Method and system for creating word-level differential privacy using feature hashing techniques

1. A method of creating word-level differential privacy with the hashing trick to protect confidentiality of a textual data, the method comprising:

receiving, by a differential privacy creating device, a list of a plurality of hashes with a weight (or weights) associated with each of the plurality of hashes;

updating, by said differential privacy creating device, said list with new hashes that are within the range of allowable hash values but not included in said received list of hashes;

updating, by said differential privacy creating device, said list with a new weight to each of said plurality of hashes that are missing said weight, wherein in cases where said received list of hashes is a newer version of a previously-existing list of hashes, updating said list with adjusted weights with respect to said hashes weights in said previously-existing list of hashes;

fitting, by said differential privacy creating device, a probability distribution to said list of said weights of said plurality of hashes; and

generating, by said differential privacy creating device, said new weights and said adjusted weights based on sampling of said probability distribution.

2. The method as claimed in claim 1 , wherein a corpus of textual data is converted to said plurality of hashes.

3. The method as claimed in claim 1 , wherein said probability distribution is continuous when said weight of said plurality of hashes are continuous, and wherein said probability distribution is discrete value when said weight of said plurality of hashes are discrete values.

4. The method as claimed in claim 1 , wherein said probability distribution is multi-dimensional when number of types of said weights is more than one.

5. The method as claim in claim 4 , wherein said probability distribution is 2d dimensional when type of said weight is d dimensional and previously-existing weights are being adjusted.

6. A non-transitory computer-readable medium storing computer-executable instructions for generating strategy and roadmap for end-to-end information technology (IT) infrastructure cloud implementation, the computer-executable instructions configured for:

receiving a list of a plurality of hashes with a weight (or weights) associated with each of the plurality of hashes;

updating said list with new hashes that are within the range of allowable hash values but not included in said received list of hashes;

updating said list with a new weight to each of said plurality of hashes that are missing said weight, wherein in cases where said received list of hashes is a newer version of a previously-existing list of hashes, updating said list with adjusted weights with respect to said hashes weights in said previously-existing list of hashes;

fitting a probability distribution to said list of said weights of said plurality of hashes; and

generating said new weights and said adjusted weights based on sampling of said probability distribution.

7. The non-transitory computer-readable medium as claimed in claim 6 , wherein said probability distribution is continuous when said weight of said plurality of hashes are continuous, and wherein said probability distribution is discrete value when said weight of said plurality of hashes are discrete values.

8. The non-transitory computer-readable medium as claimed in claim 6 , wherein said probability distribution is multi-dimensional when the number of types of said weights is more than one.

9. The non-transitory computer-readable medium as claimed in claim 6 , wherein said probability distribution is 2d dimensional when type of said weight is d dimensional and previously-existing weights are being adjusted.