I am a beginner in the ML/data science world.

I have a group of training data with the following characteristics.

Each training data is located in a separate file (this is just for elaborating the problem, not directly related to the question).
Each of these files contains tab-separated values of variables that progress over the course of an event - from the epoch of the event to the end of the event. For example,

1 5 8 3
9 4 3 11

... and so on.

Here 1, 5, 8, and 3 are values of 4 different variables at the event epoch. 9, 4, 3, and 11 are the values of the same variables after a certain progression of the event, and so on.

Each of these data is correlated to the values of all variables of the previous row. For example, 9 (Row 2, Col 1) is dependent on 1, 5, 8, 3 (each value of Row 1). Same for value in any other cell in Row 2.

I have many instances (i.e., files) of such training data.
I would like to train a generative network with the available training data to produce a data set that starts from an event epoch to the end of the event.

Which ML algorithm should I use?

I saw in other posts people have recommended using LSTM for such time series training data. I also saw recommendations for using any supervised GAN algorithm. If someone has solved a similar problem, will greatly appreciate any feedback and pointers to a code reference.

________________________
Pursuing data science certification in bangalore.