beyond precision and recall
play

Beyond Precision and Recall Considerations for better search - PowerPoint PPT Presentation

Sep 15, 2022 •1.35k likes •1.9k views

Beyond Precision and Recall Considerations for better search experience Andreas Brckner (Sr. Search and Merchandising Consultant) ATTRAQT provides cloud-based and server-to-server SaaS tools that maximize Onsite Search, Online

  1. Beyond Precision and Recall Considerations for better search experience Andreas Brückner (Sr. Search and Merchandising Consultant)

  2. ATTRAQT provides cloud-based and server-to-server SaaS tools that maximize Onsite Search, Online Merchandising and eCommerce Personalization for eCommerce retailers. With easy-to-use interfaces and effective technology, ATTRAQT helps retailers make their online shop as attractive and successful as their brick- and-mortar stores. ATTRAQT offers two distinct products, Freestyle Merchandising and Fredhopper, that enhance and extend retail platform capabilities and put the power and control back in the hands of the merchandising teams. ATTRAQT are able to work with all eCommerce platforms and in house systems including …

  3. Some of the industry’s leading retailers are already harnessing the power of ATTRAQT’s services, including: ASOS, House of Fraser, The North Face, Boulanger, Intergamma, Eddie Bauer, Timberland, Volcom, Waitrose and many more … For more information please contact sayhello@attraqt.com or visit www.attraqt.com

  4. Relevance? 96% of relevant 92% of results results are are relevant returned 100% 80% 60% Score 40% 20% 0% Precision Recall

  5. How to Change the Perception of Search?

  6. Search as conversation Browsing The need for Expressing a User search search question Expectations results

  7. Why we Search

  8. Reasons why we use search: Why we search

  9. Reasons why we use search: • The user knows what they are looking Why we search for

  10. Why we search Searches for specific products per industry 53% 3% Fashion Electronics

  11. Reasons why we use search: • The user knows what they are looking Why we search for • Search is more visible than navigation

  12. Why we search

  13. Why we search Size of Search Box

  14. Why we search Size of Search Box Search Use +75%

  15. Reasons why we use search: • The user knows what they are looking Why we search for • Search is more visible than navigation • Search is easier to use

  16. Expressing a Search Question

  17. Expressing a question

  18. Expressing a question 55% 32% 13% Additional Category or brand Combination filters

  19. Expressing a question

  20. Expressing a question 50% Share Fuzzy Matches Share of all searches 25% decrease in fuzzy matches 25% 0% -2 days -1 days 0 days 1 days 2 days Search suggest go-live

  21. Expectations about results

  22. Reasons why expectations are not met: Expectations?

  23. Reasons why expectations are not met: • Product not sold Expectations?

  24. Expectations: not sold

  25. Expectations: not sold

  26. Expectations: not sold

  27. Expectations: not sold

  28. Reasons why expectations are not met: • Product not sold Expectations? • Fuzzy matching

  29. Expectations: fuzzy search

  30. Expectations: fuzzy search

  31. Expectations: fuzzy search

  32. Expectations: fuzzy search

  33. Expectations: fuzzy search

  34. Reasons why expectations are not met: • Product not sold Expectations? • Fuzzy matching • Too generic search

  35. Expectations: generic

  36. Expectations: generic

  37. Expectations: generic

  38. Simplifying browsing products

  39. Browsing results Page 1 Page 2 Page 3 Page 4

  40. Browsing results

  41. Browsing results

  42. Browsing results

  43. Browsing results

  44. Browsing results dress Name Name Summer dress Evening dress Category Category Casual dresses , Evening wear Beach dresses Style Style Formal Skater dress

  45. Browsing results

  46. Browsing results

  47. Browsing results

  48. Browsing results

  49. Search as conversation Browsing The need for Expressing a User search search question Expectations results

  50. Thank you

Recommend

Precision and recall John Goldsmith June 26, 2015 1 Document retrieval Precision How well do

Precision and recall John Goldsmith June 26, 2015 1 Document retrieval Precision How well do

Precision and recall John Goldsmith June 26, 2015 1 Document retrieval Precision How well do the documents that your system gives you actually satisfy what you are looking for? Recall How sure are you that you got back all of the documents

594 views • 11 slides
Evaluation Metrics Presented by Dawn Lawrie 1 Some Possibilities Precision Recall F-measure

Evaluation Metrics Presented by Dawn Lawrie 1 Some Possibilities Precision Recall F-measure

Evaluation Metrics Presented by Dawn Lawrie 1 Some Possibilities Precision Recall F-measure Mean Average Precision Mean Reciprocal Rank 2 Precision Set Proportion of things of interest in some set Example: Im interested in apples

555 views • 28 slides
Beyond Precision and Recall: Understanding Uses (and Misuses) of Similarity Hashes in Binary

Beyond Precision and Recall: Understanding Uses (and Misuses) of Similarity Hashes in Binary

Beyond Precision and Recall: Understanding Uses (and Misuses) of Similarity Hashes in Binary Analysis Fabio Pagani 1 , Matteo DellAmico 2 , Davide Balzarotti 1 1 EURECOM 2 Symantec Research Labs ACM Conference on Data and Application Security

