System for evaluating journal articles

A computer program product for determining likelihood that a medical journal article is classified as high value, the computer program product comprising a computer readable storage medium having instructions executable by one or more processors to cause a computer system to:

• procure, in a medical articles database on the computer system, a training set including medical articles recommended by subject matter experts associated with medical journals that have published the medical articles, wherein each recommended medical article is associated with two randomly selected articles associated with a source of the corresponding recommended medical article;
• identify, by a natural language processor, n-grams corresponding to entities, terms, and natural language data in titles, abstracts, and text of the medical articles, wherein the n-grams include disease types, patient characteristics, and drug types and corresponding frequencies in the titles, abstracts, and text of the medical articles associated with the list of possible treatments for the patient;
• identify, by a feature extraction controller of the computer, first features in the medical articles via a remote annotation service and a remote article information service, wherein the identification of the first features is based on the entities, terms, and natural language data;
• reduce, by the feature extraction controller, the first features to form a set of relevant features based on a classification label associated with each medical article utilizing a chi-squared algorithm, wherein the classification label indicates a value relative to the other medical articles;
• adapt, by a classifier of the computer system, a classification function based on the relevant features of the medical articles in the training set and on the classification label of each medical article utilizing supervised learning;
• weigh, by a weight unit, each treatment in the list of possible treatments based on the adapted classification function, procured medical articles, patient allergies, and previously tried treatments;
• and
• recommend, by an insight resolution unit, a treatment based on the weighted list of possible treatments;
• provide, by a business and domain insight generator, the recommended treatment to one or more entities, wherein the provided recommended treatment includes include annotations information, wherein annotations information constitutes aggregations of the set of relevant features.

An apparatus comprising:

• one or more processors;
• a computer program product comprising a computer readable storage medium having instructions to determine likelihood that a medical journal article is classified as high value, the instructions executable by the one or more processors to cause the one or more processors to:
• procure, in a medical articles database, a training set including medical articles recommended by subject matter experts associated with medical journals that have published the medical articles, wherein each recommended medical article is associated with two randomly selected articles associated with a source of the corresponding recommended medical article;
• identify, by a natural language processor, n-grams corresponding to entities, terms, and natural language data in titles, abstracts, and text of the medical articles, wherein the n-grams include disease types, patient characteristics, and drug types and corresponding frequencies in the titles, abstracts, and text of the medical articles associated with the list of possible treatments for the patient;
• identify, by a feature extraction controller, first features in the medical articles via a remote annotation service and a remote article information service wherein the identification of the first features is based on the entities, terms, and natural language data;
• reduce, by the feature extraction controller, the first features to form a set of relevant features based on a classification label associated with each medical article utilizing a chi-squared algorithm, wherein the classification label indicates a value relative to the other medical articles;
• adapt, by a classifier, a classification function based on the relevant features of the medical articles in the training set, and on the classification label of each medical article utilizing supervised learning;
• weigh, by a weight unit, each treatment in the list of possible treatments based on the adapted classification function, procured medical articles, patient allergies, and previously tried treatments;
• and
• recommend, by an insight resolution unit, a treatment based on the weighted list of possible treatments;
• provide, by a business and domain insight generator, the recommended treatment to one or more entities, wherein the provided recommended treatment includes include annotations information, wherein annotations information constitutes aggregations of the set of relevant features.