This is a multi-step job involving machine learning (ML) with recurrent neural networks (RNN) and Long Short-Term Memory (LSTM) algorithms. (UPDATE: I've now learned that our data set (30 years of monthly data = only 360 sets of data) is too small for RNN and LSTM, so now the steps below have been updated.)
Step 1 (this step - UPDATED) will be evaluating the problem and conceptualizing how to solve it. The problem is described below. It will not involve writing code but should explain which tool you would use to address the task.
Step 2 will then involve writing and testing the code, and we'll discuss the effort necessary for this step after Step 1 is completed.
For our machine learning task, we will be using economic time-series data. We'll have about 30 years of monthly data, and each period will have about 30 values for things like GDP (gross domestic product), size of the workforce, housing starts, interest rates, etc.
STEP 1 (UPDATED) is to determine which ML algorithms will best allow us to predict future values AND perform "what if" analysis.
For example, we want to input specific values for the interest rate and for government spending (two independent values) for the next 12 months and then rerun the algorithm and see the result on inflation (a key dependent value).
Step 1 would be to evaluate this task and propose how to accomplish it. For example, you might say that all we need to do is use a certain command in python to insert the values. Or, at the other end of the spectrum, you might say that we need to alter the code in an algorithm to insert the new values.
25 pekerja bebas membida secara purata $180 untuk pekerjaan ini
Hello, I am a machine learning expert. I would use Keras for your problem as this library has some wonderful neural network to use on Time-Series data. I am excited to do this project. Best Regards, Borut
Hey, I am a DL / ML professional with 5 years of NLP experience. I can do this task. I have also written papers . Let us talk soon on this. Thanks, Pranay