Social-media monitoring for cold-start recommendations

Generating personalized movie recommendations to users is a problem that most commonly relies on user-movie ratings. These ratings are generally used either to understand the user preferences or to recommend movies that users with similar rating patterns have rated highly. However, movie recommenders are often subject to the Cold-Start problem: new movies have not been rated by anyone, so, they will not be recommended to anyone; likewise, the preferences of new users who have not rated any movie cannot be learned. In parallel, Social-Media platforms, such as Twitter, collect great amounts of user feedback on movies, as these are very popular nowadays. This thesis proposes to explore feedback shared on Twitter to predict the popularity of new movies and show how it can be used to tackle the Cold-Start problem. It also proposes, at a finer grain, to explore the reputation of directors and actors on IMDb to tackle the Cold-Start problem. To assess these aspects, a Reputation-enhanced Recommendation Algorithm is implemented and evaluated on a crawled IMDb dataset with previous user ratings of old movies,together with Twitter data crawled from January 2014 to March 2014, to recommend 60 movies affected by the Cold-Start problem. Twitter revealed to be a strong reputation predictor, and the Reputation-enhanced Recommendation Algorithm improved over several baseline methods. Additionally, the algorithm also proved to be useful when recommending movies in an extreme Cold-Start scenario, where both new movies and users are affected by the Cold-Start problem.

On and off air: an ethnographic exploration of minority radio in Portugal

Taking a Media Anthropology’s approach to dynamics of mediated selfrepresentation in migratory contexts, this thesis starts by mapping radio initiatives produced by, for and/or with migrants in Portugal. To further explore dynamics of support of initial settlement in the country, community-making, cultural reproduction, and transnational connectivity - found both in the mapping stage and the minority media literature (e.g. Kosnick, 2007; Rigoni & Saitta, 2012; Silverstone & Georgiou, 2005) - a case study was selected: the station awarded with the first bilingual license in Portugal. The station in question caters largely to the British population presenting themselves as “expats” and residing in the Algarve. The ethnographic strategy to research it consisted of “following the radio” (Marcus, 1995) beyond the station and into the events and establishments it announces on air, so as to relate production and consumption realms. The leading research question asks how does locally produced radio play into “expats” processes of management of cultural identity – and what are the specificities of its role? Drawing on conceptualizations of lifestyle migration (Benson & O’Reilly, 2009), production of locality (Appadurai 1996) and the public sphere (Butsch, 2007; Calhoun & et al, 1992; Dahlgren, 2006), this thesis contributes to valuing radio as a productive gateway to research migrants’ construction of belonging, to inscribe a counterpoint in the field of minority media, and to debate conceptualizations of migratory categories and flows. Specifically, this thesis argues that the station fulfills similar roles to other minority radio initiatives but in ways that are specific to the population being catered to. Namely, unlike other minority stations, radio facilitates the process of transitioning between categories along on a continuum linking tourists and migrants. It also reflects and participates in strategies of reterritorialization that rest on functional and partial modes of incorporation. While contributing to sustain a translocality (Appadurai, 1996) it indexes and fosters a stance of connection that is symbolically and materially connected to the UK and other “neighborhoods” but is, simultaneously, oriented to engaging with the Algarve as “home”. Yet, besides reifying a British cultural identity, radio’s oral, repetitive and ephemeral discourse particularly trivializes the reproduction of an ambivalent stance of connection with place that is shared by other “expats”. This dynamic is related to migratory projects driven by social imaginaries fostered by international media that stimulate the search for idealized ways of living, which the radio associates with the Algarve. While recurrently localizing and validating the narrative projecting an idealized “good life”, radio amplifies dynamics among migrants that seem to reaffirm the migratory move as a good choice.

