Creating a survey sample group according to a desired participant distribution in real time

A method of creating a balanced survey group in real time with multiple goals, the method comprising:

• creating a vector representation of an entry, the vector representation being a multidimensional vector representation identifying an exclusion or inclusion of a candidate from each of a plurality of categories;
• calculating a first similarity value indicating a similarity between the vector representation and a first histogram vector, the first histogram vector representing a dynamically balanced emptiness associated with each of a plurality of categories determined based on a first goal and during a creation of an initial survey group, the similarity value being a result of a dot product between the vector representation and the first histogram vector;
• calculating a second similarity value indicating a similarity between the vector representation and a second histogram vector;
• calculating a third similarity value indicating a similarity between the vector representation and a third histogram vector;
• integrating the first, second, and third similarity values to calculate a final similarity value;
• and
• integrating the entry vector representation into the first histogram vector, the second histogram vector, and the third histogram vector based on a comparison of the first similarity value, the second similarity value, and the third similarity value with at least one similarity threshold, wherein the first histogram relates to the first goal, the second histogram relates to a second goal, and the third histogram relates to a third goal, and wherein the integrating enables, at least in part, dynamic balancing of the first histogram vector, the second histogram vector, and the third histogram vector while at least one initial survey group is being created.

A computing apparatus for creating a balanced survey group in real time, the apparatus comprising:

• means for creating a vector representation of an entry, an entry being a potential respondent in a survey, and the vector representation being a multidimensional vector representation identifying an exclusion or inclusion of the potential respondent from each of a plurality of categories;
• means for calculating a similarity value indicating a similarity between the entry vector representation and a first histogram vector, the first histogram vector representing a dynamically balanced emptiness associated with each of a plurality of categories determined based on a first goal and during a creation of an initial survey group, the similarity value being a result of a dot product between the vector representation and the first histogram vector;
• means for determining whether the similarity value is greater than a similarity threshold by comparing the result of the dot product with the similarity threshold, the similarity threshold being determined based on a linear superposition of a most-preferred entry vector and a least-preferred entry vector representing a type of respondent most needed and least needed;
• and
• integrating the entry vector representation into the first histogram vector in real time, wherein the first histogram relates to the first goal, and wherein the integrating enables, at least in part, dynamic balancing of the first histogram vector while the initial survey group is being created.