The Datapred blog

On machine learning, time series and how to use them.

Cost function

Custom loss functions - What they are and why you need them

Machine learning models often disappoint in production for a simple reason - they have the wrong target.
Read More
Productizing models

Productizing machine learning models - What is required?

on Mar 19, 2019 12:37:43 PM By | Datapred | 0 Comments | machine learning continuous intelligence streaming data productization
We talk a lot with the data science teams of industrial companies, and it is striking how unaware they usually are of the gap between a good machine learning model and a production-ready machine learning application.
Read More
seasons

How should you handle seasonality?

on Jan 14, 2019 2:07:43 PM By | Datapred | 0 Comments | predictive analysis seasonality holt-winters
Seasonality - recurring but not necessarily periodic data patterns - is a staple of time series modeling. Since capturing true seasonality greatly enhances model accuracy, we wanted to share our thoughts and experience on the detection and modeling of such data patterns.  
Read More
Backtesting-1

Advanced cross-validation tips for time series

In a previous post, we explained the concept of cross-validation for time series, aka backtesting, and why proper backtests matter for time series modeling.
Read More
Fish & Chips

Best practices for bulletproof time series modeling

on Oct 30, 2018 9:03:51 AM By | Datapred | 0 Comments | time series machine learning backtesting backtest predictive analysis modeling
Our goal in this post is to discuss our standard strategy (beyond respecting basic time series modeling principles) for building accurate predictive models. We will use the example of commodity purchasing optimization.  
Read More
Prophet

A better (Facebook) Prophet

on Oct 9, 2018 2:30:17 PM By | Datapred | 0 Comments | machine learning demand prediction Prophet predictive analysis facebook
What is Facebook's Prophet? Prophet is a forecasting (i.e. time-series specific) algorithm open-sourced by Facebook in February 2017, and belonging to the GAM family of algorithms.
Read More
Forest

Random Forest for predictive maintenance: Try harder

on Sep 19, 2018 2:16:56 PM By | Datapred | 0 Comments | random forest predictive maintenance
(Based on too many true stories) You are the Operations Manager of an industrial company. You have read a lot about Industry 4.0, and asked your favorite consultancy to build a predictive maintenance prototype for your key industrial asset.
Read More
Demand prediction

Machine learning for demand prediction: what works and what doesn’t

on Aug 24, 2018 10:15:31 AM By | Datapred | 0 Comments | time series machine learning demand prediction
Machine learning for demand prediction is all the rage: industrial companies are suddenly waking up to the potential of machine learning in that area, proofs of concept are being launched everywhere, consulting companies are making millions…
Read More
Backtesting

The basics of backtesting

on Aug 14, 2018 5:03:00 PM By | Datapred | 0 Comments | time series machine learning backtesting backtest
Time series modeling requires the understanding of concepts that feel both basic and abstract:
Read More
neural_network

The problem with deep learning and time series

Industry is finally waking up to the potential of machine learning for predictive analysis. But data scientists faced with a prediction challenge (demand prediction, predictive maintenance, churn prediction…) usually favor one of two equally inadequate approaches:
Read More
All posts

Recent Posts

Topics

see all

Subscribe