Downloaded 83 times
More Related Content
![[BrightonSEO 2022] Unlocking the Hidden Potential of Product Listing Pages](https://cdn.slidesharecdn.com/ss_thumbnails/abuali-brightonseo2022slides-221003112121-157abb52-thumbnail.jpg?width=640&height=640&fit=bounds)
Similar to BrightonSEO - NLP for SEOs - How to optimise your content for BERT.pptx
![How to Salely Buylinkedin Connection-[Stap by Stap Guide].pdf](https://cdn.slidesharecdn.com/ss_thumbnails/howtosalelybuylinkedinconnection-stapbystapguide-251128160456-020f6973-thumbnail.jpg?width=640&height=640&fit=bounds)
BrightonSEO - NLP for SEOs - How to optimise your content for BERT.pptx
- 1. NLP for SEOs:How to optimise your content for BERT @jcbhaagen Josephine Haagen Journey Further https://www.slideshare.net/ JosephineHaagen
- 2. 2 What is NLP? ©2022 Journey Further.All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 3. How machines understand language. 3 ©2022Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 4. 4 Or, in SEOterms: Takes unstructured data and attempts to structure it. ©2022 Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 5. ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd 5 What can NLP do for you?
- 6. 6 A content strategyled by NLP ©2022 Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 7. 7 A content strategyled by NLP ©2022 Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 8. ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd 8 What did we do?
- 9. 9 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd 01 Prioritised key pages based on potential traffic
- 10. 10 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd 02 Optimised them (based on prioritised order)
- 11. 11 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd 03 Watched the line go up
- 12. 12 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd How does NLP work?
- 13. 13 Processing ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd INPUT TOKENISATION
- 14. 14 Processing ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd TOKENISATION POS TAG SALIENCE STEMS DEPENDENCY
- 15. 15 NLP in action ©2022Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 16. 16 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 17. 17 What do weknow about Alice? ©2022 Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Alice is swimming against the current. Noun (aux) V Verb Preposition Determiner Noun Parsed by Google’s NLP tool Parsed manually
- 18. 18 BERT ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 19. ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd 19 BERT NLU icon Token masking Next sentence prediction
- 20. 20 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd “In autumn the ______ fall from the trees.”
- 21. 21 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd At dinner, Ernie ordered a pizza and tiramisu. Do we still call it link juice? After a long day of work, Ernie went out to dinner with friends. IsNext NotNext
- 22. 22 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd How to optimise content with NLP
- 23. 23 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd “First-time buyer” Adjective (first-time) supporting a noun (buyer) Maximum three word phrase Noun is not also an adjective No numbers in the keyword
- 24. 24 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 25. 25 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd In the rugged Colorado Desert of California, there lies buried a treasure ship sailed there hundreds of years ago by either Viking or Spanish explorers. Some say this is legend; others insist it is fact. A few have even claimed to have seen the ship, its wooden remains poking through the sand like the skeleton of a prehistoric beast. Among those who say they’ve come close to the ship is small-town librarian Myrtle Botts. In 1933, she was hiking with her husband in the Anza-Borrego Desert, not far from the border with Mexico. Text Relevance rugged Colorado Desert of California 0.89 treasure ship 0.78 Anza- Borrego Desert 0.75 Spanish explorers 0.6
- 26. 26 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Villas in Spain are incredibly popular, and for good reason too! With beautiful beaches and a charming countryside, bustling cities and idyllic towns, a Spanish villa holiday will certainly be one to remember.
- 27. 27 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Topic Original Updated Difference holiday villas 0.00 0.58 0.58 Spanish villa properties 0.64 0.69 0.05 In Salient, salience is measured as a score out of 1.
- 28. 28 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Changes go through
- 29. 29 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Introduce focus keywords in the first paragraph Experiment with sentence structure Look at what the tool sees as relevant Do
- 30. 30 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Don’t Keyword stuff Write for robots
- 31. 31 So why use NLP? ©2022Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 32. 32 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd 01 Because you can 02 Because Google is 03 Because it works!
- 33.
- 34. 34 Appendix ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd
- 35. 35 ©2022 Journey Further. Allrights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Tools mentioned: SALIENT Google’s natural language API demo Understanding BERT: ● Open Sourcing BERT: State-of-the-Art Pre-training for Natural Language Processing ● BERT Explained: State of the art language model for NLP ● What’s My Line? Next Sentence Prediction in RunwayML with BERT ● BERT For Next Sentence Prediction
- 36. 36 Syntax & Semantics ©2022Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Syntax is the grammatical structure of a sentence, the order each word goes into. Semantics is the intent part of a sentence, how do the words in the sentence affect each other? How does the order in which they come in change your meaning? NLP turns syntax and semantics into its own data structure which specifies the relationship between words and phrases. This is known as ontology. NLU takes this ontology and attempts to prescribe intent to it. “Patient’s mother was involved in a car accident and injured her hand.”
