Content modification using natural language processing to include features of interest to various groups

1. A method of modifying content associated with an item comprising:

determining, via a processor, features of interest of the item to a plurality of different groups based on user comments produced by members of the plurality of different groups, wherein the members within each group have a common characteristic;

identifying, via the processor, the features of interest to each group within the content associated with the item; and

modifying, via the processor, the content associated with the item by balancing the features of interest to the plurality of different groups within the content associated with the item, wherein the balancing includes:

scrolling the content associated with the item on a display, wherein a size of the content associated with the item exceeds a presentation size for the display; and

adjusting a quantity of the features of interest to one or more of the plurality of different groups in a scrolled portion of the content associated with the item presented on the display, wherein the adjusted scrolled portion of the content associated with the item includes a quantity of the features of interest to each group that is within a tolerance of a predetermined number of features from quantities of features of interest for others of the plurality of different groups.

2. The method of claim 1 , further comprising:

determining, via the processor, that the content associated with the item is of greater interest to one of the plurality of different groups based on a distance between entities within the content associated with the item and sets of words describing the plurality of different groups.

3. The method of claim 1 , wherein determining the features of interest of the item to the plurality of different groups comprises:

determining a set of reference features for the item;

learning vector representations of words using the user comments produced by one or more members of a group; and

selecting one or more reference features as the features of interest to the group based on distances between the learned vector representation corresponding to each reference feature within the set of reference features and the learned vector representation corresponding to the item.

4. The method of claim 3 , wherein selecting one or more reference features as the features of interest to the group comprises:

ranking the reference features based on the distances, wherein the reference features with a shorter distance to the item are assigned a higher ranking; and

selecting one or more of the ranked features in order of the ranking as the features of interest to the group.

5. The method of claim 1 , wherein determining the features of interest of the item to the plurality of different groups comprises:

classifying the user comments into the plurality of different groups using a machine learning classifier; and

re-training the machine learning classifier based on the features of interest to the plurality of different groups.

6. The method of claim 1 , wherein modifying the content associated with the item comprises:

determining the features of interest to the plurality of different groups absent from the content associated with the item; and

incorporating one or more of the absent features of the plurality of different groups into the content associated with the item, wherein the quantity of the features of interest to each group is balanced within the content associated with the item, and wherein incorporating the one or more absent features of the plurality of different groups comprises:

generating one or more sentences for each of the one or more absent features of the plurality of different groups; and

inserting the one or more sentences into the content associated with the item.

7. The method of claim 1 , wherein the predetermined number is in a range of one to five.

8. A system of modifying content associated with an item comprising:

at least one processor configured to:

determine features of interest of the item to a plurality of different groups based on user comments produced by members of the plurality of different groups, wherein the members within each group have a common characteristic;

identify the features of interest to each group within the content associated with the item; and

modify the content associated with the item by balancing the features of interest to the plurality of different groups within the content associated with the item, wherein the balancing includes:

scrolling the content associated with the item on a display, wherein a size of the content associated with the item exceeds a presentation size for the display; and

adjusting a quantity of the features of interest to one or more of the plurality of different groups in a scrolled portion of the content associated with the item presented on the display, wherein the adjusted scrolled portion of the content associated with the item includes a quantity of the features of interest to each group that is within a tolerance of a predetermined number of features from quantities of features of interest for others of the plurality of different groups.

9. The system of claim 8 , wherein the at least one processor is further configured to:

determine that the content associated with the item is of greater interest to one of the plurality of different groups based on a distance between entities within the content associated with the item and sets of words describing the plurality of different groups.

10. The system of claim 8 , wherein determining the features of interest of the item to the plurality of different groups comprises:

determining a set of reference features for the item;

learning vector representations of words using the user comments produced by one or more members of a group; and

selecting one or more reference features as the features of interest to the group based on distances between the learned vector representation corresponding to each reference feature within the set of reference features and the learned vector representation corresponding to the item.

11. The system of claim 10 , wherein selecting one or more reference features as the features of interest to the group comprises:

ranking the reference features based on the distances, wherein the reference features with a shorter distance to the item are assigned a higher ranking; and

selecting one or more of the ranked features in order of the ranking as the features of interest to the group.

12. The system of claim 8 , wherein determining the features of interest of the item to the plurality of different groups comprises:

classifying the user comments into the plurality of different groups using a machine learning classifier; and

re-training the machine learning classifier based on the features of interest to the plurality of different groups.

13. The system of claim 8 , wherein modifying the content associated with the item comprises:

determining the features of interest to the plurality of different groups absent from the content associated with the item; and

incorporating one or more of the absent features of the plurality of different groups into the content associated with the item, wherein the quantity of the features of interest to each group is balanced within the content associated with the item, wherein incorporating the one or more absent features of the plurality of different groups comprises:

generating one or more sentences for each of the one or more absent features of the plurality of different groups; and

inserting the one or more sentences into the content associated with the item.

14. A computer program product for modifying content associated with an item, the computer program product comprising one or more computer readable storage media having program instructions collectively stored on the one or more computer readable storage media, the program instructions executable by a processor to cause the processor to:

determine features of interest of the item to a plurality of different groups based on user comments produced by members of the plurality of different groups, wherein the members within each group have a common characteristic;

identify the features of interest to each group within the content associated with the item; and

modify the content associated with the item by balancing the features of interest to the plurality of different groups within the content associated with the item, wherein the balancing includes:

scrolling the content associated with the item on a display, wherein a size of the content associated with the item exceeds a presentation size for the display;

adjusting a quantity of the features of interest to one or more of the plurality of different groups in a scrolled portion of the content associated with the item presented on the display, wherein the adjusted scrolled portion of the content associated with the item includes a quantity of the features of interest to each group that is within a tolerance of a predetermined number of features from quantities of features of interest for others of the plurality of different groups.

15. The computer program product of claim 14 , wherein the program instructions further cause the processor to:

determine that the content associated with the item is of greater interest to one of the plurality of different groups based on a distance between entities within the content associated with the item and sets of words describing the plurality of different groups.

16. The computer program product of claim 14 , wherein determining the features of interest of the item to the plurality of different groups comprises:

determining a set of reference features for the item;

learning vector representations of words using the user comments produced by one or more members of a group; and

selecting one or more reference features as the features of interest to the group based on distances between the learned vector representation corresponding to each reference feature within the set of reference features and the learned vector representation corresponding to the item.

17. The computer program product of claim 16 , wherein selecting one or more reference features as the features of interest to the group comprises:

ranking the reference features based on the distances, wherein the reference features with a shorter distance to the item are assigned a higher ranking; and

selecting one or more of the ranked features in order of the ranking as the features of interest to the group.

18. The computer program product of claim 14 , wherein determining the features of interest of the item to the plurality of different groups comprises:

classifying the user comments into the plurality of different groups using a machine learning classifier; and

re-training the machine learning classifier based on the features of interest to the plurality of different groups.

19. The computer program product of claim 14 , wherein modifying the content associated with the item comprises:

determining the features of interest to the plurality of different groups absent from the content associated with the item; and

incorporating one or more of the absent features of the plurality of different groups into the content associated with the item, wherein the quantity of the features of interest to each group is balanced within the content associated with the item.

20. The computer program product of claim 19 , wherein incorporating the one or more absent features of the plurality of different groups comprises:

generating one or more sentences for each of the one or more absent features of the plurality of different groups; and

inserting the one or more sentences into the content associated with the item.