| 摘要 |
The present invention is a computer implemented system for analyzing text, for computing aspects of sentiment via chord transitions, such as prissiness, bombast and sarcasm via underlying multiple dimensions of sentiment, computing compassionate rhetoric via successive areas of sentiment in text, for automatically computing normalization of polysemous dictionary entries to provide lookup of meanings for text analysis of words and phrases, and for automatically computing the degree of theme interweave in text via computation of conceptual crossings of paragraphs. |
| 主权项 |
1. A computer implemented search engine which automatically generates a plurality of search results from an input text and a query text, the search engine comprising a processor, a search index for retrieving matches in meaning between an input text and query text, and a user interface for receiving the query text and displaying search engine results, the method comprising:
receiving the input text and query text; performing, via the processor, and a token-by-token analysis of the query text and input text, a computation of a map of sentiment valences to successive areas of text therein; computing, via the processor, a summation of a negative area from negative valences and a summation of a resolution area from positive valences, based on the map of sentiment valences; computing an unacceptable area, via the processor, based on a difference between the negative area and the positive area; computing an unacceptable running area, via the processor, based on an excess of the unacceptable area which is beyond a max acceptable running imbalance area; computing, via the processor, an acceptable negative area for 60% resolution, based on the positive area, wherein the acceptable negative area is within 60% of the positive area; computing a total compassion, via the processor, based on the acceptable negative area, the negative area, the positive area, wherein total compassion is the subtraction of the unacceptable running area from the sum of the positive area and the acceptable negative area; computing, an array of paragraph clusters ordered by compassion in similar paragraphs, via the processor, based on the total compassion in token-by-token analysis of the query text and input text; computing a query cluster compassion trigrams index based on the array of paragraph clusters ordered by compassion in similar paragraphs; retrieving a set of segment results sorted by relevance based on search engine intersection of input text and query text within the query cluster compassion trigrams index; outputting, via the user interface, the set of segment results sorted by relevance; wherein the computer analysis includes a classification, a categorization, or sorting of the segment results ordered by compassion. |
