Adjustment of document relationship graphs

A tangible, non-transitory, machine-readable medium storing instructions that when executed by one or more processors effectuate operations comprising:

• gauging, with one or more processors, document pair-wise semantic similarity of documents;
• sending, with one or more processors, instructions to a client computing device web browser to display a visual representation that indicates the pair-wise semantic similarity of the documents;
• receiving, with one or more processors, from the client computing device web browser, an input n-gram in association with a command to increase or decrease semantic similarity between documents that include the n-gram;
• obtaining, with one or more processors, access to a corpus including natural language text documents;
• determining, with one or more processors, an expansion n-gram by determining that the expansion n-gram is related to the input n-gram based on statistical analysis of co-occurrence of candidate n-grams with the input n-gram in the natural language text of the corpus, wherein:
• determining the expansion n-gram comprises determining expansion n-grams based on respective conditional probabilities indicative of the input n-gram co-occurring with the respective candidate n-grams in bodies of text in the corpus,
• the expansion n-gram is determined before or after gauging document pair-wise semantic similarity, and
• the document pair-wise semantic similarity relates to documents in the corpus or a different corpus;
• storing, with one or more processors, the expansion n-gram in memory;
• adjusting, with one or more processors, the pair-wise semantic similarity of pairs of documents including the input n-gram or the expansion n-gram, wherein at least some adjustments are responsive to documents including the expansion n-gram;
• and
• sending, with one or more processors, instructions to the client computing device web browser to display an adjusted visual representation that indicates resulting adjusted pair-wise semantic similarity of the documents.