Evaluating text understanding based on context
Ning Fang, Xiangfeng Luo, Weimin Xu · 2008
Based on minimization of Boolean complexity in human concept learning, a computational method of text understanding is proposed. Based on defining the complexity of text understanding, the difficulty of text understanding is used to mimic human’s reading experience and measure the understanding degree of a text by reader, and the sum of the complexity of text understanding derived from the addition of one-by-one sequential sentences is used to mimic human’s reading process and measure the energy consumed by reader. Experiments verify that more contexts are added, more easily the text is understood by a machine, which is consistent with linguistic viewpoint that context can help to understand a text; furthermore, experiments verify that author-given sentence sequence is one of the sequences most easily understood by a machine, in other words, the principle of simplicity actually exists in text writing process, which is consistent with cognitive science viewpoint that human concept learning abides by the principle of simplicity. Therefore, our method is validated from the linguistic and cognitive perspectives, and it could provide a theoretical foundation for machine-based text understanding.