Sales Predictions

Papad

Hello everyone,
I am working on a dataset which contains sales of some products. The format is like:
           product1_sales |  product2_sales |  product3_sales | factor1 | factor2
Date | 
I want to predict future sales, and the only thing I have are the factors for future.
What do you suggest me? Do you think that a linear model is ok?
I also wanted to ask, how can I do this with auto model? I can predict only the data I already know, so have I got to 
add some empty queries to my dataset in order to predict them?


Secondly, in another dataset , I have this kind of format:
                   Speed | Pressure | Temperature
Timestamp|

This is about a kind of machines. I also have another dataset which gives alarms if an event happens in a specific timestamp.
So I want to find correlations at first, and then I want to built a model in order to understand every time if something will happen before I have an alarm, in order not to stop the machine. Generally I want my model to recognize the flow of my data and predict what is gonna happen. 
To sum up, how can I say to my model for example that if speed is under 10 and pressure over 5 I want you to recognize it as a problem of a kind?
Thanks in advance.

varunm1

Hello @Papad

Looks like you are looking to forecast which can be done by time series forecasting techniques in rapidminer. I don't think the current automodel deals with time series forecasting as it is not distinguishing time series data from regular data. There is a similar discussion on forecasting below. As @jacobcybulski informed, you can use Deep learning algorithms like LSTM as well.

https://community.rapidminer.com/discussion/comment/58999#Comment_58999

Hope this helps thanks

Telcontar120

You should take a look at the time series operators and their associated tutorials.  There are a lot of different approaches you could take to this type of data, but the time series operators are going to be your best starting point.