1

Topic: Forecasting on the contrary

All greetings!
The task:
To predict, for example, NOT action course on parameters, and to produce possibly optimal parameters at which the action course will have the necessary values (maximum/minimaln)...
On an input the same sampling, as at a prediction of course of the action (for example).
Where to dig, prompt?
What algorithms it is more preferable to the similar task?

2

Re: Forecasting on the contrary

SergeiSvetliy;
.

3

Re: Forecasting on the contrary

SergeiSvetliy wrote:

Where to dig, prompt?
What algorithms it is more preferable to the similar task?

You should solve the equation simply.  (= and
The decision can be analytical. Depends from  ().
If at you  model
(0, to,) = 0 + *
That decision is obvious. Numerical methods also are not necessary to you.

4

Re: Forecasting on the contrary

1) theoretical methods for the decision of the given task do not exist
2) the practical approach

#!/usr/bin/python3
# OPTIMALITY PREDICTOR
# Prerequisites: ask your line manager about optimal stock parameters
def get_optimal_stock_parameters (ask_min, ask_max, *ask_other_questions):
min_daily_rate = ask_min
max_daily_rate = ask_max
avg_daily_rate = (ask_min + ask_max)/2
other_optimal_parameters = ask_other_questions
res = {}
res [' Min Daily Rate '] = min_daily_rate
res [' Max Daily Rate '] = max_daily_rate
res [' Average Daily Rate '] = avg_daily_rate
res [' Other Optimal Stock Market Parameters'] = other_optimal_parameters
return res
# TEST PREDICTOR FUNCTION:
ask_min = 100
ask_max = 110
volatility ='High'
market_activity_level = ' High'
test = get_optimal_stock_parameters (ask_min, ask_max, volatility, market_activity_level)
print ("Optimal parameters:")
print ("Min Daily Rate: {:.2f}".format (test [' Min Daily Rate ']))
print ("Average Daily Rate: {:.2f}".format (test [' Average Daily Rate ']))
print ("Max Daily Rate: {:.2f}".format (test [' Max Daily Rate ']))
print ("Other important parameters: {}".format (test [' Other Optimal Stock Market Parameters']))

5

Re: Forecasting on the contrary

Above  any .
Tried to calculate separately on value and parameter, for example:
1... Parameter x1
1.... Parameter y1
...
Here it is possible
But so the correlation (possible) parameters is lost...
It is possible certainly naive , or a priori  for detection of a sheaf of groups of parameters, but...
It is all not that... ((
an average precisely does not approach... And if the anomalous surge to 9999 at mean values 1-10 then generally all flies?!...