Media names matching and normalization
Methods and apparatus to facilitate matching of names for same media content are disclosed. Example methods include analyzing first data associated with first media content and, when a program name/identifier and/or episode name/identifier is not identified in the first data, supplementing the data with second data to form third data and processing the third data with respect to fourth data associated with second media content and calculating a composite match score including a program match score and an episode match score based on processing the third data with respect to the fourth data. When the first media content is determined to match the second media content based on the processing of the third data with respect to the fourth data and a comparison of the composite match score to a threshold is satisfied, a normalized media name is generated for the first media content and the second media content.
1. A system comprising:
memory to store instructions; and
at least one processor to execute the instructions to at least:
extract media names of at least two types from a first data set to form a set of extracted media names;
apply pattern matching to the set of extracted media names and the first data set to normalize a name for two or more media content items represented in the first data set to form a set of normalized names using at least a first scoring pass and a second scoring pass, the first scoring pass applying a first weighting and the second scoring pass applying at least one of the first weighting or a second weighting, wherein, when a score of the first pass is greater than a threshold, the second scoring pass applies the first weighting, and, when the score of the first pass is less than the threshold, the second scoring pass applies the second weighting;
train a pattern matching model to perform the pattern matching based on the pattern matching applied to the set of extracted media names and the first data set to form the set of normalized names;
output the set of normalized names associated with the first data set; and
deploy the trained pattern matching model for a second data set.
2. The system of claim 1 , wherein the pattern matching model includes at least one of machine learning or deep learning.
3. The system of claim 1 , wherein the first data set includes a media data file, the media data file to be analyzed by the at least one processor including a match service to normalize the name for each media content item.
4. The system of claim 3 , wherein the match service is to communicate with an affiliated provider to form the set of normalized names.
5. The system of claim 1 , wherein the at least one processor is to apply pattern matching by generating at least one of a program match score, an episode match score, or a composite match score among the set of extracted media names to form the set of normalized names.
6. The system of claim 5 , wherein the composite match score is associated with one of three classifications: a good match, a possible match or a no match, wherein a good match corresponds to a composite match score greater than or equal to a first threshold, a possible match corresponds to a composite match score greater than or equal to a second threshold less than the first threshold, and a no match corresponds to a composite match score less than the second threshold.
7. The system of claim 1 , wherein the at least one processor is to apply pattern matching to form the set of normalized names using meta-data extracted the first data set and the set of extracted media names.
8. The system of claim 1 , wherein the media names of at least two types include at least two of a) linear media names, b) video-on-demand media names, or c) digital media names.
9. A non-transitory computer readable storage medium comprising instructions that, when executed, cause at least one processor to:
extract media names of at least two types from a first data set to form a set of extracted media names;
apply pattern matching to the set of extracted media names and the first data set to normalize a name for two or more media content items represented in the first data set to form a set of normalized names using at least a first scoring pass and a second scoring pass, the first scoring pass applying a first weighting and the second scoring pass applying at least one of the first weighting or a second weighting, wherein, when a score of the first pass is greater than a threshold, the second scoring pass applies the first weighting, and, when the score of the first pass is less than the threshold, the second scoring pass applies the second weighting;
train a pattern matching model to perform the pattern matching based on the pattern matching applied to the set of extracted media names and the first data set to form the set of normalized names;
output the set of normalized names associated with the first data set; and
deploy the trained pattern matching model for a second data set.
10. The computer readable storage medium of claim 9 , wherein the pattern matching model includes at least one of machine learning or deep learning.
11. The computer readable storage medium of claim 9 , wherein the first data set includes a media data file, the media data file to be analyzed by a match service to normalize the name for each media content item.
12. The computer readable storage medium of claim 11 , wherein the match service is to communicate with an affiliated provider to form the set of normalized names.
13. The computer readable storage medium of claim 9 , wherein the instructions, when executed, cause the at least one processor to apply pattern matching by generating at least one of a program match score, an episode match score, or a composite match score among the set of extracted media names to form the set of normalized names.
14. The computer readable storage medium of claim 13 , wherein the composite match score is associated with one of three classifications: a good match, a possible match or a no-match, wherein a good match corresponds to a composite match score greater than or equal to a first threshold, a possible match corresponds to a composite match score greater than or equal to a second threshold less than the first threshold, and a no match corresponds to a composite match score less than the second threshold.
15. The computer readable storage medium of claim 9 , wherein the instructions, when executed, cause the at least one processor to apply pattern matching to form the set of normalized names using meta-data extracted from the first data set and the set of extracted media names.
16. A method comprising:
extracting, using at least one processor, media names of at least two types from a first data set to form a set of extracted media names;
applying, using the at least one processor, pattern matching to the set of extracted media names and the first data set to normalize a name for two or more media content items represented in the first data set to form a set of normalized names using at least a first scoring pass and a second scoring pass, the first scoring pass applying a first weighting and the second scoring pass applying at least one of the first weighting or a second weighting, wherein, when a score of the first pass is greater than a threshold, the second scoring pass applies the first weighting, and, when the score of the first pass is less than the threshold, the second scoring pass applies the second weighting;
training, using the at least one processor, a pattern matching model to perform the pattern matching based on the pattern matching applied to the set of extracted media names and the data set to form the set of normalized names;
outputting, using the at least one processor, the set of normalized names associated with the first data set; and
deploying, using the at least one processor, the trained pattern matching model for a second data set.