Prediktiv Analyse av Data fra BaneNOR - en mulighetsstudie

Abstract

Predictive Analysis indicates patterns in the subsets of data trained on standardized behavior of the railway systems, indicating deterioration or fault waiting to occur, which again accesses the state of life of the equipment. Machine learning (ML) is a technique used to handle complex data samples. The processing sequence is simplified using dimensional reduction methods that transform the data to lower dimensions while retaining meaningful properties. A semi-systematic review is performed to gain insight and knowledge on using predictive analysis in the railway system. The autoencoder (AE), Linear Discriminant Analysis (LDA), and Principle Component Analysis (PCA) methods are created and deployed using three validating scenarios, with K-means clustering deployed on each. In the end, the performance metrics calculation examines the clustering quality. It is observed shows that PCA has its strengths, particularly in reducing dimensionality and still capture the most variance in the data, AE not only reduces the dimensionality but also learns a representation of the data that is better at capturing complex non-linear relationships, which can be more beneficial for the clustering task. Hence, based on the results, AE would be preferred over PCA and LDA if the goal is to obtain distinct and well-separated clusters.

Predictive Analysis indicates patterns in the subsets of data trained on standardized behavior of the railway systems, indicating deterioration or fault waiting to occur, which again accesses the state of life of the equipment. Machine learning (ML) is a technique used to handle complex data samples. The processing sequence is simplified using dimensional reduction methods that transform the data to lower dimensions while retaining meaningful properties. A semi-systematic review is performed to gain insight and knowledge on using predictive analysis in the railway system. The autoencoder (AE), Linear Discriminant Analysis (LDA), and Principle Component Analysis (PCA) methods are created and deployed using three validating scenarios, with K-means clustering deployed on each. In the end, the performance metrics calculation examines the clustering quality. It is observed shows that PCA has its strengths, particularly in reducing dimensionality and still capture the most variance in the data, AE not only reduces the dimensionality but also learns a representation of the data that is better at capturing complex non-linear relationships, which can be more beneficial for the clustering task. Hence, based on the results, AE would be preferred over PCA and LDA if the goal is to obtain distinct and well-separated clusters.