System and method for providing prediction models for predicting changes to placeholder values
The present disclosure relates to a method and non-transitory machine-readable storage medium encoded with instructions for using prediction models for predicting values, the medium comprising instructions for receiving a plurality of entry identifiers, instructions for receiving a value for each of the plurality of entry identifiers, instructions for determining whether the value for each of the plurality of entry identifiers has changed and a magnitude of the change, instructions for building a model for predicting a time-to-value change, instructions for building a model for predicting a future magnitude of change, instructions for performing a simulation using the model for predicting the time-to-value change and the model for predicting the future magnitude of change and instructions for outputting a confidence interval.
1. A non-transitory machine-readable storage medium for using prediction models for predicting values related to medical insurance claims for one or more patients, the medium comprising:
instructions for receiving a plurality of entry identifiers for the one or more patients;
instructions for receiving a value for each of the plurality of entry identifiers;
instructions for determining whether the received value for each of the plurality of entry identifiers has changed and a magnitude of the change;
instructions for building a model for predicting a time-to-value change for costs of medical insurance claims for the one or more patients,
wherein predicting the time-to-value includes outputting a probability in relation to a time to change for each of the plurality of entry identifiers and
wherein the time to change is defined as
D
=
{
d
1
,
…
,
d
n
t
max
}
where d m =min(argmax i c m i , t max )-e m and wherein, d 1 is a first time-to-change value for a first entry identifier of the plurality of entry identifiers,
d
n
t
max
is a second time-to-change value for a second entry identifier of the plurality of entry identifiers at a maximum censoring time, c m i is a change-indicator for an entry identifier of the plurality of entry identifiers at a time i, t max is a censoring time, e m is an entry time for an entry identifier, and d m is a minimum time-to-change value;
instructions for building a model for predicting a future magnitude of change for the costs of the medical insurance claims, wherein predicting the future magnitude of change includes outputting a probability in relation to the future magnitude of change for each of the plurality of entry identifiers;
instructions for performing a simulation using both the model for predicting the time-to-value change and the model for predicting the future magnitude of change, and
instructions for outputting a confidence interval,
wherein the output probabilities and confidence interval comprises payment information for the costs of the medical insurance claims and a measure of margin of error.
2. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , the medium further comprising:
instructions for defining an entry time, the entry time being a first time when at least one of the plurality of entry identifiers was received.
3. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 2 , the medium further comprising:
instructions for defining a time to change, the time to change being a lag time between when at least one of the plurality of entry identifiers was received and when a change in the value for at least one of the plurality of entry identifiers was received.
4. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , wherein the change, C i is determined by
C
i
=
{
c
p
j
i
}
where
c
p
j
i
=
{
1
if
v
p
j
i
≠
v
p
j
i
+
1
0
else
,
where v is the value for each of the plurality of entry identifiers, and wherein c p j i is a change-indicator for a pair of successive values for each of the plurality of entry identifiers, v p j i is the value for a given entry identifier of the plurality of entry identifiers, and v p j i+1 is a successive value for the given entry identifier of the plurality of entry identifiers.
5. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , wherein the model for predicting a time-to-change model is a Poisson process model.
6. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , wherein the model for predicting a time-to-change model is a Cox regression model.
7. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , wherein the model for predicting the time-to-change outputs a probability that the value for each of the plurality of entry identifiers will change.
8. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , wherein the model for predicting the future magnitude of change is a first order statistic model of the magnitude of change.
9. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , wherein the magnitude of change, ΔV is defined as
Δ
V
=
{
v
p
(
t
m
+
d
m
)
-
v
p
t
m
}
for each of the plurality of entry identifiers with c p =1 in any C i , where v is the value for each of the plurality of entry identifiers, c is the change, t is the time value,
v
p
(
t
m
+
d
m
)
is the value for a first entry identifier of the plurality of entry identifiers at a minimum time period and a minimum time to change value, and
v
p
t
m
is the value for the first entry identifier of the plurality of entry identifiers at the minimum time period.
10. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 1 , wherein the simulation using the model for predicting the time-to-value change and the model for predicting the future magnitude of change is a Monte Carlo simulation.
