Name: SPbR #17. International
Start: 2020-12-11T18:00:00+03:00
End: 2020-12-11T20:00:00+03:00

Вы этот раз мы соберемся онлайн вместе с нашими друзьями из Applied R Munich.

С организацией трансляции нам в очередной раз помогает JetBrains. Спасибо им за это.

Поговорим про использовании R в лингвистике, Fuzzy Matching и машинном обучении в R. Все доклады будут на английском.

Анонсы докладов:

Speaker 1:
Anton Antonov, dxFeed Solutions GmbH,
R in linguistic studies
I have always known that linguistics and mathematics have a lot in common. Still, I was surprised to discover the true extent of that similarity. The complexity, the aesthetics, seemingly impossible problems that are resolved by establishing intrinsic relations between abstract objects... Finally, the helpfulness of R in dealing with these challenges.
In this obviously non-serious talk, I will be speaking about my amateurish attempts at passing a particularly complex online course in linguistics using only my head, a pen, a sheet of paper, and R.

Speaker 2:
Alexey Osipov, Sidenis,
Linking via Fuzzy Matching
Quite often we have to link datasets by text fields, in which the same information is represented in different ways.
We will consider several test cases for the problem of linking various datasets:

linking of datasets by text fields with meaningful information,
linking of datasets with short text fields with non-meaningful information (e.g., company names),
linking of a dataset with unstructured text with a dataset with non-meaningful information (e.g., company names).
We will discuss the theory, look into the relevant R packages, and outline the applications in insurance.
Speaker 3:
Martin Binder, creator of the mlr3piplines package,
mlr3: Machine Learning, Pipelines, and AutoML in R
One of R's strong points for data science is the availability of many different machine learning algorithms and implementations. This diversity, however, comes at the cost of inconsistent APIs, which often require writing custom glue code for everyday tasks such as resampling or data preprocessing. 'mlr3' and accompanying packages remedy this by offering a uniform interface to many machine learning methods, while at the same time providing a broad range of further functionality.
We will present the possibilities offered by 'mlr3': From basics like model fitting and inference, over model performance evaluation, benchmarking, and hyperparameter tuning, to the capability to build complex data pre- and post-processing pipelines with simple building blocks and a natural data flow paradigm. Finally, we will provide links to further resources on how mlr3 can be used most effectively.

Arkady Kalakutsky

St. Petersburg R User Group

Technology

Cloud Computing

Statistical Computing

Data Analytics

Big Data

Machine Learning

Data Visualization

Data Mining

R Project for Statistical Computing

Data Science

Programming in R

Applied Statistics

SPbR #17. International

Online event

St. Petersburg R User Group

SPbR #17. International

St. Petersburg R User Group

Детали

Участники также заинтересованы в