The principle of paper duplicate checking is:
1. The detection system only accepts the text content of the paper, and the contents such as pictures, tables and formula editing will be ignored.
2. The system first divides the article into paragraphs according to line breaks, then divides the search paragraphs into sentences according to punctuation marks, and then checks each sentence again. The similarity between two sentences mainly depends on which words are contained in the sentence and their positions in the sentence;
3. Overall similarity = number of similar words/number of detected words. Non-text parts (such as table of contents, titles, formulas, charts, references, etc.). ) the system automatically recognizes that it does not participate in the detection, and the number of words detected is generally slightly less than that of the paper.
General colleges and universities require that the duplicate checking rate of graduation thesis is less than 30%, and some require 20%. If a paper is published, the publishing house (such as Hans Publishing House) usually requires that the duplicate checking rate be less than 15%. Therefore, we must find ways to reduce the weight of papers with high duplicate checking rate!
As for the means of losing weight, I recommend the simplest one.
China culture is extensive and profound. We can express the same meaning in many ways, or we can expand or reduce a sentence to something with the same meaning but different words.
However, with the development of machine learning and artificial mental retardation, HowNet's weight check system is constantly improving, and some artificial weight loss content can be seen, which will test friends' language organization ability-how to turn words into something that people can understand but machines can't understand.
In addition, some majors need a lot of charts, and there are a lot of technical terms on the charts. You can take screenshots of some charts, but you can't recognize the contents of the pictures at present (this is a folk prescription and is not recommended).
It is worth noting that after passing the test, please don't make any big additions to avoid adding a lot of untested words.