Ask your data
  • Blog
  • About & Contact

What fundraisers can learn from machines

2/19/2017

0 Comments

 
BildDaintree, Australia
Artificial intelligence (AI) has been an influential concept not only in the scientific community but also in popular culture. Depending on the respective attitude towards AI, one might associate characters such as inhibited but likeable android Data from Star Trek or neurotic and vicious HAL 9000 from the Space Odysee movies with it. The form of AI those two represent is called General AI, i.e. the intelligence of a machine that enables it to perform intellectual tasks as good as or better than a human. General AI is – at least for the moment – still a topic for science fiction. What is interesting for various industries is the notion of Narrow AI (also termed Weak AI). These are technologies that enable humans to fulfil specific tasks in an automated manner just like or even better than they could.

In the context of data, machine learning can be seen as an approach to achieve artifical intelligence. Machine learning is about analyzing data, learning from it and using the insights gained for decision making or predicitions about something. 

The learning in machine learning can be attempted in two generic forms:
  • The first is so called supervised learning. Supervised learning algorithms aim to make predictions based upon a given set of data examples for which properties are known. A machine learning algorithm can for instance look for patterns in historic stock prices and take any possibly relevant information into account that is available with it, be it days of week, indicators that describe company performance, the state of the economy or even the weather. As soon as an algorithm has found the relationship between the stock price (dependent variable) and independet variables (the ones that allegedly influence stock prices), it can be used to predict future outcomes by plugging in values for the variables. This might ring a bell for many who have come across regression models in their education. In fact linear regression is said to be the most popular machine learning model.
  • In unsupervised learning data comes without pre-defined labels, i.e. any kind of classifications or categorizations are not included. The goal of unsupervised learning algorithms is to discover hidden structures. A practical example of this is the functionality of Apple’s iOS that tries to sort photos using the actual faces on them. Unsupervised learning seems much harder as we ask the algorithm to do something for us without telling it (or knowing) how to do it.
Going through all possibly relevant algorithms would go beyond the scope of this post that is intended to be an introductory one (more specific ones are planned, though). As I have started diving deeper into machine learning recently, I can recommend this cheat sheet by Mithun Sridharan before you start your desk research and do the googeling. You might also find Laura Hamilton's overview of the pros and cons of the most popular algorithms helpful for a start. If you can´t wait to start trying things, you might find this blog post on KD Nuggest worth taking a look at as it offers heaps of links to machine learning cheat sheets for different platforms. As a fan of R, I can particularly recommend this overview by Yanchang Zhao.

What can fundraisers do? I have to say that I did not come across many practice sharing posts or articles when I conducted research for this post. I doubt that either the availability of data or the competence portfolio of analysts and data scientists limit the possibilities of fundraising organizations in the context of machine learning techniques. However, there might be a certain level of insecurity regarding where and how to start. I found an inspiring blog post by Stephen W. Lambert in which he explains that all you basically need is a computer, a database with relevant data and your brain to start diving into machine learning techniques. I think Lambert’s text invites fundraising organizations to do their theoretical and conceptual homework, process and prepare their data accordingly and start experimenting with maching learning techniques. So - go ahead and try.





0 Comments



Leave a Reply.

    Categories

    Alle
    Artificial Intelligence
    Because It´s Fun!
    Churn
    Clustering
    Data Sciene @NPO
    Data Visualization
    Facebook
    Machine Learning
    Maps
    Natural Language Processing
    Neural Nets
    Power BI
    Predictive Analytics
    Social Media
    Time Series
    Twitter

    Archive

    Januar 2021
    November 2020
    September 2020
    August 2020
    Mai 2020
    April 2020
    Februar 2020
    Dezember 2019
    November 2019
    September 2019
    Juni 2019
    April 2019
    März 2019
    Januar 2019
    Dezember 2018
    Oktober 2018
    August 2018
    Juni 2018
    Mai 2018
    März 2018
    Februar 2018
    Dezember 2017
    November 2017
    Oktober 2017
    September 2017
    August 2017
    Juli 2017
    Mai 2017
    April 2017
    März 2017
    Februar 2017
    Januar 2017

About

Copyright © 2018
  • Blog
  • About & Contact