Detecting spam users automatically with a neural network
Future
The method described in this article has some limitations. Although a neural network can come close to any complex function, it may be the case that the optimization processes do not produce the optimum solution. In this case, the network only achieves a low accuracy level.
A further potential problem is caused by unbalanced or contradictory training data, which, for instance, might quite accidentally involve only the spammers having hyphens in their names. There is also the previously mentioned risk in large networks of overfitting, where the network learns the training data by heart but doesn't gain the ability to evaluate new, unknown data.
Despite these limitations, you can check far more pages than before using the method described in this article, because the neural network pre-sorts potential spammers. If additional spammers are found manually, you can feed them into the network later in the form of training data.
Infos
- All listings for the article: http://www.linux-magazin.de/static/listings/magazin/2016/12/machine_learning/
- TensorFlow: https://www.tensorflow.org
- TensorFlow: Large-scale machine learning on heterogeneous systems? (2015): http://download.tensorflow.org/paper/whitepaper2015.pdf
- TFLearn: http://tflearn.org
- Bengio, Yoshua, Practical recommendations for gradient-based training of deep architectures. In G. Montavon, G.B. Orr, and K.-R. Müller (eds.), Neural Networks: Tricks of the Trade, 2nd ed. Springer-Verlag, 2012, pp. 437-478
- Overfitting: https://www.ibm.com/developerworks/community/blogs/jfp/entry/Overfitting_In_Machine_Learning
- Installing TensorFlow: https://www.tensorflow.org/versions/r0.10/get_started/os_setup.html#pip-installation
- Installing TFLearn: http://tflearn.org/installation/
The Author
Chris Hinze studies IT at the University of Erlangen-Nuremberg, Germany, and works at Benjamin Lochmann New Media GmbH as a web developer. His work there involves back ends for smartphone apps, and he has collaborated on automated spam recognition for http://homepage-baukasten.de.
« Previous 1 2 3 4
Buy this article as PDF
(incl. VAT)
Buy Linux Magazine
US / Canada
UK / Australia
Direct Download
Read full article as PDF:
Price $2.95
News
-
KaOS 2022.06 Now Available With KDE Plasma 5.25
The newest iteration of KaOS Linux not only adds the latest KDE Plasma desktop but sets LibreOffice as the default.
-
Manjaro 21.3.0 Is Now Available
Manjaro “Ruah” has been released and includes the latest Calamares installer, GNOME 42, and much more.
-
SpiralLinux is a New Linux Distribution Focused on Simplicity
A new Linux distribution, from the creator of GeckoLinux, is a Debian-based operating system with a focus on simplicity and ease of use.
-
HP Dev One Linux Laptop is Now Available for Pre-Order
The System76/HP collaboration Dev One laptop, geared toward developers, is now available for pre-order.
-
NixOS 22.5 Is Now Available
The latest release of NixOS with a much-improved package manager and a user-friendly graphical installer.
-
System76 Teams up with HP to Create the Dev One Laptop
HP and System76 have come together to develop a new laptop, powered by Pop!_OS and aimed toward developers.
-
Titan Linux is a New KDE Linux Based on Debian Stable
Titan Linux is a new Debian-based Linux distribution that features the KDE Plasma desktop with a focus on usability and performance.
-
Danielle Foré Has an Update for elementary OS 7
Now that Ubuntu 22.04 has been released, the team behind elementary OS is preparing for the upcoming 7.0 release.
-
Linux New Media Launches Open Source JobHub
New job website focuses on connecting technical and non-technical professionals with organizations in open source.
-
Ubuntu Cinnamon 22.04 Now Available
Ubuntu Cinnamon 22.04 has been released with all the additions from upstream as well as other features and improvements.