The duplicate checking results in PDF format and word format are generally the same, but there are different results, which are caused by the following reasons:

If the table of contents in the paper in word format is correctly typeset, the HowNet detection system will divide the chapters according to the table of contents, and then check each chapter. The table of contents is displayed in gray and does not participate in duplicate checking. Papers in pdf format need to be converted into text, and the directory format is incorrect, so the directory should be tested and compared. Therefore, the detection rate of PDF is high;

In the case of correct typesetting, references do not compare data and do not participate in data comparison. In this way, the paper references in pdf format will be converted into text format, and the format is incorrect, so they are basically marked in red, so the PDF detection is on the high side.

Endnotes and footnotes are generally not detected by HowNet in Word, but in PDF, HowNet is detected as text for the same reason as 3. Therefore, the PDF detection is on the high side.

PDF format has strong confidentiality. If the contents in the PDF document are not correctly identified during repeated checking and parsing, garbled codes will appear, resulting in a result of 0.