Systems and methods for managing organizational structures

A system for managing organizational structures, comprising:

• a memory to store executable instructions;
• and,
• a processor adapted to access the memory, the processor further adapted to execute the executable instructions stored in the memory to:
• receive organizational datasets representing an organizational structure for an enterprise, the received organizational datasets including human resource records that are stored in the memory, each human resource record including a role dataset representing an enterprise role for the enterprise;
• generate role tokens based on at least a designated portion of the role dataset, the generated role tokens including adjusted unique role datasets having text or values contained in the designated portion of the role dataset that are truncated and deduped, wherein the generated role tokens are stored in the memory;
• assign prioritization scores to the generated role tokens using a hierarchical clustering method and a statistical modelling method;
• receive, via an user interface, a training sample size, a training cycle count, a threshold prediction certainty, and a prediction iterations count;
• extract the generated role tokens having threshold prioritization scores;
• assign training labels to a training portion of the extracted tokens based on the assigned prioritization scores, the training labels representing a categorical level associated with the extracted tokens in the training portion, the training portion defined by the training sample size, the training labels repeatedly assigned to the training portion one or more training cycles, and the training cycles defined by the training cycle count;
• generating a sparse matrix based on the training portion of the extracted tokens, wherein the assigned training labels for the training portion are assigned to token elements in the sparse matrix;
• assign predictive labels to a remaining portion of the extracted tokens via a logistic regression classifier based on model parameters and the sparse matrix, each of the predictive labels representing the categorical level associated with the extracted tokens in the remaining portion, each of the predictive labels including an assigned prediction certainty generated by the logistic regression classifier, the predictive labels are repeatedly assigned by the logistic regression classifier to each one of the extracted tokens in the remaining portion for one or more prediction iterations, the prediction iterations are based on the prediction iterations count, and the prediction iterations for at least one of the extracted tokens terminating if the assigned prediction certainty for the at least one of the extracted tokens is above the threshold prediction certainty;
• adjust the model parameter of the logistic regression classifier after each of the prediction iterations;
• generate a model organizational dataset based on the assigned training labels and the assigned predictive labels, the model organizational dataset including the extracted tokens, each of the extracted tokens associated with one or more categorical levels;
• and,
• adjust the extracted tokens in the model organizational dataset that are associated with more than one of the one or more categorical levels and at least one flagged categorical level, the flagged categorical level of one of the extracted tokens differing from the first categorical level associated with the one of the extracted tokens, the adjusted token having an adjusted categorical level.

A non-transitory computer-readable medium including instructions configured to be executed by a processor, wherein the executed instructions are adapted to cause the processor to:

• receive organizational datasets representing an organizational structure for an
• enterprise, the received organizational datasets including human resource records that are stored in the memory, each human resource record including a role dataset representing an enterprise role for the enterprise;
• generate role tokens based on at least a designated portion of the role dataset, the generated role tokens including adjusted unique role datasets having text or values contained in the designated portion of the role dataset that are truncated and deduped, wherein the generated role tokens are stored in the memory;
• assign prioritization scores to the generated role tokens using a hierarchical clustering method and a statistical modelling method;
• receive, via an user interface, a training sample size, a training cycle count, a threshold prediction certainty, and a prediction iterations count;
• extract the generated role tokens having threshold prioritization scores;
• assign training labels to a training portion of the extracted tokens based on the assigned prioritization scores, the training labels representing a categorical level associated with the extracted tokens in the training portion, the training portion defined by the training sample size, the training labels repeatedly assigned to the training portion one or more training cycles, and the training cycles defined by the training cycle count;
• generating a sparse matrix based on the training portion of the extracted tokens, wherein the assigned training labels for the training portion are assigned to token elements in the sparse matrix;
• assign predictive labels to a remaining portion of the extracted tokens via a logistic regression classifier based on model parameters and the sparse matrix, each of the predictive labels representing the categorical level associated with the extracted tokens in the remaining portion, each of the predictive labels including an assigned prediction certainty generated by the logistic regression classifier, the predictive labels are repeatedly assigned by the logistic regression classifier to each one of the extracted tokens in the remaining portion for one or more prediction iterations, the prediction iterations are based on the prediction iterations count, and the prediction iterations for at least one of the extracted tokens terminating if the assigned prediction certainty for the at least one of the extracted tokens is above the threshold prediction certainty;
• adjust the model parameter of the logistic regression classifier after each of the prediction iterations;
• generate a model organizational dataset based on the assigned training labels and the assigned predictive labels, the model organizational dataset including the extracted tokens, each of the extracted tokens associated with one or more categorical levels;
• and,
• adjust the extracted tokens in the model organizational dataset that are associated with more than one of the one or more categorical levels and at least one flagged categorical level, the flagged categorical level of one of the extracted tokens differing from the first categorical level associated with the one of the extracted tokens, the adjusted token having an adjusted categorical level.