Method for predictive analytics
A computer implemented method for predicting the behaviour of a population is described. The method comprising: analysing qualitative data for a transactional event encountered by at least one member of the population, to determine a topic of the qualitative data; identifying one or more emotions for the at least one member of the population associated with the qualitative data; determining a transactional event outcome for the identified emotion and associated qualitative data; assigning a score for the at least one member of the population; repeating the above steps for a plurality of events for the at least one member of the population to provide a population database; updating the score as the steps are repeated; and storing a score history and updated score in the population database; and analysing the population database and providing an output to predict the behaviour of a member of the population for new qualitative data.
1 . A computer implemented method for predicting the behaviour of one or more members of a population using predictive analytics method performed by one or more computer processor comprising the steps of:
receiving and storing in a population history database, structured and unstructured qualitative and quantitative data for one or more transactional events for one or more members of the population, where the population history database stores for each member, member identification, associated topic of qualitative and quantitative data, emotion, transactional event outcome, score and score history;
ingesting the structured and unstructured qualitative and quantitative data, including cleaning, integrating, and transforming the structured and unstructured qualitative and quantitative data to produce a structured dataset for modelling and predictive analysis;
analysing using Natural Language Processing the qualitative and quantitative transactional event data in the dataset for a transactional event encountered by at least one member of the population, to determine a topic of the qualitative and quantitative data, wherein the Natural Language processing comprises topic modelling, emotion analytics and sentiment analytics for calibration of the predictive analysis where the Natural Language Processing generates an ontology and taxonomy table, and then passes textual analyses to a behavioural analytics module, which generates a results table, and segment table, for use by the analytics module which stores predictive models;
using the behaviour analytics for identifying one or more emotions for the at least one member of the population associated with the qualitative and quantitative data;
determining using one or more computer processors a transactional event outcome for the identified emotion and associated qualitative and quantitative data using one or more predictive analytic techniques selected from correlation models, grouping models, explanatory models and predictive models;
assigning, using one or more computer processors a score for the at least one member of the population based on the transactional event outcome;
repeating the above analysing, identifying, determining and assigning steps for a plurality of transactional events for the at least one member of the population to provide updated entries for the population history database;
providing feedback from the scores about the performance of the models used for the predictive analytics to refine the models of the predictive analytic techniques;
updating the score for the at least one member of the population as the steps are repeated; and storing a score history and updated score in the population history database; and
analysing the population history database and providing an output to predict the behaviour of a member of the population for new qualitative and quantitative data.
2 . A computer implemented method as claimed in claim 1 , where the steps of the method are repeated for a plurality of members of the population.
3 . A computer implemented method as claimed in claim 1 , wherein the steps of the method are repeated at a set frequency.
4 . A computer implemented method as claimed in claim 3 , wherein the set frequency is one of: daily, weekly, monthly, or annually.
5 . A computer implemented method as claimed in claim 1 , where the qualitative data is a narrative.
6 . A computer implemented method as claimed in claim 1 wherein the narrative is at least one of: a written narrative or a spoken narrative.
7 . A computer implemented method as claimed in claim 6 , wherein the spoken narrative is received as an audio or video recording.
8 . A computer implemented method as claimed in claim 1 , wherein the one or more emotions is explicit from the qualitative and quantitative data.
9 . A computer implemented method as claimed in claim 1 , wherein the one or more emotions is implicit from the qualitative and quantitative data.
10 . A computer implemented method as claimed in claim 1 , further comprising obtaining metadata about the qualitative and quantitative data.
11 . A computer implemented method as claimed in claim 1 , wherein the one of more emotions identified are one of more of: anger, anticipation, contempt, disgust, fear, joy, love, sadness, or surprise.
12 . A computer implemented method as claimed in claim 1 , wherein the qualitative data is analysed to also identify a sentiment for the qualitative data.
13 . A computer implemented method as claimed in claim 1 , wherein, when the qualitative data from different members of the population contains common topics, the common topics are grouped for further analysis.
14 . A computer implemented method as claimed in claim 1 , further comprising determining an emotional profile for at least one member of the population.
15 . A method as claimed in claim 14 , wherein the emotional profile is determined for a plurality of members of the population.
16 . A computer implemented method according to claim 1 , wherein the steps are performed using cloud computing infrastructure.
17 . A computer implemented method according to claim 1 , further comprising the step of determining an increase or decrease of at least one specified emotion over a defined time period.
18 . A computer implemented method as claimed in claim 17 , wherein the time period is between 1 day to 90 days.
19 . A computer implemented method as claimed in claim 1 , further comprising the step of determining the time period over which a subject will revert to a baseline emotional state.
20 . A computer implemented method as claimed in claim 1 , further comprising the step of calculating a coefficient of emotional gain, which determines absolute changes in an arousal level of a subject.