Data driven analysis, modeling, and semi-supervised machine learning for qualitative and quantitative determinations
Systems and methods are provided for data driven analysis, modeling, and semi-supervised machine learning for qualitative and quantitative determinations. The systems and methods include obtaining data associated with individuals, and determining features associated with the individuals based on the data and similarities among the individuals based on the features. The systems and methods can label some individuals as exemplary, generate a graph wherein nodes of the graph represent individuals, edges of the graph represent similarity among the individuals, and nodes associated labeled individuals are weighted. The disclosed system and methods can apply a weight to unweighted nodes of the graph based on propagating the labels through the graph where the propagation is based on influence exerted by the weighted nodes on the unweighted nodes. The disclosed systems and methods can provide output associated with the individuals represented on the graph and the associated weights.
1. A non-transitory computer readable storage medium storing instructions that are executable by a first computing device that includes one or more processors to cause the first computing device to perform a method comprising:
obtaining, from one or more data sources, one or more data sets associated with a plurality of professionals;
determining features associated with the plurality of professionals;
determining similarities among the plurality of professionals based on the determined features;
generating data representing a connected graph based on the determined similarities and the determined features, wherein nodes of the graph are associated with the plurality of professionals;
determining a first set of labels for a first subset of the plurality of professionals;
annotating a first subset of nodes of the graph with the first set of labels;
annotating a second subset of nodes of the graph with a second set of labels by propagating the first set of labels to the second subset of nodes of the graph;
ranking the plurality of professionals based on the first set of labels and the second set of labels; and
providing output associated with the ranking.
2. The non-transitory computer readable storage medium of claim 1 , wherein generating data representing a connected graph utilizes a random graph model to select a subset of nodes to connect using edges on the graph.
3. The non-transitory computer readable storage medium of claim 1 , wherein edges of the graph are weighted and the weight of the edges are based on the determined similarities.
4. The non-transitory computer readable storage medium of claim 1 , wherein the set of labels includes at least one of positive labels and negative labels.
5. The non-transitory computer readable storage medium of claim 1 , wherein propagating the first set of labels to the second subset of nodes of the graph further comprises:
identifying a second subset of the plurality of professionals that are similar to at least one professional of the first subset of the plurality of professionals based on determined similarities;
determining the second subset of labels for the second subset of the plurality of professionals based on determined features among the first subset of the plurality of professionals and the second subset of the plurality of professionals;
annotating the second subset of nodes of the graph using the second subset of labels.
6. The non-transitory computer readable storage medium of claim 1 , wherein the graph includes weighted nodes and unweighted nodes, wherein the weighted nodes are based on the first set of labels.
7. The non-transitory computer readable storage medium of claim 6 , wherein propagating the first set of labels to the second subset of nodes of the graph further comprises:
weighting the unweighted nodes of the graph wherein the weights are based on an influence exerted by the weighted nodes on the unweighted nodes.
8. The non-transitory computer readable storage medium of claim 7 , wherein the influence of the weighted nodes on the unweighted nodes is determined using Poisson's equation, Laplace's equation, a Laplacian exponential diffusion kernel, a regularized Laplacian kernel, or a Von Neumann diffusion kernel.
9. The non-transitory computer readable storage medium of claim 1 , wherein providing the output further comprises providing the data representing the graph and the output associated with the ranking for processing by a client device.
10. The non-transitory computer readable storage medium of claim 1 , wherein providing output further comprises providing the output for display on a graphical user interface.
11. A data-driven analysis system comprising:
one or more memory devices storing processor executable instructions; and
one or more processors configured to execute the instructions to cause the data-driven analysis system to perform:
obtaining, from one or more data sources, one or more data sets associated with a plurality of professionals;
determining features associated with the plurality of professionals;
determining similarities among the plurality of professionals based on the determined features;
generating data representing a connected graph based on the determined similarities and the determined features, wherein nodes of the graph are associated with the plurality of professionals;
determining a first set of labels for a first subset of the plurality of professionals;
annotating a first subset of nodes of the graph with the first set of labels;
annotating a second subset of nodes of the graph with a second set of labels by propagating the first set of labels to the second subset of nodes of the graph;
ranking the first set of plurality of professionals based on the first set of labels and the second set of labels; and
providing output associated with the ranking.
12. The data-driven analysis system of claim 11 , wherein edges of the graph are weighted and the weight of the edges are based on the determined similarities.
13. The data-driven analysis system of claim 11 , wherein propagating the first set of labels to the second subset of nodes of the graph further comprises:
identifying a second subset of the plurality of professionals that are similar to at least one professional of the first subset of the plurality of professionals based on determined similarities;
determining a second subset of labels for the second subset of the plurality of professionals, wherein the second subset of labels are based on determined feature among the first subset of the plurality of professionals and the second subset of the plurality of professionals;
annotating the second subset of nodes of the graph using the second subset of labels.
14. The data-driven analysis system of claim 11 , wherein the graph includes weighted nodes and unweighted nodes, wherein the weighted nodes are based on the first set of labels.
15. The data-driven analysis system of claim 11 , wherein propagating the first set of labels to the second subset of nodes of the graph further comprises:
weighting the unweighted nodes of the graph wherein the weights are based on an influence exerted by the weighted nodes on the unweighted nodes.
16. The data-driven analysis system of claim 11 , wherein providing the output further comprises providing the data representing the graph and the output associated with the ranking for processing by a client device.
17. A method performed by one or more processors and comprising:
obtaining, from one or more data sources, one or more data sets associated with a plurality of professionals;
determining features associated with the plurality of professionals;
determining similarities among the plurality of professionals based on the determined features;
generating data representing a connected graph based on the determined similarities and the determined features, wherein nodes of the graph are associated with the plurality of professionals;
determining a first set of labels for a first subset of the plurality of professionals;
annotating a first subset of nodes of the graph with the first set of labels;
annotating a second subset of nodes of the graph with a second set of labels by propagating the first set of labels to the second subset of nodes of the graph;
ranking the first set of plurality of professionals based on the first set of labels and the second set of labels; and
providing output associated with the ranking.
18. The method of claim 17 , wherein the set of labels includes at least one of positive labels and negative labels.
19. The method of claim 18 , wherein propagating the first set of labels to the second subset of nodes of the graph further comprises;
identifying a second subset of the plurality of professionals that are similar to at least one professional of the first subset of the plurality of professionals based on determined similarities;
determining a second subset of labels for the second subset of the plurality of professionals, wherein the second subset of labels are based on determined features among the first subset of the plurality of professionals and the second subset of the plurality of professionals;
annotating the second-subset of nodes of the graph using the second subset of labels.
20. The method of claim 19 , wherein providing output further comprises providing the output for display on a graphical user interface.