System and method for creating and searching medical ontologies
View Patent ↗A method for creating and searching medical ontologies includes providing a semi-structured information source comprising a plurality of articles linked to each other, each article having one or more sections and each article is associated with a concept, creating a directed unlabeled graph representative of the information source, providing a plurality of labels, labeling a subset of edges, and assigning each unlabeled edge an equal probability of being assigned one of the labels. For each node, the probability of each outgoing edge is updated by smoothing each probability by an overall probability distribution of labels over all outgoing edges of each node, and the probability of each incoming edge is updated the same way. A label with a maximum probability is assigned to an edge if said maximum probability is greater than a predetermined threshold to create a labeled graph.
1. A method for creating and searching medical ontologies comprising the steps of:
providing a semi-structured information source comprising a plurality of articles linked to each other, each article having one or more sections, wherein each article is associated with a concept;
creating a directed unlabeled graph representative of said information source, wherein each graph node is associated with a concept and each graph edge is associated with a link between concepts;
providing a plurality of labels, and labeling a subset of said edges of said graph;
assigning, by a processor, each unlabeled edge an equal probability of being assigned one of said plurality of labels;
updating, for each node, the probability of each outgoing edge by smoothing each said probability by an overall probability distribution of labels over all outgoing edges of each node;
updating, for each node, the probability of each incoming edge by smoothing each said probability by an overall probability distribution of labels over all incoming edges of each node; and
assigning a label with a maximum probability to an edge if said maximum probability is greater than a predetermined threshold to create a labeled graph, wherein said labeled graph represents a new ontology configured to be searched.
2. The method of claim 1 , further comprising repeating said steps of updating, for each node, the probability of each outgoing edge and updating, for each node, the probability of each incoming edge.
3. The method of claim 1 , further comprising discarding an edge if said maximum probability for a label is less than said predetermined threshold.
4. The method of claim 1 , wherein updating, for each node, the probability of each outgoing or incoming edge comprises calculating
P
ik
←
P
ik
×
∑
j
=
1
n
P
jk
∑
l
=
1
m
(
P
il
×
∑
j
=
1
n
P
jl
)
where P ik is the probability of edge i having label k, the sum in the numerator and the inner sum in the denominator are over all outgoing or incoming edges, and the outer sum in the denominator is over all labels.
5. The method of claim 1 , wherein labeling a subset of said edges of said graph comprises finding a first article having a list a concepts, finding an instance of a concept in said list of concepts in a section of a second article, and assigning a title of said section of said second article as a label of an edge from said second article to said concept.
6. The method of claim 1 , wherein labeling a subset of said edges of said graph comprises finding a concept in a list of concepts in a section of an article, and assigning a title of said section as a label of an edge from said article to said concept.
7. The method of claim 1 , further comprising the step of extracting from said information source a plurality of concepts and a plurality of links, each link representative of a relationship between two or more of the plurality of concepts.
8. The method of claim 1 , wherein each label of the plurality of labels represents a relationship between two concepts.
9. In a non-transitory computer readable program storage device having stored therein a program of instructions executable by the computer for creating and searching medical ontologies, the program storage device comprising instructions for:
providing a semi-structured information source comprising a plurality of articles linked to each other, each article having one or more sections, wherein each article is associated with a concept;
creating a directed unlabeled graph representative of said information source, wherein each graph node is associated with a concept and each graph edge is associated with a link between concepts;
providing a plurality of labels, and labeling a subset of said edges of said graph;
assigning each unlabeled edge an equal probability of being assigned one of said plurality of labels;
updating, for each node, the probability of each outgoing edge by smoothing each said probability by an overall probability distribution of labels over all outgoing edges of each node;
updating, for each node, the probability of each incoming edge by smoothing each said probability by an overall probability distribution of labels over all incoming edges of each node; and
assigning a label with a maximum probability to an edge if said maximum probability is greater than a predetermined threshold to create a labeled graph, wherein said labeled graph represents a new ontology configured to be searched.
10. The non-transitory computer readable program storage device of claim 9 , the method further comprising repeating said steps of updating, for each node, the probability of each outgoing edge and updating, for each node, the probability of each incoming edge.
11. The non-transitory computer readable program storage device of claim 9 , the method further comprising discarding an edge if said maximum probability for a label is less than said predetermined threshold.
12. The non-transitory computer readable program storage device of claim 9 , wherein updating, for each node, the probability of each outgoing or incoming edge comprises calculating
P
ik
←
P
ik
×
∑
j
=
1
n
P
jk
∑
l
=
1
m
(
P
il
×
∑
j
=
1
n
P
jl
)
where P ik is the probability of edge i having label k, the sum in the numerator and the inner sum in the denominator are over all outgoing or incoming edges, and the outer sum in the denominator is over all labels.
13. The non-transitory computer readable program storage device of claim 9 , wherein labeling a subset of said edges of said graph comprises finding a first article having a list a concepts, finding an instance of a concept in said list of concepts in a section of a second article, and assigning a title of said section of said second article as a label of an edge from said second article to said concept.
14. The non-transitory computer readable program storage device of claim 9 , wherein labeling a subset of said edges of said graph comprises finding a concept in a list of concepts in a section of an article, and assigning a title of said section as a label of an edge from said article to said concept.