509 views • 36 slides
Precision, Recall, and Sensitivity of Monitoring Partially Synchronous Distributed Systems

Precision, Recall, and Sensitivity of Monitoring Partially Synchronous Distributed Systems

RunHme VerificaHon 2016 (RV16) Precision, Recall, and Sensitivity of Monitoring Partially Synchronous Distributed Systems Sorrachai Yingchareonthawornchai 1 , Duong Nguyen 1 , Vidhya Tekken Valapil 1 , Sandeep Kulkarni 1 , and Murat Demirbas 2

1.01k views • 55 slides
Data Mining Lecture 05: Overfitting Evaluation: accuracy, precision, recall, ROC

Data Mining Lecture 05: Overfitting Evaluation: accuracy, precision, recall, ROC

CISC 4631 Data Mining Lecture 05: Overfitting Evaluation: accuracy, precision, recall, ROC Theses slides are based on the slides by Tan, Steinbach and Kumar (textbook authors) Eamonn Koegh (UC Riverside) Raymond Mooney

748 views • 62 slides
and Evaluation CMSC 678 UMBC Central Question: How Well Are We Doing? Precision, Recall,

and Evaluation CMSC 678 UMBC Central Question: How Well Are We Doing? Precision, Recall,

Experimental Setup, Multi-class vs. Multi-label classification, and Evaluation CMSC 678 UMBC Central Question: How Well Are We Doing? Precision, Recall, F1 Accuracy Log-loss Classification ROC-AUC

982 views • 74 slides
Panel: Context-Dependent Evaluation of Tools for NL RE Tasks: Recall vs. Precision, and Beyond

Panel: Context-Dependent Evaluation of Tools for NL RE Tasks: Recall vs. Precision, and Beyond

Panel: Context-Dependent Evaluation of Tools for NL RE Tasks: Recall vs. Precision, and Beyond Daniel Berry, Jane Cleland-Huang, Alessio Ferrari, Walid Maalej, John Mylopoulos, Didar Zowghi 2017 Daniel M. Berry RE 2017 R vs P Panel Pg.

563 views • 53 slides
Panel: Context-Dependent Evaluation of Tools for NL RE Tasks: Recall vs. Precision, and Beyond

Panel: Context-Dependent Evaluation of Tools for NL RE Tasks: Recall vs. Precision, and Beyond

Panel: Context-Dependent Evaluation of Tools for NL RE Tasks: Recall vs. Precision, and Beyond Daniel Berry, Jane Cleland-Huang, Alessio Ferrari, Walid Maalej, John Mylopoulos, Didar Zowghi 2017 Daniel M. Berry RE 2017 R vs P Panel Pg.

1.02k views • 77 slides
VLVK EHF. VLVK EHF. Precision machining Precision machining Professional precision for

VLVK EHF. VLVK EHF. Precision machining Precision machining Professional precision for

VLVK EHF. VLVK EHF. Precision machining Precision machining Professional precision for 26 years From the start Vlvk was established in 1988 By Danel Gumundsson One person working for the first year (DG) Magns

564 views • 9 slides
Evaluation of Classifiers Evaluation of Classifiers ROC Curves ROC Curves Reject Curves Reject

Evaluation of Classifiers Evaluation of Classifiers ROC Curves ROC Curves Reject Curves Reject

Evaluation of Classifiers Evaluation of Classifiers ROC Curves ROC Curves Reject Curves Reject Curves Precision- -Recall Curves Recall Curves Precision Statistical Tests Statistical Tests Estimating the error rate of a classifier

565 views • 32 slides
GPU tuning, part 1 (updated) CSE 6230: HPC Tools & Apps Fall 2014 September 30 &

GPU tuning, part 1 (updated) CSE 6230: HPC Tools & Apps Fall 2014 September 30 &

vuduc.org/cse6230 GPU tuning, part 1 (updated) CSE 6230: HPC Tools & Apps Fall 2014 September 30 & October 2 Recall: 2 Recall: 6 GB/s 2 Recall: 3 Recall: 4 Recall: 5 Recall: 6 Recall: 7 Recall: 8 Recall: 9

1.19k views • 99 slides
Mixed Precision Training PAI Overview What is mixed-precision

Mixed Precision Training PAI Overview What is mixed-precision

Mixed Precision Training PAI Overview What is mixed-precision & Why mixed-precision How mixed-precision Mixed-precision tools on PAI-tensorflow Experimental results 1 What is mixed-precision

1.38k views • 45 slides
AM P A R CudA Multiple Precision ARithmetic librarY When do we need more precision?

AM P A R CudA Multiple Precision ARithmetic librarY When do we need more precision?

