Abstract
This article describes our approach for the Author Identification task introduced in PAN 2015. Given a set of documents written by the same author and a questioned document with an unknown author, the task is to decide whether the questioned document was written by the same author as the other documents or not. Our approach uses Random Forest and a feature-encoding scheme based on the Universal Background Model strategy, building different feature vectors that describe: 1) the complete population of authors in a dataset, 2) the known author, 3) the questioned document and combines the three of them in a single representation.

Abstract
The goal of this project, one of the proposals of the EPS@ISEP 2014 Spring, was to develop an Aquaponics System. Over recent years Aquaponics systems have received increased attention due to its possibilities in helping reduce strain on resources within 1st and 3rd world countries. Aquaponics is the combination of Hydroponics and Aquaculture and mimics a natural environment in order to successfully apply and enhance the understanding of natural cycles within an indoor process. By using this knowledge of natural cycles it was possible to create a system with the capabilities similar to that of a natural environment with the benefits of electronic adaptions to enhance the overall efficiency of the system. The multinational team involved in its development was composed of five students, from five countries and fields of study. This paper covers their solution, involving overall design, the technology involved and the benefits it could bring to the current market. The team was able to achieve the final rendered Computer Aided Design (CAD) drawings, successfully performed all the electronic testing, and designed a solution under budget. Furthermore, the solution presented was deeply studied from the sustainability viewpoint and the team also developed a product specific marketing plan. Finally, the students involved in this project obtained new knowledge and skills.

Abstract
The Special Issue of Information Processing and Management includes research papers on the intersection between time and information retrieval. In 'Evaluating Document Filtering Systems over Time', Tom Kenter and Krisztian Balog propose a time-aware way of measuring a system's performance at filtering documents. Manika Kar, SeAa7acute;rgio Nunes and Cristina Ribeiro present interesting methods for summarizing changes in dynamic text collections over time in their paper 'Summarization of Changes in Dynamic Text Collection using Latent Dirichlet Allocation Model.' Hideo Joho, Adam Jatowt and Roi Blanco report on the temporal information searching behaviour of users and their strategies for dealing with searches that have a temporal nature in 'Temporal Information Searching Behaviour and Strategies', a user study. In controlled settings, thirty participants are asked to perform searches on an array of topics on the web to find information related to particular time scopes. Adam Jatowt, Ching-man Au Yeung and Katsumi Tanaka present a 'Generic Method for Detecting Content Time of Documents'. The authors propose several methods for estimating the focus time of documents, i.e. the time a document's content refers to. Xujian Zhao, Peiquan Jin and Lihua Yue present an approach to determining the time of the underlying topic or event in their article entitled 'Discovering Topic Time from Web News'.

Abstract
This article deals with a real-life waste collection routing problem. To efficiently plan waste collection, large municipalities may be partitioned into convenient sectors and only then can routing problems be solved in each sector. Three diverse situations are described, resulting in three different new models. In the first situation, there is a single point of waste disposal from where the vehicles depart and to where they return. The vehicle fleet comprises three types of collection vehicles. In the second, the garage does not match any of the points of disposal. The vehicle is unique and the points of disposal (landfills or transfer stations) may have limitations in terms of the number of visits per day. In the third situation, disposal points are multiple (they do not coincide with the garage), they are limited in the number of visits, and the fleet is composed of two types of vehicles. Computational results based not only on instances adapted from the literature but also on real cases are presented and analyzed. In particular, the results also show the effectiveness of combining sectorization and routing to solve waste collection problems. (c) 2015 Wiley Periodicals, Inc. NETWORKS, Vol. 65(2), 155-165 2015

Abstract
In this paper we introduce the dial-a-ride problem with split requests and profits (DARPSRP). Users place transportation requests, specifying a pickup location, a delivery location, and a time window for either of the two. Based on maximum user ride time considerations, the second time window is generated. A given fleet of vehicles, each with a certain capacity, is available to serve these requests, and maximum route duration constraints have to be respected. Each request is associated with a revenue and the objective is to maximize the total profit, that is, the total revenue minus the total costs. Transportation requests involving several persons may be split if it is beneficial to do so. We formulate the DARPSRP as a mixed-integer program using position variables and in terms of a path-based formulation. For the solution of the latter, we design a branch-and-price algorithm. The largest instance solved to optimality, when applied to available instances from the literature, has 40 requests; when applied to newly generated instances, the largest instance solved to optimality consists of 24 requests. To solve larger instances a variable neighborhood search algorithm is developed. We investigate the impact of request splitting under different geographical settings, assuming favorable settings for request splitting in terms of the number of people per request. The largest benefits from request splitting are obtained for problem settings exhibiting clustered customer locations.

Abstract
Social networks are making part of the daily routine of millions of users. Twitter is among Facebook and Instagram one of the most used, and can be seen as a relevant source of information as users share not only daily status, but rapidly propagate news and events that occur worldwide. Considering the dynamic nature of social networks, and their potential in information spread, it is imperative to find learning strategies able to learn in these environments and cope with their dynamic nature. Time plays an important role by easily out-dating information, being crucial to understand how informative can past events be to current learning models and for how long it is relevant to store previously seen information, to avoid the computation burden associated with the amount of data produced. In this paper we study the impact of longstanding messages in micro-blogging classification by using different training time-window sizes in the learning process. Since there are few studies dealing with drift in Twitter and thus little is known about the types of drift that may occur, we simulate different types of drift in an artificial dataset to evaluate and validate our strategy. Results shed light on the relevance of previously seen examples according to different types of drift.