Papers by Kaj-Mikael Björk
The Internet is moving from the traditional desktop network paradigm to a ubiquitous paradigm whe... more The Internet is moving from the traditional desktop network paradigm to a ubiquitous paradigm where a multitude of small computing devices such as computer chips and smart sensors are involved in daily activities and routines. This means that a rapidly growing amount of devices are connected to the Internet. At the same time, infrastructure-less and self-configuring systems like Mobile Ad hoc Networks (MANET) are gaining popularity since they provide a possibility for mobile devices to share information with each other without being dependent on a core infrastructure. Routing security in MANETs is, however, a significant challenge to wide scale adoption. One of the most severe security threats to MANET routing is the wormhole attack due to its ability to disrupt a significant proportion of network traffic, while simultaneously being difficult to detect. This paper provides an overview of recent research findings on wormhole attack detection in MANETs collected from a joint research project with
A Novel ELM Ensemble for Time Series Prediction
Proceedings in Adaptation, Learning and Optimization, 2019
This paper presents a novel methodology for time series prediction. It is based on Extreme Learni... more This paper presents a novel methodology for time series prediction. It is based on Extreme Learning Machines and an adaptive ensemble techniques. It is tested successfully on the CIF 2016 competition datasets which are composed of 72 time series in total. Among those, 48 time series are artificial with each having 108 training data points and 12 testing points. So for each artificial time series, there are 120 values, which is more than that of the rest 24 real time series.
Big Data Analytics, a research project in the complex digital era
To make sense of the enormous amount of data is a very important challenge in our society today. ... more To make sense of the enormous amount of data is a very important challenge in our society today. This short paper presents the activities in, as well as some results from, the project Big Data Analytics. The main focus of the project is set on dealing with both big data and the underlying analytics. The examples are found from different business contexts, such as retail, industrial and the financial sector.
ELM Feature Selection and SOM Data Visualization for Nursing Survey Datasets
Proceedings of ELM2019, 2020
Proceedings in Adaptation, Learning and Optimization, 2018
The paper proposes to analyze a data set of Finnish ranks of academic publication channels with E... more The paper proposes to analyze a data set of Finnish ranks of academic publication channels with Extreme Learning Machine (ELM). The purpose is to introduce and test recently proposed ELM-based mislabel detection approach with a rich set of features characterizing a publication channel. We will compare the architecture, accuracy, and, especially, the set of detected mislabels of the ELM-based approach to the corresponding reference results in .
Proceedings of the Fifth International Conference on Informatics in Control, Automation and Robotics Service, 2008
Finding the optimal OWA (ordered weighted averaging) operators is important in many decision supp... more Finding the optimal OWA (ordered weighted averaging) operators is important in many decision support problems. The OWA-operators enables the decision maker to model very different kinds of aggregator operators. The weights need to be, however, determined under some criteria, and can be found through the solution of some optimization problems. The important parameter called the level of orness may, in many cases, be uncertain to some degree. Decision makers are often able to estimate the level using fuzzy numbers. Therefore, this paper contributes to the current state of the art in OWA operators with a model that can determine the optimal (minimum variability) OWA operators under a (unsymmetrical triangular) fuzzy level of orness.
ArXiv, 2018
The paper proposes a new variant of a decision tree, called an Extreme Learning Tree. It consists... more The paper proposes a new variant of a decision tree, called an Extreme Learning Tree. It consists of an extremely random tree with non-linear data transformation, and a linear observer that provides predictions based on the leaf index where the data samples fall. The proposed method outperforms linear models on a benchmark dataset, and may be a building block for a future variant of Random Forest.
Distance Estimation for Incomplete Data by Extreme Learning Machine
Proceedings in Adaptation, Learning and Optimization, 2018
Data with missing values are very common in practice, yet many machine learning models are not de... more Data with missing values are very common in practice, yet many machine learning models are not designed to handle incomplete data. As most machine learning approaches can be formulated in terms of distance between samples, estimating these distances on data with missing values provides an effective way to use such models. This paper present a procedure to estimate the distances using the Extreme Learning Machine. Experimental comparison shows that the proposed approach achieves competitive accuracy with other methods on standard benchmark datasets.
Proceedings of the 12th ACM International Conference on PErvasive Technologies Related to Assistive Environments, 2019
This concept paper highlights a recently opened opportunity for large scale analytical algorithms... more This concept paper highlights a recently opened opportunity for large scale analytical algorithms to be trained directly on edge devices. Such approach is a response to the arising need of processing data generated by natural person (a human being), also known as personal data. Spiking Neural networks are the core method behind it: suitable for a low latency energy-constrained hardware, enabling local training or re-training, while not taking advantage of scalability available in the Cloud.
Scikit-ELM: An Extreme Learning Machine Toolbox for Dynamic and Scalable Learning
Proceedings of ELM2019, 2020
This paper presents a novel library for Extreme Learning Machines (ELM) called Scikit-ELM (https:... more This paper presents a novel library for Extreme Learning Machines (ELM) called Scikit-ELM (https://github.com/akusok/scikit-elm, https://scikit-elm.readthedocs.io). Usability and flexibility of the approach are the main focus points in this work, achieved primarily through a tight integration with Scikit-Learn, a de facto industry standard library in Machine Learning outside Deep Learning. Methodological advances enable great flexibility in dynamic addition of new classes to a trained model, or by allowing a model to forget previously learned data.
