Introduction to Forecasting in Machine Learning and Deep Learning




[ad_1]

QCon Plus covers the trends, best practices, and solutions leveraged by the world’s most innovative software shops. Taking place between November 4-18, the event is thoughtfully designed with shorter, focused technical sessions spread over 3 weeks. You’ll learn from 54 speakers and 4 keynotes across 18 tracks. The event includes highly interactive sessions, Q&As, AMAs, breakouts, and real-time collaborative action.

Find out more about QCon Plus (Nov 4-18, 2020) and save your spot now: https://bit.ly/3kMT5bo

————————————————————————————————————————————————

Forecasts are critical in many fields, including finance, manufacturing, and meteorology. At Uber, probabilistic time series forecasting is essential for marketplace optimization, accurate hardware capacity predictions, marketing spend allocations, and real-time system outage detection across millions of metrics.

In this talk, Franziska Bell provides an overview of classical, machine learning and deep learning forecasting approaches. In addition fundamental forecasting best practices will be covered.

This video was recorded at QCon.ai 2018: https://bit.ly/2piRtLl

If you are a software engineer that wants to learn more about machine learning check our dedicated introductory guide https://bit.ly/2HPyuzY .

For more awesome presentations on innovator and early adopter, topics check InfoQ’s selection of talks from conferences worldwide http://bit.ly/2tm9loz

Source


[ad_2]

Comment List

  • InfoQ
    December 10, 2020

    What are the limitations of classical time series according to the video?
    What forecasting challenges was Uber facing in 2018 ?

  • InfoQ
    December 10, 2020

    Short but full of learning ideas for me! Appreciate it.

  • InfoQ
    December 10, 2020

    Gives you wellness, Professor ♥
    Please help .. I need data set .. to train Model to predict the occurrence of fires due to the large number of fires that happen today in Syria (more than 100 fires)
    Please Help

  • InfoQ
    December 10, 2020

    This video was a complete waste of time. Nothing of value was imparted.

  • InfoQ
    December 10, 2020

    Hello Sir,

    I have a question about forecasting. Should plan corrections be determined during forecasting? Or is it not necessary to determine them? Why is there an extrapolation? An extrapolation is absolutely necessary.

  • InfoQ
    December 10, 2020

    just a commercial presentation. Lost of time.

  • InfoQ
    December 10, 2020

    Excellent delivery

  • InfoQ
    December 10, 2020

    Beautifully explained 👍😊

  • InfoQ
    December 10, 2020

    There are some time-travelling mistakes shown here. "passes" are being run https://youtu.be/bn8rVBuIcFg?t=315 – that's the mistake. Pass 2 has been affected by pass 1 (double bad there's even overlap in the test). In other words, pass-2 has the benefit of knowledge from the future. Here is where the mistake manifests: https://youtu.be/bn8rVBuIcFg?t=619 – that is plotting the equivalent of "pass 5", which has had multiple generations of insight into the future. If an entirely new and never-before-seen test set had been chosen, and had been run and shown FOR THE FIRST TIME EVER during this talk, then the output is meaningful. Right now, it's just fools-gold. The "FIRST TIME EVER" statement is very important. If it was run before walking on stage, and it didn't work, it would not get shown to us… yet ANOTHER time-traveller mistake. This talk is not useful to say whether or not it's superior to the other methods shown for solving the problem (two reasons: the time-traveller mistake in the ML, and the lack of giving the other models similar insight, or at least an even playing field). It also lacks comparison against any non-ML bespoke solution built by a statistician. If it's doing worse than a human-built design made by a team spending the same number of hours on the problem, then it's a step backwards, right?

  • InfoQ
    December 10, 2020

    Nice presentation!

  • InfoQ
    December 10, 2020

    Did anyone have success in finding the open source book by Rob Heinemann? Thanks

  • InfoQ
    December 10, 2020

    The machine learning holds the highest CAGR of 44.86% during the forecast period 2019-2025.

    Request a sample @ https://www.envisioninteligence.com/industry-report/global-machine-learning-market/?utm_source=yt-chitti

  • InfoQ
    December 10, 2020

    Can you use this model in predicting stock prices?

  • InfoQ
    December 10, 2020

    The machine learning holds the highest CAGR of 44.86% during the forecast period 2019-2025.

    Request a sample @ https://www.envisioninteligence.com/industry-report/global-machine-learning-market/?utm_source=yt-chitti

  • InfoQ
    December 10, 2020

    Here is also the document referenced in one of the slides which provide a lot of detail behind their architecture. https://arxiv.org/pdf/1709.01907.pdf

  • InfoQ
    December 10, 2020

    Waste of time, there are other videos better than this. No substance.

  • InfoQ
    December 10, 2020

    What was the book the lecturer recommanded at last, plz? I found it hard to figure out the author's name…

Write a comment