Energy is a requirement for electronic devices.  A processor is a substantial part of computer components in terms of energy consumption. A great concern has risen over recent years about computers with regard to the energy consumption.  Taking accurate information about energy consumption of a processor allows us to predict energy flow features. However, using traditional classifiers may not enhance the accuracy of the prediction of energy consumption. Deep learning shows great promise for predicting energy consumption of a processor. Stacked auto-encoders has emerged a robust type of deep learning. This work investigates the effects of tuning stacked auto-encoder in computer processor with regard to the energy consumption. To search parameter space, a grid search based training method is adopted. To prepare data to prediction, a data preprocessing algorithm is also proposed. According to the obtained results, on average, the method provides 0.2% accuracy improvement along with a remarkable success in reducing parameter tuning error. Further, in receiver operating curve analysis, tuned stacked auto-encoder was able to increase value of are under the curve up to 0.5.

AAL (Ambient Assisted Living) existing architectures lack the sense of abstraction. The overall existing designs propose a set of elements combined with specific technologies. These visions of AAL systems narrows the possibilities and the choices ahead of the engineers and strict the range of using new technologies, which are likely to be easier and affordable. In this paper, we propose a context-aware RA (Reference Architecture) suitable for the design of distributed AAL systems. Our design is standardized and technology independent. Our aim is to provide a common background for developers and deployers to achieve a common understanding while designing the systems. The major gain is to reduce the efforts made while integrating several systems into one complete and stable environment. Ignoring all the specifications, the details and the objectives of the systems, we introduce the standard qualifications, practices and experiences that assimilate the core of every AAL oriented system. Our perception is global, unified and standard. In addition, it presents an infrastructure that would survive the evolution of technologies. It is adjustable and adaptable to the different possibilities of AAL applications.

The Air Pressure System (APS) is a type of function used in heavy vehicles to assist braking and gear changing. The APS failure dataset consists of the daily operational sensor data from failed Scania trucks. The dataset is crucial to the manufacturer as it allows to isolate components which caused the failure. However, missing values and imbalanced class problems are the two most challenging limitations of this dataset to predict the cause of the failure. The prediction results can be affected by the way of handling these missing values and imbalanced class problem. In this paper, we have examined and presented the impact of five different missing value imputation techniques namely: Expectation Maximization, Mean Imputation, Soft Impute, MICE, and Iterative SVD in producing significantly better results. We have also performed an empirical comparison of their performance by applying five different classifiers namely: Naive Bayes, KNN, SVM, Random Forest, and Gradient Boosted Tree on this highly imbalanced dataset. The primary aim of this study is to observe the impact of the mentioned missing value imputation techniques in the enhancement of the prediction results, performing an empirical comparison to figure out the best classification model and imputation technique. We found that the MICE imputation and the random under-sampling techniques are the highest influential techniques for improving the prediction performance and false negative rate.

The rapid population growth in Dar Es Salaam has prompted the demand of effective transport system in the city.  This tremendous rise of population led to serious road traffic congestions, which brings a number of challenges into the city and other growing urban areas. City authorities attempted various solutions to control the traffic congestions such as construction of new roads, expansion of existing roads, installation of traffic lights and other transportation infrastructures such as re-establishment of commuter train to operate within the city but they couldn’t effectively relieve the problem. Eventually, the Government of Tanzania (GoT) supported the city’s effort by establishing the organ called Dar Es Salaam Rapid Transit (DART) to supervise the implementation and operation of Bus Rapid Transit (BRT) system. The BRT system provided direct benefits to passengers such as minimal travel time, improved reliability as compared to other public transport commonly known as daladala, and reduced accident as BRT buses travel in their dedicated lanes. Despite these benefits there still persist transportation challenges with the BRT, where passengers still suffer from waiting on very long queue during ticket booking, shortage of smart cards, they are unable to check balance direct from their mobile phones, as well as they fail to top-up onto their card’s balance using their smart phones. This paper presents a software technology approach that would help passengers to check balance, send request specifying station to board a bus and check the bus arrival time at any station.

Among natural disasters observed each year, flood represents 40% and remains one of the most important problems that many governments want to solve. Each year flood is responsible for many damages that cost a lot of money and even lot of people’s life. To reduce these damages caused, flood forecasting and warning systems which are able to alert people when a flood occurs have been built. However, most of these flood forecasting systems(FFS) are usually designed for specific regions and mostly for developed countries and are not suitable for developing countries because of climatological and environmental parameters difference. The problem of flood forecasting in developing countries could be explained in one part by the lack of meteorological stations and hydraulic stations necessary for flood forecasting systems to make predictions. Moreover, existing flood forecasting systems, have forecast accuracy problem because of constant changes of the environment and climate usually caused by anthropic factors. To face these problems, this work proposes an auto-adaptive flood forecasting system based on hydraulic models and data analysis techniques on meteorological and wireless sensors networks data to realize reliable forecast. The large number of experiments conducted show that the solutions proposed in this work performed well.

Recommender systems that possess adequate information about users and analyze their information, are capable of offering appropriate items to customers. Collaborative filtering method is one of the popular recommender system approaches that produces the best suggestions by identifying similar users or items based on their previous transactions. The low accuracy of suggestions is one of the major concerns in the collaborative filtering method. Several methods have been introduced to enhance the accuracy of this method through the discovering association rules and using evolutionary algorithms such as particle swarm optimization. However, their runtime performance does not satisfy this need, thus this article proposes an efficient method of producing cred associations rules with higher performances based on a genetic algorithm. Evaluations were performed on the data set of MovieLens. The parameters of the assessment are: run time, the average of quality rules, recall, precision, accuracy and F1-measurement. The experimental evaluation of a system based on our algorithm outperforms show than the performance of the multi-objective particle swarm optimization association rule mining algorithm, finally runtime has dropped by around 10%.