Manifold Blog

Manifold Blog

Preparing Your Data for Predictive Analytics

Posted by Sourav Dey on Nov 16, 2018 2:17:23 PM

By Kyle Seaman, Head of Product at Sentenai, and Sourav Dey, Co-Founder and CTO at Manifold

Predictive analytics is an undeniably valuable technology, with research indicating its market size could top $12 billion USD by 2022. Across a range of industries, businesses, and applications, using historical data to predict future outcomes can lead to greater operational efficiency in a variety of ways. Predictive analytics can enable organizations to streamline their operational processes, optimize their demand forecasting, drastically reduce downtime, and better understand their customers’ propensity to buy.

Read More

Topics: Data science

Intimidated by AI? Ask Yourself These 5 Questions, And You’re Halfway to Implementation

Posted by Vinay Seth Mohta on Oct 4, 2018 1:11:28 PM

Do you ever feel like machine learning is moving so fast that it’s impossible to keep up? You’re not alone — that’s what the hype cycle has lots of people thinking.

Hype bubbles seem to build up every few years around a specific technology, like the cloud, big data, or, in this case, artificial intelligence.

Read More

Topics: Data science

Lean AI: A 6-Step Guide to Making a Tangible Business Impact as Efficiently as Possible

Posted by Vinay Seth Mohta on Oct 1, 2018 3:10:34 PM

Lean AI is a new, innovative practice and its principles should be widely recognizable. A number of existing systems inspired us in the development of Lean AI, including human-centered design at IDEO, the Lean Startup methodology, agile software development principles, and the CRISP-DM approach pioneered by the data-mining community.

Read More

Topics: Data science

Custom Loss Functions for Gradient Boosting

Posted by Prince Grover on Sep 28, 2018 3:27:51 PM

By Prince Grover and Sourav Dey

 

Gradient boosting is widely used in industry and has won many Kaggle competitions. The internet already has many good explanations of gradient boosting (we've even shared some selected links in the references), but we've noticed a lack of information about custom loss functions: the why, when, and how. This post is our attempt to summarize the importance of custom loss functions in many real-world problems — and how to implement them with the LightGBM gradient boosting package.

Read More

Topics: Data science, Data engineering

Applications of Matrix Decompositions for Machine Learning

Posted by Prince Grover on Jul 25, 2018 9:00:00 AM

In machine learning and statistics, we often have to deal with structural data, which is generally represented as a table of rows and columns, or a matrix. A lot of problems in machine learning can be solved using matrix algebra and vector calculus. In this blog, I’m going to discuss a few problems that can be solved using matrix decomposition techniques. I’m also going to talk about which particular decomposition techniques have been shown to work better for a number of ML problems.

Read More

Topics: Data science

Distance Matrix Vectorization Trick

Posted by Sourav Dey on Aug 15, 2016 7:00:00 AM

A common problem that comes up in machine learning is needing to find the l2-distance between two sets of vectors. For example, in implementing the k-nearest-neighbors algorithm, we have to find the l2-distance between the a set of test vectors, held in a matrix X (MxD), and a set of training vectors, held in a matrix X_train (NxD). Our goal is to create a distance matrix D (MxN) that contains the l2-distance from every test vector to every training vector. How can we do this efficiently?

Read More

Topics: Data science

Sensor Edge Finding at Cortex

Posted by Sourav Dey on Feb 22, 2016 9:00:00 AM

This article is Part 2 of a three-part series that we are writing about work that Manifold did with one of our clients, Cortex Building Intelligence. In our previous post, we talked about finding edges in sensor signals so we could use them to help us estimate a building’s start time.

We want to find sharp transitions in the various HVAC sensors—rising edges for electricity, steam, and static pressure and falling edges for supply air temperature (SAT). It’s easy for a human to pick out edges—but how do we teach a computer to do it?

Read More

Topics: Data science, Signal processing, AI at the edge

Data Science at Cortex

Posted by Sourav Dey on Feb 17, 2016 9:00:00 AM

This article is Part 1 of a three-part series that we are writing about work Manifold did with one of our clients, Cortex Building Intelligence. Cortex’s vision is to use data-science to make commercial building heating, ventilation, and air conditioning (HVAC) operations more efficient.

Over the next few posts, we want to give you a look "under the hood” of our data-science operations. To that end, we’ll discuss how we solved one of the foundational problems at Cortex: figuring out when a building’s HVAC systems were turned on.

Read More

Topics: Signal processing, Data science, AI at the edge

Never Miss a Post

Get the Manifold Blog in Your Inbox

We publish occasional blog posts about our client work, open source projects, and conference experiences. We focus on industry insights and practical takeaways to help you accelerate your data roadmap and create business value.


Subscribe Here


Recent Posts