Elmano Ramalho Cavalcanti

On improving temporal and spatial mobility metrics for wireless ad hoc networks

This work shows that two well-known spatial and temporal mobility metrics for wireless multi-hop networks have limitations, possibly resulting in misleading results. Based on the concept of spatial dependence among nodes including transient periods of no movement, we propose mobility metrics able to promptly capture spatial and temporal dependence among mobile nodes. Through simulation, we compared the metrics over an extensive set of synthetic mobility models. The results revealed that our spatial metric can accurately capture spatial dependence in scenarios having different levels of node pause time. Our temporal metric also demonstrated to be better suited for capturing different levels of temporal dependence, without being biased by node speed. Moreover, we also proposed and validated a regression model for predicting our spatial mobility metric with a 94.8% goodness-of-fit.

Predicting mobility metrics through regression analysis for random, group, and grid-based mobility models in MANETs

In Mobile ad hoc Networks (MANETs), some mobility metrics are directly related to the performance of routing protocols. Creating accurate predict models for mobility metrics is an important advance for designing better mobility-adaptive protocols. Through regression analysis, we propose predictive formulas for three mobility metrics: link duration, node degree, and network partitioning, considering a set of random, group, and grid-based mobility models. We propose specific derived parameters for group and grid-based models, and show that they are good predictors for the metric values. The results also show that link duration and node degree are more predictive for random and grid-based, and less predictive for group-based models.

Degree of node proximity: a spatial mobility metric for manets

Spatial mobility metrics are used in several areas in mobile ad hoc networks (MANETs), such as protocol evaluation, and the design of clustering algorithms. We show that current spatial mobility metrics do not take into account spatial dependence in the absence of node movement. In order to provide a better understanding of spatial dependence, we propose a more comprehensive mobility metric, Degree of Node Proximity (DNP), based on the average distance among mobile nodes. Through simulation, we compared our metric against other well-known spatial metric over several classes of mobility models. DNP is shown able to capture spatial dependence in scenarios with different levels of node pause time. Besides that, DNP is not biased towards node speed, being able to distinguish group-based mobility models from others. As an additional contribution, we proposed several regression models for predicting DNP for the random, group, and grid-based mobility models under consideration in this paper.

A Proposal to Prevent Click-Fraud Using Clickable CAPTCHAs

Advertising on the Internet is a key factor for the success of several businesses nowadays. The Internet has evolved to a point where it has become possible to develop a business model completely based on Web advertising, which is important for the consolidation of such a model and the continuity of the Internet itself. However, it is often observed that some content publishers are dishonest and employ automated tools to generate traffic and profit by defrauding advertisers. Similarly, some advertisers use automated tools to click on the ads of their competitors, aiming to exhaust the budget of the competitors marketing departments. In this paper, differently of recent click fraud detection mechanisms, that take place after the fraud has already occurred, we propose an approach for preventing automated click-fraud, based on the use of click able CAPTCHAs.

Estimating the Impact of Mobility Models' Parameters on Mobility Metrics in MANETs

Aiming to analyze and classify the different types of existing mobility models, various mobility metrics have been proposed. However, little is known about the impact caused by changes in input parameters of a particular model over the mobility metrics. In this paper, several major mobility models and metrics are evaluated using a new methodology. Data collected were analyzed statistically using correlation as the statistical technique to explore the relationship between parameters and metrics. The results revealed that, depending on the configuration of the parameters of the models, it is possible that the metrics are not able to differentiate the models. Furthermore, the results also revealed that identical input parameters among the models caused different impacts on the mobility metrics. In particular, the minimum speed, in general, affects the metrics as much as the maximum speed does.

Enhancing OLSR protocol performance through improved detection of Spatial Dependence

One of the key factors of the OLSR routing protocol is the MPR selection algorithm, which is based only on the reachability of each neighbor, not taking into consideration how they are moving. As a result, the selected MPR set may be unstable. One way to improve the stability of the MPR set is through spatial mobility metrics that are able to promptly monitor the degree of movement correlation between a node and its neighbors. However, we show that current metrics have limitations on capturing the spatial correlation in the various states of collective motion. Through an enhanced spatial mobility metric, we propose a MPR selection algorithm, which was integrated into a new mobility-aware OLSR protocol. The new technique provided an performance gain in terms of packet delivery ratio and end-to-end delay, besides presenting fewer out of order packets.

On the applicability of mobility metrics for user movement pattern recognition in MANETs

In this paper we propose a set of mobility metrics, which are employed in the generation of supervised classification learning methods through the decision tree algorithm, with the goal to recognize user movement patterns in mobile ad hoc networks. Hundreds of scenarios produced by several well-known mobility models were employed for training and testing the supervised algorithms. The most suitable classification model showed an accuracy of 99.20% and Kappa index of 0.991, which indicates a high level of agreement between the classification model and real classification.

On realistic evaluation of recent spatial and temporal mobility metrics for mobile ad hoc networks

Results from recent studies have pointed out that two well-known spatial and temporal mobility metrics for mobile ad hoc networks have critical drawbacks. At the same time, two new spatial and one temporal mobility metrics have been proposed and validated through purely synthetic mobility models. However, due to the limitations that such models have to characterize the real trace provided by human mobility, this paper presents a comprehensive validation of these new metrics employing a recently well-accepted realistic mobility model. Furthermore, we discuss some potential applications for these recent mobile metrics.

Detecção e Avaliação de Cola em Provas Escolares Utilizando Mineração de Texto: um Estudo de Caso

A mineracao de texto tem sido utilizada para diversos propositos, como na extracao, classificacao e categorizacao de documentos. Dentro desse contexto, este trabalho apresenta um estudo de caso realizado na Universidade Federal de Campina Grande, em que se utilizou do modelo de vetor de espacos e da metrica de similaridade por cosseno na construcao de um modelo classificador capaz de inferir o nivel de cola existente entre provas escolares. A inferencia do modelo foi comparada a deteccao tradicional (humana) de cola e quantificada estatisticamente pelas seguintes metricas: acuracia, variacao da acuracia, precisao, recall, F-measure e indice Kappa. Os resultados apontam que o modelo criado apresentou uma boa qualidade na deteccao do nivel de cola entre provas, tornando possivel sua utilizacao em situacoes de ensino reais.