I have a large amount of data. I need to make a neural network that trains itself from an excel data sheet and validates its output from another excel data set. There will be an output coloumn in the validate excel file. The model needs to give the output near about the givern row.
You are mentioning two different problems:
- data gathering: here you want to get your data from an excel file for both the training and the validation
- modeling your data: here you want to do it with a neural network.
For the data gathering, I suggest you to use pandas, a python library which can extract data from excel files with
pandas.read_excel as well as many other source (csv files, parquet files). Then once you have the data in python you can try to understand them with a neural network. The book I used to learn it was handson machine learning by Aurelien Gueron (https://github.com/ageron/handson-ml). Hope you will find this useful.
Before going to neural network you should also have a look at scikit learn a python library which is used for machine learning when you do not have enough data to train a neural network, if you are getting you data from excel files I guess you have less than 100 thousand data rows which might be a bit too few to train a neural net.
You have to break your task in 3 parts:
- load the data
- train the model
- use the model on validation data
The first and third tasks identical in your case, I think - you have to find a way to load the excel file. As mentioned in the other answer
pandas.read_excel is probably the function you need.
- If both excel files have the same structure, then you are good to go
- If the validation file is different, you have to load it and manually adjust the column names (and the set of columns that exist) in order to match the training data. E.g. the model would expect the same input
If part of your task is to also output the prediction in the validation file, in a separate column, then see how to use
pandas.DataFrame.to_excel. This would be a subtask for task
3 or maybe a new task on itself: save the prediction in an output file.