ML Classification from Crawled Data for Client Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Companies face challenges in efficiently evaluating and recommending a subset of potential clients or partners from a large pool, due to limited information and resource-intensive brute force approaches.
Innovation Solution
A method and system that utilize crawled data from websites to generate classifications by applying a machine learning algorithm. This algorithm processes text bodies parsed from website information, weights critical terms, and determines categorization labels, ultimately recommending suitable clients or partners.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a brute force approach is used to evaluate each prospective client/partner, then evaluation completeness is improved, but resource consumption and time increase significantly
Solution Approach 1:
The patent segments the evaluation process into multiple stages: initial filtering using machine learning models to identify promising candidates, followed by more detailed analysis only for those that pass the filter. This divides the large pool of prospects into segments that require different levels of evaluation resources, resolving the contradiction between thorough evaluation and resource efficiency.
Solution Approach 2:
The system performs preliminary actions by pre-processing and analyzing basic information about all prospects using automated machine learning models before human reviewers invest time in detailed evaluation. This preliminary filtering identifies high-potential candidates, ensuring that comprehensive evaluation resources are focused only on the most promising cases.
2Measurement precision
If more information is extracted and parsed for each entity, then classification accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The patent applies partial action by extracting and parsing only the most relevant information fields for initial classification, rather than analyzing every piece of available data. The machine learning model is trained to make accurate classifications based on a selective subset of features, achieving good classification accuracy without the computational cost of processing all possible information.
Solution Approach 2:
The system applies local quality by adjusting the depth and type of information extraction based on the specific entity and context. For some entities, more detailed parsing is performed, while for others, a lighter extraction suffices. This adaptive approach maintains classification accuracy while optimizing processing time based on local characteristics of each case.
3Productivity
If automated machine learning models are used to process evaluations, then processing speed is improved, but risk of bias from varying information availability increases
Solution Approach 1:
The patent implements equipotentiality by designing the machine learning model to normalize and standardize the evaluation criteria across all prospects. The model is trained to recognize that different amounts of information do not necessarily indicate different quality levels, and it adjusts its assessment to account for varying information availability, ensuring that all entities are evaluated on an equal footing.
Solution Approach 2:
The system incorporates feedback mechanisms where the machine learning model's predictions are continuously refined based on outcomes and human reviewer corrections. This feedback loop helps the model learn to compensate for biases related to information availability, improving fairness over time while maintaining high processing speed.
Data Source
AI summary
Disclosed herein are systems and method for extrapolating from crawled data to generate classifications. In one aspect, a method may receive an evaluation request with an input list comprising at least one entity and a respective website identifier of the at least one entity, identify, from a plurality of categorization labels, a subset of categorization labels that correspond to a requesting entity that generated the evaluation request, crawl information using the website identifier, generate at least one text body from the crawled information, apply a machine learning algorithm on the at least one text body, wherein the machine learning algorithm is configured to generate an output vector indicating categorization labels from the subset of categorization labels that the at least one entity corresponds to, based on terms in the at least one text body; and transmit the output vector to a computing device of the requesting entity.