Text based prediction of psychological cohorts
Computer implemented techniques for classifying mental states of individuals are described. The techniques determine sets of words that are associated with multiple groups having different mental status, and a classification model is used to classify one group against another group. Furthermore, by determining points of intersection of words between a first group and second group, words that are statistically predictive terms and that are unique to each group, to provide further predictive features for differentiating the multiple cohorts.
1. A computer implemented process comprises:
producing by one or more computers plural sets of training data that comprise words associated with multiple groups of individuals, with the multiple groups being of individuals having different classifications of mental states that define a like number of multiple cohorts;
executing by one or more computers a classification model using the training data to classify a first group of individuals in relation to a second, different group of individuals, with the first and second groups of individuals being of the multiple groups of individuals, and different ones of the multiple cohorts;
determining by one or more computers points of intersection of words between the first group and the second group of individuals;
determining by one or more computers words in the training data that are unique to the first group of individuals to provide isolated statistically predictive terms for the first group to identify predictive features for a first cohort; and
determining by one or more computers words in the training data that are unique to the second group of individuals to provide isolated statistically predictive terms for the second group to identify predictive features for a second one of the cohorts.
2. The method of claim 1 further comprising:
determining by the one or more computers a prediction of mental health status of the groups.
3. The method of claim 2 wherein the prediction is of suicidality risk.
4. The method of claim 2 wherein the prediction is suicidal ideation.
5. The method of claim 1 further comprising applying by the one or more computers a military veteran population dataset for suicidality prediction.
6. The method of claim 1 further comprising;
forming by the one or more computers ‘contextualized word pairs’ to improve predictive accuracy.
7. The method of claim 1 wherein determining the model comprises
applying a machine learning system, including Bayesian algorithms to the dataset to predict mental state.
8. The method of claim 1 further comprises
applying a machine learning system that includes genetic algorithms and genetic programming systems to predict mental state.
9. A computer implemented process comprises:
producing by one or more computers plural sets of training data that comprise words associated with multiple groups of individuals, with the multiple groups being of individuals having different mental states that define multiple cohorts;
executing by one or more computers a classification model using the training data to classify a first group of individuals in relation to a second, different group of individuals, with the first and second groups of individuals being of the multiple groups of individuals;
determining by one or more computers points of intersection of words between the first group and the second group of individuals;
determining by one or more computers words in the training data that are unique the first group of individuals and the second group of individuals to provide isolated statistically predictive terms for the first and second groups to identify predictive features for corresponding first and second cohorts; and
applying a workflow process to reduce complexity in classification of the first and second groups:
producing by the one or more computers a visualization of mental state cohorts for classification of risk.
10. The method of claim 9 further comprising:
determining by the one or more computers a prediction of mental health status of the groups.
11. A computer program product tangibly stored on a computer readable hardware storage device, the computer program product for mental state classification comprises instructions for causing a processor to:
produce plural sets of training data that comprise words associated with multiple groups of individuals, with the multiple groups of individuals having different classifications of mental states that define a like number of multiple cohorts;
execute a classification model on each of the sets of training data to classify a first group of individuals in relation to a second, different group of individuals, with the first and second groups of individuals being of the multiple groups of individuals, and different ones of the multiple cohorts;
determine points of intersection of words between the first group and the second group of individuals;
determine words in the training data that are unique to the first group of individuals to provide isolated statistically predictive terms for the first group to identify predictive features for a first cohort; and
determine words in the representations of the training data that are unique to the second group of individuals to provide isolated statistically predictive terms for the second group to identify predictive features for a second one of the cohorts.
12. The product of claim 11 further comprising instructions to:
determining by the one or more computers a prediction of mental health status of the groups.
13. The product of claim 11 wherein the prediction is of suicidality risk.
14. The product of claim 11 wherein the prediction is suicidal ideation.
15. The product of claim 11 further comprising applying by the one or more computers a military veteran population dataset for suicidality prediction.
16. The product of claim 11 further comprising instructions to:
form contextualized word pairs to improve predictive accuracy.
17. The product of claim 1 wherein determining the model comprises
apply a machine learning system, including Bayesian algorithms to the dataset to predict mental state.
18. The product of claim 11 further comprising instructions to
apply a machine learning system that includes genetic algorithms and genetic programming systems to predict mental state.
19. Apparatus, comprising:
a processor;
a memory coupled to the processor; and
a computer readable storage device storing a computer program product for mental state classification, the computer program product comprises instructions for causing the processor to:
produce plural sets of training data that comprise words associated with multiple groups of individuals, with the multiple groups of individuals having different classifications of mental states that define a like number of multiple cohorts;
execute a classification model on each of the sets to classify a first group of individuals in relation to a second, different group of individuals, with the first and second groups of individuals being of the multiple groups of individuals, and different ones of the multiple cohorts;
determine points of intersection of words between the first group and the second group of individuals;
determine words in the sets of the training data that are unique to the first group of individuals to provide isolated statistically predictive terms for the first group to identify predictive features for a first cohort; and
determine words in the sets of the training data that are unique to the second group of individuals to provide isolated statistically predictive terms for the second group to identify predictive features for a second one of the cohorts.
20. The apparatus of claim 19 wherein the product further comprising instructions to
apply a machine learning system that includes genetic algorithms and genetic programming systems to predict mental state.