Hypothesis Validation via Causal-Consequence Sentence Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for extracting relevant sentences from documents to validate hypotheses about economic trends, particularly those involving causes and consequences, face challenges in balancing recall and precision, with simple word searches leading to low recall and OR searches resulting in low precision, and fail to effectively extract related but not explicitly mentioned information.
Innovation Solution
A computer-implemented method that extracts sentences from a document database by identifying documents related to the causal part of a hypothesis, obtaining high co-occurrence words, selecting sentences positionally close to these words, and relating them to the consequence part, thereby balancing recall and precision by adjusting weights for similarity and polarity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AND search is used to search for sentences including cause and consequence words, then precision is improved, but recall is lowered
Solution Approach 1:
The patent segments the hypothesis into causal part and consequence part, then separately processes each part through different extraction paths (positional extraction for cause, semantic extraction for consequence), finally merging results to achieve both precision and recall
Solution Approach 2:
The patent introduces an intermediary word list W containing high co-occurrence words that act as mediators between the causal part and relevant sentences. These intermediary words expand the search beyond exact keyword matches, improving recall while maintaining precision through the structured merging process
2Loss of information
If OR search is used to search for sentences including cause or consequence, then recall is improved, but precision is lowered
Solution Approach 1:
The patent divides the search process into two segmented paths: one for causal part (set S1) and one for consequence part (set S2), then merges them with intersection operation. This segmentation prevents irrelevant results from OR search while maintaining comprehensive coverage
Solution Approach 2:
The patent dynamically adjusts the extraction strategy based on the hypothesis structure - using positional proximity for causal part extraction and semantic similarity for consequence part extraction, optimizing both precision and recall adaptively
Data Source
AI summary
A method is provided for validating a hypothesis sentence. The method extracts, from a document database D using a hypothesis sentence that includes a causal part and a consequence part, a set D1 of documents related to the causal part. The method extracts, from set D1, a set S of sentences that include expressions of opinion. The method obtains a word list W of words that have a high co-occurrence in the set D1. The method selects, from the set S, a set S1 of sentences that are positionally close to any of the words in the word list W. The method selects, from the set S, a set S2 of sentences that are related to the words in the consequence part of the hypothesis. The method extracts and displays sentences included in both the set S1 and the set S2 as opinion sentences in relation to the hypothesis sentence.