This work approaches the text document classification problem derived from the contest “Identify Keywords and Tags from Millions of Text Questions”, published on the website Kaggle. Using data from the StackOverflow website, the problem is to predict the tags assigned to questions. This categorization is multi-class and multi-tag, which means, a question can be assigned to different topics and can also have several tags. To solve this problem, we propose a 5-way multi-class classifier system. The results obtained by this classification scheme are discussed, by analysing certain score metrics of the classifier system. Competitive results were obtained by the 5-way classifier system, obtaining F1 scores ranging from 0.59 to 0.76. The main contribution of this paper lies on the preprocessing (which implements the feature extraction phase) and the multi-tag multi-class classification scheme.

This article presents a comparison of wind speed forecasting techniques, starting with the Auto-regressive Integrated Moving Average, followed by Artificial Intelligence-based techniques. The objective of this article is to compare these methods and provide readers with an idea of what method(s) to apply to solve their forecasting needs. The Artificial Intelligence-based techniques included in the comparison are Nearest Neighbors (the original method, and a version tuned by Differential Evolution), Fuzzy Forecasting, Artificial Neural Networks (designed and tuned by Genetic Algorithms), and Genetic Programming. These techniques were tested against twenty wind speed time series, obtained from Russian and Mexican weather stations, predicting the wind speed for 10 days, one day at a time. The results show that Nearest Neighbors using Differential Evolution outperforms the other methods. An idea this article delivers to the reader is: what part of the history of the time series to use as input to a forecaster? This question is answered by the reconstruction of phase space. Reconstruction methods approximate the phase space from the available data, yielding m (the system’s dimension) and τ (the sub-sampling constant), which can be used to determine the input for the different forecasting methods.

The kNN time series forecasting method is based on a very simple idea: similar states, observed in the past, most likely have produced similar output values. One has to look for nearest neighbors, according to some distance measure. The first idea that comes to mind when we see the nearest neighbor time series forecasting technique is to weigh the contribution of the different neighbors, according to distance to the present observation. The fuzzy version of the nearest neighbor time series forecasting technique (FNN), implicitly weighs the contribution of the different neighbors to the prediction, using the fuzzy membership of the linguistic terms as a kind of distance to the current observation. The training phase compiles all different scenarios of what has been observed in the time series past as a set of fuzzy rules. When we encounter a new situation and need to predict the future outcome, just like in normal fuzzy inference systems, the current observation is fuzzified, the set of rules is traversed to see which ones of them are activated (i.e., their antecedents are satisfied) and the outcome of the forecast is defuzzified by the common center of gravity rule. FNN was tested with a set of nonlinear chaotic time series, and the results compared with ARIMA. Its performance is satisfactory at this point and encouraging to continue evaluating the method, poising and improving it.

Load forecasting provides essential information for engineers and operators of an electric system. Using the forecast information, an electric utility company’s engineers make informed decisions in critical scenarios. The deregulation of energy industries makes load forecasting even more critical. In this article, the work we present, called Nearest Neighbors Load Forecasting (NNLF), was applied to very short-term load forecasting of electricity consumption at the national level in Mexico. The Energy Control National Center (CENACE—Spanish acronym) manages the National Interconnected System, working in a Real-Time Market system. The forecasting methodology we propose provides the information needed to solve the problem known as Economic Dispatch with Security Constraints for Multiple Intervals (MISCED). NNLF produces forecasts with a 15-min horizon to support decisions in the following four electric dispatch intervals. The hyperparameters used by Nearest Neighbors are tuned using Differential Evolution (DE), and the forecaster model inputs are determined using phase-space reconstruction. The developed models also use exogenous variables; we append a timestamp to each input (i.e., delay vector). The article presents a comparison between NNLF and other Machine Learning techniques: Artificial Neural Networks and Support Vector Regressors. NNLF outperformed those other techniques and the forecasting system they currently use.