Discover and read the best of Twitter Threads about #Stats

Most recents (24)

1/ "Software is eating the world. Machine learning is eating software. Transformers are eating machine learning."

Let's understand what these Transformers are all about

#DataScience #MachineLearning #DeepLearning #100DaysOfMLCode #Python #pythoncode #AI #DataAnalytics
2/ #Transformers architecture follows Encoder and Decoder structure.

The encoder receives input sequence and creates intermediate representation by applying embedding and attention mechanism.

#DataScience #MachineLearning #DeepLearning #100DaysOfMLCode #Python #pythoncode #AI
3/ Then, this intermediate representation or hidden state will pass through the decoder, and the decoder starts generating an output sequence.

#DataScience #MachineLearning #DeepLearning #100DaysOfMLCode #Python #pythoncode #AI #DataScientist #DataAnalytics #Statistics
Read 14 tweets
2/ It is important to standardize variables before running Cluster Analysis. It is because cluster analysis techniques depend on the concept of measuring the distance between the different observations we're trying to cluster.

#DataScience #MachineLearning #DeepLearning
3/ If a variable is measured at a higher scale than the other variables, then whatever measure we use will be overly influenced by that variable.

#DataScience #MachineLearning #DeepLearning #100DaysOfMLCode #Python #pythoncode #AI #DataScientist #DataAnalytics #Statistics
Read 16 tweets
1/ Can you classify something without seeing it before - that's what Zero-Shot Learning is all about - A Thread

👉 One of the popular methods for zero-shot learning is Natural Language Inference (NLI).

#DataScience #DeepLearning #MachineLearning #100DaysOfMLCode #Pytho
3/ In Zero-shot classification, we ask the model to classify a sentence to one of the classes (label) that the model hasn't seen during training.

#DataScience #MachineLearning #100DaysOfMLCode #DataScientist #Statistics #programming #ArtificialIntelligence #Data #pythoncode #AI
Read 13 tweets
1/ Why do we need the bias term in ML algorithms such as linear regression and neural networks ? - A thread

#DataScience #MachineLearning #100DaysOfMLCode #DataScientist #Statistics #programming #ArtificialIntelligence #Data #pythoncode #AI #Stats #DeepLearning #100DaysOfCode Image
2/ In linear regression, without the bias term your solution has to go through the origin. That is, when all of your features are zero, your predicted value would also have to be zero.

#DataScience #MachineLearning #100DaysOfMLCode #DataScientist #Statistics #programming Image
Read 7 tweets
1/ #MachineLearning #Interview questsion -
Why L1 regularizations causes parameter sparsity whereas L2 regularization does not?

#DataScience #MachineLearning #100DaysOfMLCode #DataScientist #Statistics #programming #ArtificialIntelligence #Data #pythoncode #AI #Stats
2/ L1 & L2 regularization add constraints to the optimization problem. The curve H0 is the hypothesis. The solution to this system is the set of points where the H0 meets the constraints.

#DataScience #MachineLearning #100DaysOfMLCode #DataScientist #Statistics #programming Image
3/ Regularizations in statistics or in the field of machine learning is used to include some extra information in order to solve a problem in a better way.

#DataScience #MachineLearning #100DaysOfMLCode #DataScientist #Statistics #programming #ArtificialIntelligence #Data
Read 7 tweets
Dalot vs Wan bissaka
Which is better in stats PER 90???

Let's start with passes we see that dalot and wan bissaka in passes per 90 they have many similarities,but the assist statistic and key passes is axtual very bad... Image
Then we continue and we see the defensive actions , now we see some differences in the number we can see that dalot in many things he is actually better than Wan Bissaka and he improved a lot these months in contrast wan bissaka that he is better only for blocks Image
Per 90 minutes we see that dalot helps a lit bit more in passes that lead for shot attempt or a goal , that means dalot better in offensive and defensive things than wan bissaka Image
Read 8 tweets
Z-Scores are a great #Stats tool to help give us some insight into a data set, especially in the #Bitcoin and On-Chain world.

So, what exactly is a Z-Score? What is it good for? What are its limitations? What kinds of Z-Scores are there?

Let's dive in.👇🧵
Before we can talk Z-Scores though, we need to talk Standard Deviation (SD).

What's that you say?

Well, a SD is, essentially, the "average distance from the average (mean)" for a set of data points. Image
A Z-Score, then, takes each individual point's distance from the mean and compares that to the SD.

So essentially saying, "how does this point's distance from the mean compare to the SD?". Image
Read 20 tweets
Petit thread sur l'usage des intervalles de confiance dans les sondages politiques. Ici en utilisant le dernier sondage en rolling de l'IFOP pour les présidentielles françaises 👇

#Presidentielle2022 #Sondage #stats Image
La présentation de ce sondage inclue une notice méthodologique avec un intervalle de confiance générique à utiliser. Celui-ci est exact, même s'il faut faire l'impasse sur la méthode de quota utilisée ici. Image
Mais il ne contient pas les intervalles de confiance pour chaque résultat. J'ai cherché à les ajouter à partir des données fournies leur notice méthodologique et les conclusions sont assez différentes. Première conclusion : ce n'est pas facile...
Read 24 tweets
🚨Job talk thread🚨

Title: What Can *Conformal Inference* Offer to Statistics?