11. The non-transitory machine-readable storage medium for using prediction models for predicting values of claim 10 , wherein the simulation changes the magnitude of change using the probability that the value for each of the plurality of entry identifiers will change and the magnitude of change.
12. A method for providing predictions using prediction models for predicting values related to medical insurance claims for one or more patients, the method comprising the steps of:
receiving a plurality of entry identifiers for the one or more patients;
receiving a value for each of the plurality of entry identifiers;
determining whether the received value for each of the plurality of entry identifiers has changed and a magnitude of the change;
building a model for predicting a time-to-value change for costs of medical insurance claims for the one or more patients,
wherein predicting the time-to-value includes outputting a probability in relation to a time to change for each of the plurality of entry identifiers, and
wherein the time to change is defined as
D
=
{
d
1
,
…
,
d
n
t
max
}
where d m =min(argmax i c m i , t max )-e m and wherein, di is a first time-to-change value for a first entry identifier of the plurality of entry identifiers,
d
n
t
max
is a second time-to-change value n tmax for a second entry identifier of the plurality of entry identifiers at a maximum censoring time, c m i is a change-indicator for an entry identifier of the plurality of entry identifiers at a time i, t max is a censoring time, e m is an entry time for an entry identifier, and d m is a minimum time-to-change value;
building a model for predicting a future magnitude of change for the costs of the medical insurance claims, wherein predicting the future magnitude of change includes outputting a probability in relation to the future magnitude of change for each of the plurality of entry identifiers;
performing a simulation using both the model for predicting the time-to-value change and the model for predicting the future magnitude of change, and
outputting a confidence interval,
wherein the output probabilities and confidence interval comprises payment information for the costs of the medical insurance claims and a measure of margin of error.
13. The method for providing predictions using prediction models for predicting values of claim 12 , the method further comprising the step of:
defining an entry time, the entry time being a first time when at least one of the plurality of entry identifiers was received.
14. The method for providing predictions using prediction models for predicting values of claim 13 , the method further comprising the step of:
defining a time to change, the time to change being a lag time between when at least one of the plurality of entry identifiers was received and when a change in the value for at least one of the plurality of entry identifiers was received.
15. The method for providing predictions using prediction models for predicting values of claim 13 , wherein the change, C′ is determined by
C
i
=
{
c
p
j
i
}
where
c
p
j
i
=
{
1
if
v
p
j
i
≠
v
p
j
i
+
1
0
else
,
where v is the value for each of the plurality of entry identifiers, and wherein c p j i is a change-indicator for a pair of successive values for each of the plurality of entry identifiers, v p j i is the value for a given entry identifier of the plurality of entry identifiers, and v p j i+1 is a successive value for the given entry identifier of the plurality of entry identifiers.
16. The method for providing predictions using prediction models for predicting values of claim 13 , wherein the model for predicting a time-to-change model is a Poisson process model.
17. The method for providing predictions using prediction models for predicting values of claim 13 , wherein the model for predicting a time-to-change model is a Cox regression model.
18. The method for providing predictions using prediction models for predicting values of claim 13 , wherein the model for predicting the time-to-change outputs a probability that the value for each of the plurality of entry identifiers will change.
19. The method for providing predictions using prediction models for predicting values of claim 13 , wherein the model for predicting the future magnitude of change is a first order statistic model of the magnitude of change.
20. The method for providing predictions using prediction models for predicting values of claim 13 , wherein the future magnitude of change, ΔV is defined as
Δ
V
=
{
v
p
(
t
m
+
d
m
)
-
v
p
t
m
}
for each of the plurality of entry identifiers with c p =1 in any C i , where v is the value for each of the plurality of entry identifiers, c is the change, t is the time value,
v
p
(
t
m
+
d
m
)
is the value for a first entry identifier of the plurality of entry identifiers at a minimum time period and a minimum time to change value, and
v
p
t
m
is the value for the first entry identifier of the plurality of entry identifiers at the minimum time period.
21. The method for providing predictions using prediction models for predicting values of claim 13 , wherein the simulation using the model for predicting the time-to-value change and the model for predicting the future magnitude of change is a Monte Carlo simulation.
22. The method for providing predictions using prediction models for predicting values of claim 21 , wherein the simulation changes the magnitude of change using the probability that the value for each of the plurality of entry identifiers will change and the magnitude of change.