Adaptive document curation
An adaptive document curation method. A machine learning model is trained on a first library of documents to assign staleness scores to documents, each indicating a confidence that the document to which it is assigned should be archived. If a document has a staleness score at least equal to a threshold, the document is identified as a candidate to be archived or deleted.
1. A method comprising:
training a machine learning model on a first library of documents to assign staleness scores to the documents, wherein a staleness score indicates a confidence that the document to which it is assigned should be archived;
applying the machine learning model to a first document to generate a first staleness score for the first document;
based upon a determination that the first staleness score is at least equal to a first threshold, identifying the first document as a candidate to be archived; and
archiving the first document.
2. The method of claim 1 , further comprising:
adjusting the machine learning model based upon receiving a confirmation that the first document should be archived.
3. The method of claim 1 , further comprising:
receiving an identification of a second document not included in the first library of documents;
applying the machine learning model to the second document to generate a second staleness score for the second document;
based upon a determination that the second staleness score is at least equal to the first threshold, identifying the second document as a candidate to be archived;
presenting, to a user, the second document and an indication that the second document should be archived; and
receiving an indication from the user that the second document should not be archived or that the second document should be archived.
4. The method of claim 1 , further comprising:
receiving a confirmation that the first document should be archived.
5. The method of claim 4 , wherein the confirmation that the first document should be archived is received in response to providing an indication of the staleness score.
6. The method of claim 1 , wherein, to generate the staleness score, the machine learning model uses one or more attributes selected from the group consisting of: a creation date of the first document, a last edit date of the first document, a count of unique edits of the first document during a time period, a count of unique editors of the first document during a time period, a count of unique viewers of the first document during a time period, a count of unique non-editor viewers of the first document during a time period, a count of other documents linked from the first document, an average staleness score of documents linked to the first document, a count of documents that link to the first document, and an aggregate staleness score of a folder containing the first document.
7. A non-transitory computer readable medium having instructions that when performed on at least one processor cause the at least one processor to perform steps comprising:
training a machine learning model on a first library of documents to assign staleness scores to the documents, wherein a staleness score indicates a confidence that the document to which it is assigned should be archived;
applying the machine learning model to a first document to generate a first staleness score for the first document;
based upon a determination that the first staleness score is at least equal to a first threshold, identifying the first document as a candidate to be archived; and
archiving the first document.
8. The non-transitory computer readable medium of claim 7 , the steps further comprising:
adjusting the machine learning model based upon receiving a confirmation that the first document should be archived.
9. The non-transitory computer readable medium of claim 7 , the steps further comprising:
receiving an identification of a second document not included in the first library of documents;
applying the machine learning model to the second document to generate a second staleness score for the second document;
based upon a determination that the second staleness score is at least equal to the first threshold, identifying the second document as a candidate to be archived;
presenting, to a user, the second document and an indication that the second document should be archived; and
receiving an indication from the user that the second document should not be archived or that the second document should be archived.
10. The non-transitory computer readable medium of claim 7 , the steps further comprising:
receiving a confirmation that the first document should be archived.
11. The non-transitory computer readable medium of claim 10 , wherein the confirmation that the first document should be archived is received in response to providing an indication of the staleness score.
12. The non-transitory computer readable medium of claim 7 , wherein, to generate the staleness score, the machine learning model uses one or more attributes selected from the group consisting of: a creation date of the first document, a last edit date of the first document, a count of unique edits of the first document during a time period, a count of unique editors of the first document during a time period, a count of unique viewers of the first document during a time period, a count of unique non-editor viewers of the first document during a time period, a count of other documents linked from the first document, an average staleness score of documents linked to the first document, a count of documents that link to the first document, and an aggregate staleness score of a folder containing the first document.
13. A system comprising:
a processor;
a non-transitory machine-readable storage medium storing a first library of documents; and
a set of computer-readable instructions provided by the non-transitory machine-readable storage medium, the instructions, if executed by the processor, are configurable to cause the processor to perform operations comprising:
train a machine learning model on the first library of documents to assign staleness scores to the documents, wherein a staleness score indicates a confidence that the document to which it is assigned should be archived;
apply the machine learning model to a first document to generate a first staleness score for the first document;
based upon a determination that the first staleness score is at least equal to a first threshold, identify the first document as a candidate to be archived; and
archive the first document.
14. The system of claim 13 , the processor further configured to:
adjust the machine learning model based upon receiving a confirmation that the first document should be archived.
15. The system of claim 13 , the processor further configured to:
receive an identification of a second document not included in the first library of documents;
apply the machine learning model to the second document to generate a second staleness score for the second document;
based upon a determination that the second staleness score is at least equal to the first threshold, identify the second document as a candidate to be archived;
present, to a user, the second document and an indication that the second document should be archived; and
receive an indication from the user that the second document should not be archived or that the second document should be archived.
16. The system of claim 13 , the processor further configured to:
receiving a confirmation that the first document should be archived.
17. The system of claim 16 , wherein the confirmation that the first document should be archived is received in response to providing an indication of the staleness score.
18. The system of claim 13 , wherein, to generate the staleness score, the machine learning model uses one or more attributes selected from the group consisting of: a creation date of the first document, a last edit date of the first document, a count of unique edits of the first document during a time period, a count of unique editors of the first document during a time period, a count of unique viewers of the first document during a time period, a count of unique non-editor viewers of the first document during a time period, a count of other documents linked from the first document, an average staleness score of documents linked to the first document, a count of documents that link to the first document, and an aggregate staleness score of a folder containing the first document.