This is more of an idea, than request. Once the new API's are connected up, we should be able to run the futures (dependent) input around the clock by creating multiple requests that align with the various global markets and time zones. This is an order of magnitude better than the capability we have today. However, perhaps it could be improved ...
I'm thinking if we could "tag" each "independent" input, then we could possibly concatenate the inputs from various global markets (Asia, Europe, US, etc) into a single request. For example, if we had an Input-1 tag, we could map GOOG to Input-1, DAX (German index) to Input-1 and SSE (Asia index) to Input-1. Then, the Engine request would use Input-1 as one of the independent inputs instead of the actual ticker.
In theory, this would allow training for a specific futures prediction using a single request that include multiple overnight markets in the training.
This is more of an idea, than request. Once the new API's are connected up, we should be able to run the futures (dependent) input around the clock by creating multiple requests that align with the various global markets and time zones. This is an order of magnitude better than the capability we have today. However, perhaps it could be improved ...
I'm thinking if we could "tag" each "independent" input, then we could possibly concatenate the inputs from various global markets (Asia, Europe, US, etc) into a single request. For example, if we had an Input-1 tag, we could map GOOG to Input-1, DAX (German index) to Input-1 and SSE (Asia index) to Input-1. Then, the Engine request would use Input-1 as one of the independent inputs instead of the actual ticker.
In theory, this would allow training for a specific futures prediction using a single request that include multiple overnight markets in the training.