Method and system for exchange of packets pertaining to an instrument
The present disclosure relates to a method and system for exchanging packets of information pertaining to an instrument. Data pertaining to the instrument is obtained from internal and external sources, which may be governmental or non-governmental. The obtained data undergoes a process of clustering and dimensional reduction to arrive at cleaned and optimised data attributes. A predictive model is built using those data attributes. A testing provision is included in the proposed method and system that allows for validation of the constructed model by using test data and comparing the predictions with actual values. Upon validation, the model predicts one or more packets of information that can have a bearing on the exchange of packets pertaining to the instrument.
1. A method for predicting a probability of an occurrence of an event pertaining to an instrument, said method comprising steps of:
retrieving, by a processor, instrument transaction data and any or a combination of one or more external and internal data attributes from a networked repository, and converting said data and data attributes into a plurality of variables by linking the internal data attributes with the external data attributes;
filtering, by the processor, one or more variables among the plurality of variables to get the one or more filtered variables containing information about default, wherein the filtering is performed based on similarity of information contained in attributes of the plurality of variables;
generating, by the processor, clusters of the filtered variables based on correlation amongst them, and clubbing the filtered variables based on the correlation;
transforming, by the processor, the variables from each cluster into variables, called components, to reduce dimensions of the generated clusters;
upon the reduction of the dimensions of the generated clusters, performing, by the processor, a stepwise variable selection by removing highly correlated variables among the components to generate uncorrelated variables or independent variables; and
predicting, by the processor, a probability of the occurrence of the event pertaining to the instrument, wherein said predictions are performed based on the uncorrelated variables, wherein the uncorrelated variables are formed before the step of predicting.
2. The method as claimed in claim 1 , wherein the external data attributes can be obtained from external governmental and non-governmental sources in a structured and unstructured way.
3. The method as claimed in claim 1 , wherein a back-testing process is carried out to verify the predictive capability and estimation of miss-classifications, by utilising sample data.
4. A system for predicting a probability of an occurrence of an event pertaining to an instrument, said system comprising:
a non-transitory storage device having embodied therein one or more routines operable to exchange packets in relation with the instrument; and
one or more processors coupled to the non-transitory storage device and operable to execute the one or more routines, wherein the one or more routines when executed cause the one to more processors to:
retrieve instrument transaction data and any or a combination of one or more external and internal data attributes from a networked repository, and converting said data and data attributes into a plurality of variables by linking the internal data attributes with the external data attributes;
filter one or more variables among the plurality of variables to get the one or more filtered variables containing information about default, wherein the filtering is performed based on similarity of information contained in attributes of the plurality of variables;
generate clusters of the filtered variables based on correlation amongst them, and clubbing the filtered variables based on the correlation;
transform the variables from each cluster into variables, called components, to reduce dimensions of the generated clusters;
upon the reduction of the dimensions of the generated clusters, perform a stepwise variable selection by removing highly correlated variables among the components to generate uncorrelated variables; and
predict a probability of the occurrence of the event pertaining to an instrument, wherein said predictions are performed based on the uncorrelated variables, wherein the uncorrelated variables are formed before the step of predicting.
5. The system as claimed in claim 4 , wherein the external data attributes can be obtained from one or more networked databanks, configured to receive and store information from external sources.