Gap-filling issue using ML

Hello experts,

I am new to R and Machine Learning, and I am hopeful that you can assist me with the issue I am facing. I have been endeavoring to employ Artificial Neural Networks (ANNs) for filling missing data in the target variable (T1), which has corresponding predictor variables, namely P1, P2, P3, P4, and P5.

My objective is to predict the target variable across the time period for which predictor variable data is available. I tried and excluded the "missing" months from the training and testing datasets to construct the model, which I subsequently use to predict these absent months. Nevertheless, I am uncertain whether my approach is correct or not because the model performance is very low and not improving.

Do any of you have example code or practical experience with gap-filling datasets that could guide me in addressing this issue? I have been stuck on this problem for a few months now, and I am hoping that your insightful guidance will help me overcome this hurdle.

Thank you for your assistance.

If I had missing data to impute, I would first try Multivariate Imputation by Chained Equations • mice

Yes, it is indeed possible to fill the short-term gaps. However, in the dataset I am working with, there are both short and long-term gaps. The primary focus of this task is to effectively fill the long-term gaps, spanning, for instance, an 11-month period.

does the mean that you are familiar with MICE, and know it to be inappropriate for something to do with the duration of the missing periods ? or that you have attempted to use it, and found it to be the case ?
I'm having a hard time relating the content of your post, to my statement.

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.