Before adding the new Data to your Dataset you should add a column that can be used to filter the data so it will score only the desired data. (Example add a column called "Score" which is a Boolean Data Type column 0 = exclude, 1 = include).
Then you could add new Data to you Dataset
For the already used Data to train and generate the Model the values of "Score" in all rows would be 0. However for the newly added Data to the Dataset the values of "Score" in all rows would be 1.
Then when running the Scoring Job use the Filter on column Score with 1= include.
I hope this helps.
Actually, this will cause you to have old scoring results returned with your newly scored data because every time you score a new job, you'll always be scoring against rows that have a score value of 0. The better way (and to ensure you're only scoring new data) is to keep an incrementing property on a Thing and use your incrementing property to set your score value. For example, Scoring1, Scoring2, Scoring3, etc. When you upload your new data, assign your incremented scoring value as the filter. Then when you score your data, you only score against the specific filter (Scoring4, for example). Now the ONLY results you will get back will be related to the data uploaded with the Scoring4 filter.