CAMPARY CudA Multiple Precision ARithmetic librarY Valentina Popescu Joined work with: Mioara Joldes, Jean-Michel Muller March 2015 AM P A R CudA Multiple Precision ARithmetic librarY When do we need more precision? Youtube view

912 views • 42 slides
INF 141 IR METRICS LATENT SEMANTIC ANALYSIS AND INDEXING Crista Lopes Outline Precision

INF 141 IR METRICS LATENT SEMANTIC ANALYSIS AND INDEXING Crista Lopes Outline Precision

INF 141 IR METRICS LATENT SEMANTIC ANALYSIS AND INDEXING Crista Lopes Outline Precision and Recall The problem with indexing so far Intuition for solving it Overview of the solution The Math How to measure Given the

648 views • 48 slides
MIXED PRECISION TRAINING Michael OConnor MIXED PRECISION What is the benefit? Using mixed

MIXED PRECISION TRAINING Michael OConnor MIXED PRECISION What is the benefit? Using mixed

MIXED PRECISION TRAINING Michael OConnor MIXED PRECISION What is the benefit? Using mixed precision and Volta your networks can be: 1. 3-4x faster 2. Reduce memory consumption and bandwidth pressure 3. just as powerful with no architecture

541 views • 11 slides
MIXED PRECISION TRAINING: THEORY AND PRACTICE Paulius Micikevicius What is Mixed Precision

MIXED PRECISION TRAINING: THEORY AND PRACTICE Paulius Micikevicius What is Mixed Precision

MIXED PRECISION TRAINING: THEORY AND PRACTICE Paulius Micikevicius What is Mixed Precision Training? Reduced precision tensor math with FP32 accumulation, FP16 storage Successfully used to train a variety of: Well known public

1.61k views • 37 slides
DYNAMIC PRECISION NUMERICS USING A VARIABLE-PRECISION UNUM TYPE I HW COPROCESSOR ARITH26 |

DYNAMIC PRECISION NUMERICS USING A VARIABLE-PRECISION UNUM TYPE I HW COPROCESSOR ARITH26 |

DYNAMIC PRECISION NUMERICS USING A VARIABLE-PRECISION UNUM TYPE I HW COPROCESSOR ARITH26 | BOCCO Andrea | 11 June 2019 INTRODUCTION: STATE OF THE ART Variable Precision (VP) computing has been investigated to improve convergence of

476 views • 20 slides
2018 Milken Institute Hamptons Dialogues Precision, Precision, Precision: The Future of Health

2018 Milken Institute Hamptons Dialogues Precision, Precision, Precision: The Future of Health

2018 Milken Institute Hamptons Dialogues Precision, Precision, Precision: The Future of Health and Medicine 1 1958: Americas Response to Sputnik 1958: Americas Response to Sputnik U.S. Defense budget increased to a peace- time record

4.49k views • 14 slides
Wearables for Precision Health Kay Connelly IU Grand Challenge Precision Health Initiative The

Wearables for Precision Health Kay Connelly IU Grand Challenge Precision Health Initiative The

Wearables for Precision Health Kay Connelly IU Grand Challenge Precision Health Initiative The goal of the IU Precision Health Initiative is to position Indiana University among the leading universities in discovering and developing better

237 views • 19 slides
Evaluating classifiers CS440 The 2-by-2 contingency table correct not correct positive tp fp

Evaluating classifiers CS440 The 2-by-2 contingency table correct not correct positive tp fp

10/30/19 Evaluating classifiers CS440 The 2-by-2 contingency table correct not correct positive tp fp prediction negative fn tn prediction 1 10/30/19 Precision and recall Precision : % of predicted items that are correct P = tp/

117 views • 11 slides
The Intersection between Precision Medicine and Implementation Science Precision Medicine and

The Intersection between Precision Medicine and Implementation Science Precision Medicine and

The Intersection between Precision Medicine and Implementation Science Precision Medicine and Population Health (PMPH) Interest Group www.cancercontrol.cancer.gov/ Session Logistics Questions are encouraged. You may submit your questions

403 views • 25 slides
EFFECTIVE USE OF MIXED PRECISION FOR HPC Kate Clark, Smoky Mountain Conference 2019 Why Mixed

EFFECTIVE USE OF MIXED PRECISION FOR HPC Kate Clark, Smoky Mountain Conference 2019 Why Mixed

EFFECTIVE USE OF MIXED PRECISION FOR HPC Kate Clark, Smoky Mountain Conference 2019 Why Mixed Precision Lattice Quantum Chromodynamics Mixed Precision and Krylov Solvers AGENDA Mixed Precision and Multigrid Tensor cores Future Challenges

945 views • 47 slides
Statistical Challenges Towards a Semantic Model for Precision Agriculture and Precision Livestock

Statistical Challenges Towards a Semantic Model for Precision Agriculture and Precision Livestock

www.cybele-project.eu Statistical Challenges Towards a Semantic Model for Precision Agriculture and Precision Livestock Farming Dimitris Zeginis , Evangelos Kalampokis, Konstantinos Tarabanis SemStats 2019 This project has received funding from

169 views • 14 slides
1 Bin(10, 0.3), Bin(100, 0.03) vs. Poi(3) Tender (Central) Moments with Poisson Recall: Y ~

1 Bin(10, 0.3), Bin(100, 0.03) vs. Poi(3) Tender (Central) Moments with Poisson Recall: Y ~

Whither the Binomial Binomial in the Limit Recall example of sending bit string over network Recall the Binomial distribution n ! n = 4 bits sent over network where each bit had i n i ( ) ( 1 ) P X i

191 views • 4 slides

More recommend