Classification code parser
A classification code parser and method can include: reading a classification code having a description; reading a required keyword, and a total number of keywords associated with the classification code; reading text of a note; tokenizing the text of the note to create a note token stream, the note token stream having a note token and a position of the note token within the note token stream; creating a keyword map including a total number of matched keywords; determining a match ratio from the total number of the matched keywords and the total number of the keywords; determining a proximity factor based on a shortest span of tokens within the note token stream containing all the matched keywords; and determining a strength of a match between the classification code and the note based on the match ratio being multiplied by the proximity factor.
1 . A method comprising:
processing a text to generate a stream of tokens, the processing comprising:
removing negative content identified from the text, the negative content comprising a portion of the text starting with a negative term,
tokenizing the remaining text after removing negative content to generate a candidate sequence of tokens such that each token is associated with a respective position, and
replacing each of one or more tokens from the candidate sequence of tokens with a corresponding synonym token from a set of synonym tokens to generate the stream of tokens;
mapping keywords from a set of keywords to the stream of tokens to generate a set of matched keywords such that each of the matched keywords represents the same content of the corresponding token and is assigned the respective position of the corresponding token;
determining a match ratio from the total number of the set of matched keywords and the total number of the set of keywords; and
determining a strength of a match between the set of keywords and the text based on the match ratio.
2 . The method of claim 1 , wherein the set of keywords can be generated based on a classification code, and wherein the classification code comprises an ICD-O-3 code and a behavior code.
3 . The method of claim 2 , further comprising determining a behavior factor according to the following equation:
if the behavior code is “3” or if the behavior code is “2” and the text includes “non-invasive”, then the behavior factor is 1.5;
otherwise, the behavior factor is 1.0; and wherein determining the strength comprises multiplying at least the match ratio, a proximity factor determined from a keyword span of the stream of tokens, and the behavior factor.
4 . The method of claim 1 , wherein determining the match ratio includes determining the match ratio according to the following equation:
match ratio=k 2 /m, where k is the total number of the set of matched keywords and where m is the total number of the set of keywords.
5 . The method of claim 1 , further comprising: determining a proximity factor using a keyword span associated with the stream of tokens, wherein determining the proximity factor comprises:
setting the proximity factor to be 1/(1+e/50), where e is a largest value between (s−m) and 0, where s is the shortest span of tokens within the stream of tokens including the set of matched keywords, and where m is a total number of the set of keywords.
6 . The method of claim 1 , wherein determining the strength includes multiplying the match ratio with a proximity factor determined from a keyword span of the stream of tokens.
7 . A non-transitory computer readable medium storing instructions that, when executed, cause one or more processors to perform operations, the operations comprising:
processing a text to generate a stream of tokens, the processing comprising:
removing negative content identified from the text, the negative content comprising a portion of the text starting with a negative term,
tokenizing the remaining text after removing negative content to generate a candidate sequence of tokens such that each token is associated with a respective position, and
replacing each of one or more tokens from the candidate sequence of tokens with a corresponding synonym token from a set of synonym tokens to generate the stream of tokens;
mapping keywords from a set of keywords to the stream of tokens to generate a set of matched keywords such that each of the matched keywords represents the same content of the corresponding token and is assigned the respective position of the corresponding token;
determining a match ratio from the total number of the set of matched keywords and the total number of the set of keywords; and
determining a strength of a match between the set of keywords and the text based on the match ratio.
8 . The non-transitory computer readable medium of claim 7 , wherein the set of keywords can be generated based on a classification code, and wherein the classification code comprises an ICD-O-3 code and a behavior code.
9 . The non-transitory computer readable medium of claim 8 , wherein the operations further comprise: determining a behavior factor according to the following equation:
if the behavior code is “3” or if the behavior code is “2” and the text includes “non-invasive”, then the behavior factor is 1.5;
otherwise, the behavior factor is 1.0; and wherein determining the strength comprises multiplying at least the match ratio, a proximity factor determined from a keyword span of the stream of tokens, and the behavior factor.
10 . The non-transitory computer readable medium of claim 7 , wherein the operations further comprise: determining the match ratio includes determining the match ratio according to the following equation:
match ratio=k 2 /m, where k is the total number of the set of matched keywords and where m is the total number of the set of keywords.
11 . The non-transitory computer readable medium of claim 7 , wherein the operations further comprise: determining a proximity factor using a keyword span associated with the stream of tokens, wherein determining the proximity factor comprises:
setting the proximity factor to be 1/(1+e/50), where e is a largest value between (s−m) and 0, where s is the shortest span of tokens within the stream of tokens including the set of matched keywords, and where m is a total number of the set of keywords.
12 . The non-transitory computer readable medium of claim 7 , wherein determining the strength includes multiplying the match ratio with a proximity factor determined from a keyword span of the stream of tokens.
13 . A system comprising one or more computers and one or more storage devices storing instructions that, when executed by the one or more computers, cause the one or more computers to perform operations, the operations comprise:
processing a text to generate a stream of tokens, the processing comprising:
removing negative content identified from the text, the negative content comprising a portion of the text starting with a negative term,
tokenizing the remaining text after removing negative content to generate a candidate sequence of tokens such that each token is associated with a respective position, and
replacing each of one or more tokens from the candidate sequence of tokens with a corresponding synonym token from a set of synonym tokens to generate the stream of tokens;
mapping keywords from a set of keywords to the stream of tokens to generate a set of matched keywords such that each of the matched keywords represents the same content of the corresponding token and is assigned the respective position of the corresponding token;
determining a match ratio from the total number of the set of matched keywords and the total number of the set of keywords; and
determining a strength of a match between the set of keywords and the text based on the match ratio.
14 . The system of claim 13 , wherein the set of keywords can be generated based on a classification code, and wherein the classification code comprises an ICD-O-3 code and a behavior code.
15 . The system of claim 14 , wherein the operations further comprise: determining a behavior factor according to the following equation:
if the behavior code is “3” or if the behavior code is “2” and the text includes “non-invasive”, then the behavior factor is 1.5;
otherwise, the behavior factor is 1.0; and wherein determining the strength comprises multiplying at least the match ratio, a proximity factor determined from a keyword span of the stream of tokens, and the behavior factor.
16 . The system of claim 13 , wherein the operations further comprise: determining the match ratio includes determining the match ratio according to the following equation:
match ratio=k 2 /m, where k is the total number of the set of matched keywords and where m is the total number of the set of keywords.
17 . The system of claim 13 , wherein the operations further comprise: determining a proximity factor using a keyword span associated with the stream of tokens, wherein determining the proximity factor comprises:
setting the proximity factor to be 1/(1+e/50), where e is a largest value between (s−m) and 0, where s is the shortest span of tokens within the stream of tokens including the set of matched keywords, and where m is a total number of the set of keywords.
18 . The system of claim 13 , wherein determining the strength includes multiplying the match ratio with a proximity factor determined from a keyword span of the stream of tokens.