Main points:
(1) Conformal Inference can be made applicable in many #stats problems
(2) There are lots of misconceptions about Conformal Inference
(3) Try it!

Conformal Inference was designed for generating prediction intervals with guaranteed coverage in standard #ML problems.

Nevertheless, it can be modified to be applicable in

✔️Causal inference
✔️Survival analysis
✔️Election night model
✔️Outlier detection
✔️Risk calibration

Misconceptions about conformal inference:

❌ Conformal intervals only have marginal coverage and tend to be wide
✔️ Conformal intervals w/ proper conformity scores achieve conditional coverage & efficiency (short length) if the model is correctly specified

Read 6 tweets
1/ Continuons notre série "où est passé l'argent de la recherche" avec les études du #SARS_CoV_2 dans les #eaux usées.

Sur le principe, pourquoi pas ?

C'est assez robuste pour y retrouver les #drogues. En France, ces travaux du @CNRS étaient dans @libe…
2/ Le soucis est qu'un #virus c'est pas une molécule.

La bonne nouvelle a été que l'on peut détecter le #SARS_CoV_2 dans les eaux usées !

La moins bonne a été qu'il était difficile de savoir quoi en faire...

Car savoir qu'un virus pandémique circule aide peu.
3/ Au niveau du #dépistage, à peu près aucune étude sur les #eaux usées n'a regardé la #sensibilité ou #spécificité de la méthode.

Quelle variabilité pour un même échantillon ?

Quelle variabilité pour des échantillons pris le même jour à une même station ?
Read 11 tweets
Geographic Data Science with R! 🚀🚀🚀

The R for geographic data science, by @maps4thought, provides an introduction to data science with applications for geographic data. 🧵👇

Images credit: from the book
#rstats #DataScientists #MachineLearning #Stats
The book follows the syllabus of the "R for Data Science" course at the School of Geography, @uniofleicester. The book is still a work in progress, and a draft version is available online.
The book covers to core concepts of data science such as:
✅ Intro to R programming
✅ Data wrangling
✨ Data manipulation
✨ Table operations
✨ Reproducibility
Read 8 tweets
Bayesian Generalized Linear Models with Julia! 🚀🚀🚀

TuringGLM is a new Julia package for GLM models with Bayesian flavor ❤️. As its name implies, the package uses the Turing package on the backend for the regression engine. 🧵 👇🏼

#JuliaLang #TuringLang #DataScience #Stats Image
It enables to specify Bayesian Generalized Linear Models using the formula syntax and returns an instantiated Turing model.
The package is inspired by the R's brms and Python's bambi packages (see links 👇🏼).

#rstats #Python
The package supports the following Bayesian GLM models:
✅ Linear regression
✅ Logistic regression
✅ Poisson regression
✅ Negative binomial regression
✅ Robust Regression
✅ Hierarchical models Image
Read 6 tweets
New book for Bayesian statistics with #Python! 📚📊🚀
The Bayesian Modeling and Computation in Python by
@aloctavodia, @canyon289, and @junpenglao provides an introduction to Bayesian statistics using core Python libraries for Bayesian 🧵 👇🏼

#bayesian #MachineLearning #stats ImageImageImage
The book covers the following four topics:
- Bayesian Inference concepts
- Bayesian regression methods for linear regressions, splines,
- Time series #forecasting
- Bayesian additive regression trees
- Approximate of Bayesian computation
The book's authors are the contributors of PyMC3, ArviZ, Bambi, #TensorFlow Probability, and other #Python libraries:
PyMC3 -
Tensorflow Probability -
ArviZ -
Read 6 tweets
This graph is now dynamically created based upon the latest @CDCofBC dashboard. #vaccineswork, even in the last month of this awful pandemic.

#stats #covidbc #bcpoli #epitwitter

/1 Image
Vaccination effectiveness dashboard

/2 Image
Raw data from @CDCofBC Image
Read 5 tweets
Die #Corona-#ICU-Erstaufnahmen zeigen, dass in den meisten Bundesländern nicht nur die Wachstumsrate sinkt, sondern mittlerweile auch die absoluten Zahlen:
#COVID19 #lockdown #ImpfenSchuetzt
Die Datengrundlage stammt aus dem #DIVI-#Intensivregister. Sie beinhaltet alle Personen, die ERSTMALIG auf einer #Intensivstation aufgenommen werden und SARS-CoV-2-positiv getestet sind sowie jene, die wegen Nachwirkungen einer #COVID19-Erkrankung eingeliefert wurden …
… dies schließt ausdrücklich auch Personen mit ein, die "nur" einen positiven Test haben, aber nicht an C19 erkrankt sind, worauf DIVI explizit hinweist ( Dies betrifft aber die große Mehrheit der Fälle nicht.
Read 8 tweets
@ATMNE_math panel discussions happening now! Being recorded for those who can't attend live.
Thanks @atomic_math!
Looking forward to panel on "Math 4 All" with @carloliwitter @mslailanur & Michelle McKnight
#MTBoS #iTeachMath
Michelle: Math 4 All means all Ss see themselves as math do-ers & thinkers & communicators. No obstacles are being set up by us, so all Ss see themselves as "math people"
@mslailanur : Meet all Ss where they are
@mslailanur @carloliwitter Everyone should have the access to all the math, so everyone can have the opportunities & options.
Reminded of Bob Moses, radical equations.
Read 24 tweets
hello #psychology #stats twitter!

