public repo edits

Author: ktoliver

articles/cognitive-services/personalizer/concepts-features.md

@@ -128,8 +128,7 @@ It is possible to improve your feature sets by editing them to make them larger
 
 For example, a timestamp down to the second is a very sparse feature. It could be made more dense (effective) by classifying times into "morning", "midday", "afternoon", etc.
 
-Location information also typically benefits from creating broader classifications. For example, a Latitude-Longitude coordinate such as
-Lat: 47.67402° N, Long: 122.12154° W is too precise, and forces the model to learn latitude and longitude as distinct dimensions. When you are trying to personalize based on location information, it helps to group location information in larger sectors. An easy way to do that is to choose an appropriate rounding precision for the Lat-Long numbers, and combine latitude and longitude into "areas" by makeing them into onestring. For example, a good way to represent 47.67402° N, Long: 122.12154° W in regions approximately a few kilometers wide would be "location":"34.3 , 12.1".
+Location information also typically benefits from creating broader classifications. For example, a Latitude-Longitude coordinate such as Lat: 47.67402° N, Long: 122.12154° W is too precise, and forces the model to learn latitude and longitude as distinct dimensions. When you are trying to personalize based on location information, it helps to group location information in larger sectors. An easy way to do that is to choose an appropriate rounding precision for the Lat-Long numbers, and combine latitude and longitude into "areas" by making them into one string. For example, a good way to represent 47.67402° N, Long: 122.12154° W in regions approximately a few kilometers wide would be "location":"34.3 , 12.1".
 
 
 #### Expand feature sets with extrapolated information