An agent based architecture to support monitoring in plug and produce manufacturing systems using knowledge extraction

In recent years a set of production paradigms were proposed in order to capacitate manufacturers to meet the new market requirements, such as the shift in demand for highly customized products resulting in a shorter product life cycle, rather than the traditional mass production standardized consumables. These new paradigms advocate solutions capable of facing these requirements, empowering manufacturing systems with a high capacity to adapt along with elevated flexibility and robustness in order to deal with disturbances, like unexpected orders or malfunctions. Evolvable Production Systems propose a solution based on the usage of modularity and self-organization with a fine granularity level, supporting pluggability and in this way allowing companies to add and/or remove components during execution without any extra re-programming effort. However, current monitoring software was not designed to fully support these characteristics, being commonly based on centralized SCADA systems, incapable of re-adapting during execution to the unexpected plugging/unplugging of devices nor changes in the entire system’s topology. Considering these aspects, the work developed for this thesis encompasses a fully distributed agent-based architecture, capable of performing knowledge extraction at different levels of abstraction without sacrificing the capacity to add and/or remove monitoring entities, responsible for data extraction and analysis, during runtime.

Monitoring liquidity risk at BPI GA

Field lab: Business project

A new credit monitoring framework for agents' networks

The purpose of this work is to develop a practicable approach for Telecom firms to manage the credit risk exposition to their commercial agents’ network. Particularly it will try to approach the problem of credit concession to clients’ from a corporation perspective and explore the particular scenario of agents that are part of the commercial chain of the corporation and therefore are not end-users. The agents’ network that served as a model for the presented study is composed by companies that, at the same time, are both clients and suppliers of the Telecommunication Company. In that sense the credit exposition analysis must took into consideration all financial fluxes, both inbound and outbound. The current strain on the Financial Sector in Portugal, and other peripheral European economies, combined with the high leverage situation of most companies, generates an environment prone to credit default risk. Due to these circumstances managing credit risk exposure is becoming increasingly a critical function for every company Financial Department. The approach designed in the current study combined two traditional risk monitoring tools: credit risk scoring and credit limitation policies. The objective was to design a new credit monitoring framework that is more flexible, uses both external and internal relationship history to assess risk and takes into consideration commercial objectives inside the agents’ network. Although not explored at length, the blueprint of a Credit Governance model was created for implementing the new credit monitoring framework inside the telecom firm. The Telecom Company that served as a model for the present work decided to implement the new Credit Monitoring framework after this was presented to its Executive Commission.

Assessment of NDVI, land surface temperature and precipitation anomalies for drought monitoring in Bayankhongor province, Mongolia

During the last decade Mongolia’s region was characterized by a rapid increase of both severity and frequency of drought events, leading to pasture reduction. Drought monitoring and assessment plays an important role in the region’s early warning systems as a way to mitigate the negative impacts in social, economic and environmental sectors. Nowadays it is possible to access information related to the hydrologic cycle through remote sensing, which provides a continuous monitoring of variables over very large areas where the weather stations are sparse. The present thesis aimed to explore the possibility of using NDVI as a potential drought indicator by studying anomaly patterns and correlations with other two climate variables, LST and precipitation. The study covered the growing season (March to September) of a fifteen year period, between 2000 and 2014, for Bayankhongor province in southwest Mongolia. The datasets used were MODIS NDVI, LST and TRMM Precipitation, which processing and analysis was supported by QGIS software and Python programming language. Monthly anomaly correlations between NDVI-LST and NDVI-Precipitation were generated as well as temporal correlations for the growing season for known drought years (2001, 2002 and 2009). The results show that the three variables follow a seasonal pattern expected for a northern hemisphere region, with occurrence of the rainy season in the summer months. The values of both NDVI and precipitation are remarkably low while LST values are high, which is explained by the region’s climate and ecosystems. The NDVI average, generally, reached higher values with high precipitation values and low LST values. The year of 2001 was the driest year of the time-series, while 2003 was the wet year with healthier vegetation. Monthly correlations registered weak results with low significance, with exception of NDVI-LST and NDVI-Precipitation correlations for June, July and August of 2002. The temporal correlations for the growing season also revealed weak results. The overall relationship between the variables anomalies showed weak correlation results with low significance, which suggests that an accurate answer for predicting drought using the relation between NDVI, LST and Precipitation cannot be given. Additional research should take place in order to achieve more conclusive results. However the NDVI anomaly images show that NDVI is a suitable drought index for Bayankhongor province.