a while ago i promised some graphs on why #removing #outliers using a simple cut-off (eg, >2SDs) is a #bad idea

so that I can sleep at night again, here they are


for some reason, it's very common in #psychology to remove 'outliers' from data

most common way: exclude data more than two standard deviations from mean

we spend time & money collecting data, then throw 5% away


I don't know why, or where it's taught, but there it is

Read 30 tweets
#ichimoku Pour faire suite à , le lien entre ce modèle 2A-A et pourquoi #Hosoda aurait construit #kinkohyo et #kumo de façon similaire, avec l'explication de la projection en 26 du Kumo.
(en attendant les vidéos d'@kieffer_herve 😉)
Formalisme 2A-A : spans, Kinko Hyo et Kumo, pourquoi #Hosoda n'a pas fait les choses au hasard avec son #Ichimoku
Pour compléter l'explication des spans, du #KinkoHyo et du #Kumo dans le formalisme 2A-A ( et ) avec en plus une explication possible du retracement statistique à 62.5% environ 😃
#Hosoda #Ichimoku Image
Read 40 tweets
1/ A thread on @pendle_fi, one of the most understimated projects that recently launched on mainnet. Pendle is a protocol that enables the trading and hedging of yield and currently supports two yield-generating tokens: aUSDC and cDAI, both products expire on December 29th, 2022.
2/ A deposit of aUSDC is separated into OT-aUSDC (OT represents the ownership of the underlying token deposited) and YT-aUSDC (representing the ownership of the future yield generated).
3/ @pendle_fi created a USDC liquidity pool for YT-aUSDC and YT-cDAI to price their markets, this is achieved providing incentives,
Read 14 tweets
The latest figures about the 10000s of #britsinFrance #latvia #luxembourg #malta who haven't applied for residence makes alarming reading. With a #harddeadline of #30June, my latest piece for @ukandeu examines the #stats 1/…
It addresses who might fall between the gaps - building in my experience in working with British citizens living in France over the course of the #brexitnegotiations 2/…
And calls for these countries to extend their deadlines for applications and develop communication strategies to target hard-to-reach populations. And I'm not alone @BritishInEurope have released a statement to the same effect 3/⬇️⬇️⬇️…
Read 6 tweets
There is a lot of nuance people are missing in this.

One - @Tesla and Elon still believe heavily in #bitcoin - why? They are Hodling.

Two - this signals they didn’t have a high take up on #tesla purchases with #btc - why ? 1/
Because like @elonmusk and @Tesla most people who have enough to buy a Tesla is #hodling.

Most ppl are actually selling their cars to acquire more #sats.

Three- Uninformed- #GreenTerrorists have been on a massive anti-Tesla campaign since @Tesla purchased #bitcoin 2/
This would have had a negative effect on traditional #model3 and #modelx car sales - as this green army was the backbone of positive pr for Tesla.

With #btc sales not making up the difference in sales - they need to do something to #greenwash their image. 3/
Read 6 tweets
Toute bonne bibliothèque de statistique doit avoir le livre « A Million Random Digits », 1 million de chiffres aléatoires réunis dans 400 pages. Parce qu’en 1955, c’est très très difficile de générer des chiffres vraiment aléatoires en grandes quantités. #stats #random Page de titre : A million random digitspage du livre A million random digits contenant uniquement d
Les chiffres aléatoires sont utilisés dans plusieurs domaines : la statistique mais aussi dans le développement de jeux vidéos ou si un ingénieur doit choisir au hasard les pièces métalliques à vérifier sur un pont plutôt que de toutes les vérifier par exemple...🧐
Alors plutôt que demander à son voisin un chiffre entre 0 et 1 000 000 qui ne serait pas vraiment aléatoire (puisque tout le monde a des schémas pré-établis qui pourraient rendre ce chiffre prédictible), la Rand Corporation a édité ce livre.
Read 10 tweets
Hon. Minister of State for Petroleum Resources, @HETimipreSylva will brief the press on the imminent passage of the #PetroleumIndustryBill (PIB), and what it holds for the country. Noon today. #StateHouseBriefing

Watch Live on State House YouTube Page:
Apologies for the delay in kicking off. The briefing is now on:

#StateHouseBriefing #PetroleumIndustryBill
Present at the briefing: HMS @FMPRng, @HETimipreSylva; GMD @NNPCgroup, @MKKyari; and Executive Secretary of the Petroleum Technology Development Fund (PTDF), Dr. Bello Aliyu Gusau #StateHouseBriefing
Read 19 tweets

Related hashtags

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!