Website Classification from Webpage Renders
Proceedings of ELM2019, 2020
In this paper, we present a fast and accurate method for the classification of web content. Our a... more In this paper, we present a fast and accurate method for the classification of web content. Our algorithm uses the visual information of the main homepage saved in an image format by means of a full body snapshot. Sliding windows of different sizes and overlaps are used to obtain a large subset of images for each render. For each sub-image, a feature vector is extracted by means of a pre-trained deep learning model. A Extreme Learning Machine (ELM) model is trained for different values of hidden neurons using the large collection of features from a curated dataset of 5979 webpages with different classes: adult, alcohol, dating, gambling, shopping, tobacco and weapons. Our results show that the ELM classifier can be trained without the manual specific object tagging of the sub-images by giving excellent results in comparison to more complex deep learning models. A random forest classifier was trained for the specific class of weapons providing an accuracy of 95% with a F1 score of 0.8.
Brute-force Missing Data Extreme Learning Machine for Predicting Huntington's Disease
Proceedings of the 10th International Conference on PErvasive Technologies Related to Assistive Environments, 2017
This paper presents a novel procedure to train Extreme Learning Machine models on datasets with m... more This paper presents a novel procedure to train Extreme Learning Machine models on datasets with missing values. In effect, a separate model is learned to classify every sample in the test set, however, this is accomplished in an efficient manner which does not require accessing the training data repeatedly. Instead, a sparse structure is imposed on the input layer weights, which enables calculating the necessary statistics in the training phase. An application to predicting the progression of Huntington's disease from brain scans is presented. Experimental comparisons show promising results equivalent to the state of the art in machine learning with incomplete data.
ELMVIS+: Improved Nonlinear Visualization Technique Using Cosine Distance and Extreme Learning Machines
Proceedings in Adaptation, Learning and Optimization, 2016
This paper presents ELMVIS+, a significant improvement in ELMVIS methodology that enables faster ... more This paper presents ELMVIS+, a significant improvement in ELMVIS methodology that enables faster computation, more stable results and a wider application range. The novel cost function and a fast way of estimating it speeds up the method compared to ELMVIS, especially in large-dimensional datasets. The included Genetic Algorithms add global optimization that helps ELMVIS+ to find a better optimum. The improved methodology shows state-of-the-art performance in three different benchmark datasets.
Neurocomputing, 2016
In the paper, we examine the general regression problem under the missing data scenario. In order... more In the paper, we examine the general regression problem under the missing data scenario. In order to provide reliable estimates for the regression function (approximation), a novel methodology based on Gaussian Mixture Model and Extreme Learning Machine is developed. Gaussian Mixture Model is used to model the data distribution which is adapted to handle missing values, while Extreme Learning Machine enables to devise a multiple imputation strategy for final estimation. With multiple imputation and ensemble approach over many Extreme Learning Machines, final estimation is improved over the mean imputation performed only once to complete the data. The proposed methodology has longer running times compared to simple methods, but the overall increase in accuracy justifies this trade-off.
An Optimization Model for Tactical Planning of Wood Procurement
Hawaii International Conference on System Sciences, 2009
This paper presents a new LP (Linear Programming) model to solve a tactical wood procurement and ... more This paper presents a new LP (Linear Programming) model to solve a tactical wood procurement and harvesting problem. This optimization problem occurs in several wood supply chains today. The goal with the optimization is to find a 12 month procurement, harvesting and transportation plan that minimizes total costs. The case study is found in one of the Finnish wood supply
Applied Computational Intelligence and Soft Computing, 2012
The track of developing Economic Order Quantity (EOQ) models with uncertainties described as fuzz... more The track of developing Economic Order Quantity (EOQ) models with uncertainties described as fuzzy numbers has been very lucrative. In this paper, a fuzzy Economic Production Quantity (EPQ) model is developed to address a specific problem in a theoretical setting. Not only is the production time finite, but also backorders are allowed. The uncertainties, in the industrial context, come from the fact that the production availability is uncertain as well as the demand. These uncertainties will be handled with fuzzy numbers and the analytical solution to the optimization problem will be obtained. A theoretical example from the process industry is also given to illustrate the new model.
A Generalized Scalable Software Architecture for Analyzing Temporally Structured Big Data in the Cloud
Advances in Intelligent Systems and Computing, 2014
Software architectures that allow researchers to explore advanced modeling by scaling horizontall... more Software architectures that allow researchers to explore advanced modeling by scaling horizontally in the cloud can lead to new insights and improved accuracy of modeling results. We propose a generalized highly scalable information system architecture that researchers can employ in predictive analytics research for working with both historical data and real-time temporally structured big data. The proposed architecture is fully automated and uses the same analytical software for both training and live predictions.
SOM-ELM—Self-Organized Clustering using ELM
Neurocomputing, 2015
ABSTRACT This paper presents two new clustering techniques based on Extreme Learning Machine (ELM... more ABSTRACT This paper presents two new clustering techniques based on Extreme Learning Machine (ELM). These clustering techniques can incorporate a priori knowledge (of an expert) to define the optimal structure for the clusters, i.e. the number of points in each cluster. Using ELM, the first proposed clustering problem formulation can be rewritten as a Traveling Salesman Problem and solved by a heuristic optimization method. The second proposed clustering problem formulation includes both a priori knowledge and a self-organization based on a predefined map (or string). The clustering methods are successfully tested on 5 toy examples and 2 real datasets.
Neurocomputing, 2015
This paper proposes a methodology for identifying data samples that are likely to be mislabeled i... more This paper proposes a methodology for identifying data samples that are likely to be mislabeled in a cclass classification problem (dataset). The methodology relies on an assumption that the generalization error of a model learned from the data decreases if a label of some mislabeled sample is changed to its correct class. A general classification model used in the paper is OP-ELM; it also provides a fast way to estimate the generalization error by PRESS Leave-One-Out. It is tested on two toy datasets, as well as on real life datasets for one of which expert knowledge about the identified potential mislabels has been sought.
Uploads
Papers by Kaj-Mikael Björk