial attribute data intervals, and calculating the Maximum Posteriori Probability (MAP) Bayes evaluation value for each classification, determining the attribute data intervals with the least MAP Bayes evaluation value as the final intervals, thereby improving classification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods determine attribute data intervals based on complete user attributes only, then the classification process is simple, but the classification accuracy decreases when user attributes are missing
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing attribute data intervals during an offline training phase. The system processes historical user data to determine intervals for each user attribute (e.g., age ranges, location groups) and stores these intervals in advance. When classifying new users, the system only needs to query these pre-computed intervals rather than performing complex calculations in real-time, thus improving classification accuracy while maintaining simple online processing.
Solution Approach 2:
The patent segments the continuous attribute data into discrete intervals. For example, age is divided into intervals like [0, 18), [18, 30), [30, 50), [50, 100], and locations are segmented into regional groups. This segmentation transforms complex continuous data into manageable discrete categories, enabling accurate classification even when some attributes are missing, as the system can work with the segmented intervals that have been pre-defined during training.
2Adaptability or versatility
If the system requires all user attributes to be present for classification, then the classification logic is straightforward, but the system cannot handle missing attribute data
Solution Approach 1:
The patent implements partial action by enabling classification to proceed with only a subset of user attributes. During the offline training phase, the system learns to classify users based on available attributes without requiring all attributes to be present. The online classification process can then reliably classify users even when some attributes are missing, as the pre-computed intervals are designed to work with partial data. The system performs exactly the classification needed without requiring excessive (complete) attribute information.
Data Source
AI summary
The present techniques extract attribute data of one or more classified members for one or more user attributes. With respect to a particular user attribute of the one or more user attributes, the present techniques determine initial attribute data intervals corresponding to the particular user attribute based on attribute data and classes of the classified members from the extracted attribute data. With respect to a classified member whose attribute data is missing for the particular user attribute, the present techniques set the attribute data as a preset missing value. The present techniques then merge the preset missing value into each of the initial user attribute data intervals and calculate a Maximum Posteriori Probability (MAP) Bayes estimate value respectively, and determine initial user attribute data intervals with a smallest MAP Bayes estimated value as final attribute data intervals corresponding to the particular user attribute.