I would like the detect a specific sequence of natural language elements inside a larger sequence, using a primer. The recognition algorithm should be able to deal with large portion of errors inside the sequence.
An example:
sequence : trrrrurrrrrrrrrblarrrktrrrrrrr
primer : blast
processing ---> blasrrrkt + remaining
I could not find efficient algorithm able to do it.
Than kyou very much in advance
Regards,
gioele
These may helpful for you
https://www.youtube.com/watch?v=v_L9jR8P-54
https://www.youtube.com/watch?v=7Kv1dvP72Oo
This is a great question with lots of possible answers. In addition to the helpful answer given by @Shafaq Nisar, there is a bit more to add.
One promising approach to answering your question is in the form generative sequences in
SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient
https://www.researchgate.net/publication/308324937_SeqGAN_Sequence_Generative_Adversarial_Nets_with_Policy_Gradient
For a detailed algorithm, see page 4.
Article SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient
@Shafaq Nisar thank you very much
@James F Peters Thank you very much for the useful hint. The reference is extremely interesting
@Philippe Fournier Viger you are right, the dataset I am analyzing is a sequence of process procedures (natural language tag) under the influence of uncertainty (failure, repetition, ecc,..) related to pharmaceutical manufacturing. For this two reasons unsupervised learning will not really help me. The main goal of this task is not mining, but filtering considering uncertainty. Of course a hybridization of stochastic and deterministic approach would help but requires too high computational costs for its purpose.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Algorithms