I have been tasked to build a multivariate time series forecast using Python. I am still flexible as to the best algorithm to use, however, have identified VAR, LSTM and Facebook Prophet.
I require the following help:
- identify and confirm my target variable
- identify the best algorithms to use based on my data (+/- 60 features)
- write the python code to build out the multivariate time series forecast in python
- need to complete this by 29 Nov
The fields and explanation below describe my dilemma in identifying my target variable.
Attorney_Code - Attorney nominated by the bank to get the home loan registered
Customer_Code - Each customer gets a unique code based on the loan applied for
Application_Code - customer can make several loan applications
Attorney_Date - Date when the loan was received by the attorney
Attorney_Amount - Value of the loan as at the attorney date
Registered_Date - Date when the loan finally gets registered by the attorney
Registered_Amt - final value of the loan paid to the customer
Loan_Type - specifies the purpose of the loan
Region - specifies the geographic location of the customer applying for the loan
Based on the Attorney_Date (known value), I need to be able to predict the Registered_Date (unknown value). However, many of the multivariate time series problems that I have come across use the target variable (i.e. Registered Date) to make further predictions. However, here I need to use the Attorney_date and forecast the Registration_date and thus my confusion of my target variable. In the main dataset there are +/- 60 fields and I also need to understand the best algorithm to use in doing the multivariate time series forecast in Python.
Hi, I have done time series forecasting multiple times with different clients. I have won bronze medal for my time series prediction on kaggle. For info check my profile and contact me.
8 freelancere byder i gennemsnit $123 timen for dette job
Hello, Hope you are doing well. I believe I can help you achieve your task. The project can tackle by different points of view. Hope we can have a chat and discuss further. Best,