- 37. 37 Parts of speech ©2022Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Nouns - keywords are usually nouns. Nouns are things. Verbs - doing words Adjectives - describing words Superlatives - adjectives on a scale. Best, worst, greatest, biggest, smallest. Adverbs - describes a word. Quickly, slowly, really, very. Prepositions - Links a noun to another noun. To, at, after, on. Conjunctions - Joins clauses, sentences, or words together. And, but, when.
- 38. 38 Sentence structures/syntax ©2022 Journey Further.All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Independent clause - a part of a sentence that could stand alone. A subject, a verb, and a complete idea. I like SEO. He is optimising a page. Dependent clause - Not a complete sentence. Sits with an independent clause to become a complete sentence. Because she said that, As I’m currently looking into this, Subject - Whatever’s doing the action in your sentence. I like SEO. Verb - the action in your sentence. I like SEO. Object - What receives the action in a sentence. I like SEO. Prepositional phrase - A phrase that begins with a preposition, and modifies a word in the sentence. I like SEO on smaller websites.
- 39. 39 Types of sentences ©2022Journey Further. All rights reserved. All related logos, assets messaging are trademarks of Journey Further Ltd Simple sentences - a sentence that could stand alone. A subject, a verb, and a complete idea. I like SEO. He is optimising a page. Compound sentences - A sentence that contains two or more independent clauses, they can be combined with a comma, a conjunction, or with a semicolon. I like SEO, I’m good at it. I like SEO, but I’m no good at it. Complex sentences - an independent clause and one dependent clause (a part of a sentence that can’t stand on its own). Although I like SEO, I don’t like site speed. Compound-complex sentences - a sentence that has at least two independent clauses and at least one dependent clause. These can be joined via commas or conjunctions, though they can get confusing if your comma game isn’t strong. I like SEO, even though I spent hours trying to edit a robots.txt, and my laptop crashed trying to run a crawl.
Editor's Notes
- #3 Include introduction slide - where did I start with this?
- #6 There's more - much more - you can do to create a content ideation that is based on actual insights and not just keyword research
- #7 Average rank has gone down across most of the pages we’ve optimised (this isn’t a magic bullet, using NLP doesn’t guarantee results just like the rest of SEO)
- #8 Average rank has gone down across most of the pages we’ve optimised (this isn’t a magic bullet, using NLP doesn’t guarantee results just like the rest of SEO)
- #17 Sentiment’s hokeyyyyy
- #18 Explain difference between adjective and noun
- #20 A machine learning technique, aimed at understanding intent and the relationship between queries. The blueprint for every NLP experiment since. Used in pretty much every English language query + 70 other languages across Google Search. Pre-trained on English-language Wikipedia and BooksCorpus - all the text it was trained on (its corpus) was in order and not shuffled. BERT is a pre-training model - you can use BERT as a base for your own NLP needs (provided you know Python!) There are two things that BERT does best: language modelling and next sentence prediction. Token masking: BERT was trained to predict the next word in a sentence, based off of the text that it was trained on. Bidirectional: BERT was the first NLP model that was fully bidirectional, able to use the context of both the words around masked tokens. BERT was made bidirectional by its token masking process - Input tokens were masked randomly, and the program would predict what their tokens are. Next sentence prediction: The program would be given a first sentence, then would have to guess if the next sentence given IsNext or is NotNext. 50% of second sentences would be a random sentence from the corpus, while the other half would be the actual sentence that followed the first one. This is an important process in understanding the relationship between sentences and how they work together.
- #27 Make keywords larger
- #36 Make keywords larger
- #37 Average rank has gone down across most of the pages we’ve optimised (this isn’t a magic bullet, using NLP doesn’t guarantee results just like the rest of SEO)
- #38 Average rank has gone down across most of the pages we’ve optimised (this isn’t a magic bullet, using NLP doesn’t guarantee results just like the rest of SEO)
- #39 Average rank has gone down across most of the pages we’ve optimised (this isn’t a magic bullet, using NLP doesn’t guarantee results just like the rest of SEO)
- #40 Average rank has gone down across most of the pages we’ve optimised (this isn’t a magic bullet, using NLP doesn’t guarantee results just like the rest of SEO)