Towards forecast techniques for business analysts of large commercial data sets using matrix factorization methods

Abstract

This research article suggests that there are significant benefits in exposing demand planners to forecasting methods using matrix completion techniques. This study aims to contribute to a better understanding of the field of forecasting with multivariate time series prediction by focusing on the dimension of large commercial data sets with hierarchies. This research highlights that there has neither been sufficient academic research in this sub-field nor dissemination among practitioners in the business sector. This study seeks to innovate by presenting a matrix completion method for short-term demand forecast of time series data on relevant commercial problems. Albeit computing intensive, this method outperforms the state of the art while remaining accessible to business users. The object of research is matrix completion for time series in a big data context within the industry. The subject of the research is forecasting product demand using techniques for multivariate hierarchical time series prediction that are both precise and accessible to non-technical business experts. Apart from a methodological innovation, this research seeks to introduce practitioners to novel methods for hierarchical multivariate time series prediction. The research outcome is of interest for organizations requiring precise forecasts yet lacking the appropriate human capital to develop them. 1. Non-technical Summary Companies make an effort to collect and evaluate supply chain data with the objective of providing accurate forecasts to support decision-makers in estimating the company performance. To contribute towards this objective, this research presents a method for short-term forecasts based on matrix factorization techniques that is well-suited for demand planners with domain knowledge but limited understanding of machine learning methods. Organizations up-skilling existing personnel into data scientists and seeking to increase forecast accuracy will find value in the proposed methodology.

Learning Ensembles of Anomaly Detectors on Synthetic Data

“…To perform online aggregation, we are going to use long-term aggregation strategies [17] along with approaches to model quasi-periodic data [5] and extraction of trends in the presence of non-stationary noise with long tails [3,4]. Approaches to multidimensional time-series prediction [22] and multichannel anomaly detection [13] will allow detecting complex anomalies related to change of dependencies between time-series components.…”

Topology-Based Clusterwise Regression for User Segmentation and Demand Forecasting

“…CRM models emphasizing timing patters to predict future purchase activities have been proposed previously by [2] and [3] and serve as an inspiration for this work. The use of novel machine learning methods is a promising area with little academic research and insufficient efforts to expose practitioners to them according to [4] and [5]. In addition, over 40% of analysts still use primarily traditional forecasting methods, [6].…”

“…Another contribution of this work is the presentation of an approach based on matrix factorization methods (MF). They are used in a variety of applications such as recommender systems, demand forecasting, [5], signal processing, [36], computer vision, [37], and others.…”

An Industry Case of Large-Scale Demand Forecasting of Hierarchical Components

“…Matrix Factorization (MF) methods are used in a variety of applications such as recommender systems, signal processing, [23], computer vision, [24], and others. The second contribution of this work is adapting a method discussed in [25] and [26] to demand forecasting in manufacturing. Let Y be T × n sparse or dense matrix of observations of n objects spanning the period of T time steps, i.e.…”