Pan African Conference on Science, Computing and Telecommunications (PACT) 2017

Permanent URI for this collection

http://hdl.handle.net/11071/5175

Browse

Now showing 1 - 20 of 20

Detecting scanning computer worms using machine learning and darkspace network traffic
(Strathmore University, 2017) Ochieng, Nelson; Ismail, Ateya; Waweru, Mwangi; Orero, Joseph
The subject of this paper is computer worm detection in a network. Computers worms have been defined as a process that can cause a possibly evolved copy of it to execute on a remote computer. They do not require human intervention to propagate; neither do they need to attach themselves to existing files. Computer worms spread very rapidly and modern worm authors obfuscate their code to make it difficult to detect them. This paper proposes to use machine learning to detect them. The paper deviates from existing approaches in that it uses the darkspace network traffic attributed to an actual worm attack to validate the algorithms. In addition, it attempts to understand the threat model, the feature set and the detection algorithms to explain the best combination of features and why the best algorithms succeeds where others have failed.
Reducing Infant mortality using mobile applications : a paper on the potential impact of using technology to increase vaccination in infants
(Strathmore University, 2017) Yego, Caleb Loitabei
Today’s world is characterised by professionals who have minimal time with their families and especially their young children. Working mothers find it hard to keep track of the vaccines their babies have received and there exists no mechanisms to remind the mothers of the same. This paper proposes the development of a mobile application to increase the number of children who get timely vaccination in line with the attainment of United Nations Millennium Development Goal 4 (MDG4). This application is aimed to be used by working mothers for tracking vaccination dates for infants below 12 months of age. The Chanjo App is to be located in Kenya, where the target demography is found.Using the Object Oriented Analysis and Design methodology,the plan is to develop an application that will allow mothers to plan their infant’s immunisation as well as receive timely notifications on their schedule. This will be aided by the current boom in the uptake of smartphones (android and iOS) in the Kenyan consumer space [6]. It can be rightly assumed that the system can be implemented in numbers sufficient to achieve the MDG4 goal.
Real-time monitoring model for early detection of crop diseases
(Strathmore University, 2017) Toroitich, Patrick K.; Orero, Joseph
The agricultural sector has been a key backbone to Kenya’s economy. Agriculture has played a key role in the economy through agricultural farm produce exports and job creation hence improving and maintaining good farming practices is critical in ensuring agricultural yields. Potato (Solanum tuberosum L.) is a major food and cash crop in the Kenyan highlands, widely grown by small-scale farmers. However, early detection of potato diseases still remains a challenge for both farmers and agricultural extension officers. Consequently agricultural extension officers who play a critical role in training and creating awareness on sound agricultural practices are few and often lack sufficient knowledge and tools. Current techniques used for determining and detecting of crop diseases have relied upon use human vision systems that try to examine physical and phenotypic characteristics such as leaf and stem color. This technique is indeed important for diagnosis of crop diseases, however the use of this technique is not efficient to support early detection of diseases. This study proposed the use of internet of things technology and machine learning techniques for the prediction of potato late blight disease. Temperature and humidity sensor probes placed on the potato were instrumental in monitoring conditions for potato late blight disease on a farm. These parameters constituted abiotic factors that favor the development and growth of Phytophthora infestants. Back propagation neural network model was suitable for the prediction of potato late blight disease. In designing the potato late blight prediction model, historical weather data, potato variety tolerance on late blight disease was used to build an artificial neural network disease prediction model. Incoming data streams from the sensors was used to determine level and risk of blight. This study focused on a moderate susceptible cultivator of potato in developing the model. The algorithm was preferred due to its strengths in adaptive learning. The developed model achieved an accuracy of 94%.
A Study on growth condition analysis of rice using drones
(Strathmore University, 2017) Kazuki, Murata; Atsushi, Ito; Hiroyuki, Hatano
Self-sufficiency of food is one of the most important target for Japan, however, because of decreasing the population of agriculture sector due to the progress of aging, it is difficult to realize this target. One of the main reasons of decreasing the population in agriculture is hard work in the field and requirement of experience. Our study focuses on supporting aging farmers, weekend farmers and new comers in the agriculture to reduce labor and cost, and providing system to compensate experience by using ICT. For this purpose, we are developing technology to evaluate growing condition of rice and vegetables in a field by using a drone and a multi-spectrum camera. By using this technology, it might be possible to reduce the hard work under the scorching sun and to support unexperienced young farmers. In this paper, we mention the outline of the technology that we used and the result of feasibility study to analyze growth condition of rice using drone. We measured normalized difference vegetation index (NDVI) and normalized difference red edge index (NDRE) from a drone and analyzed the data to estimate growth condition of rice. We found it was possible to find the area where growth is delaying and estimate heading day.
Foundations of an autonomic manager for maintaining quality of service in enterprise data warehouses
(Strathmore University, 2017) Omondi, Allan O.; Ateya, Ismail L.; Wanyembi, Gregory N.
Data stored in an Enterprise Data Warehouse (EDW) is an essential asset to enterprises. Through efficient access to data (where efficiency is quantitatively measured in terms of speed), SMEs can enhance their growth, productivity, and global competitiveness. This can in turn lead to a positive impact on a country's Gross Domestic Product. The purpose of this paper is to present the building blocks required to maximize the speed of data access from EDWs in a self-adaptive manner. Reinforcement Learning (RL) in a fully observable, stochastic environment is proposed. The subsequent solution to a Markov Decision Process is highlighted as the core part of the RL.
A Model for real-time remote monitoring of asthmatic patients
(Strathmore University, 2017) Oduor, Thomas J.; Orero, Joseph
Asthma affects a large number of people in the world and it is the most common chronic illness of the lung among children. Monitoring and assessing the severity of a child’s asthma at home has proven difficult and costly in many low resource settings, like Kenya. This is due to lack of easy to use and cost effective in-home monitoring tools for children suffering from asthma. In this work, we present a user-friendly wearable device that allows in-home remote monitoring of asthmatic child and real-time determination of degree of asthma severity. The device captures simultaneous measurements of heart rate and oxygen saturation using a pulse oximeter sensor. These readings are then analyzed using clinical algorithm for asthma management, and the resulting severity level sent to both the parent of the child and the assigned clinician. The device was developed using Arduino Uno microcontroller, pulse oximeter sensor and Global System for Mobile Communication (GSM) module.
Taxonomy for digital forensic evidence
(Strathmore University, 2017) Karie, Nickson M.; Kebande, Victor R.; Venter, H. S.
Modern society has increased its dependencies on digital systems and computer networks in almost every area of life today. Although this dependency is good it has opened a whole new world of possibilities for criminals to exploit. This has been seen in areas where criminals are able to use existing digital systems to share information and to reinforce their hacking techniques for nefarious purposes. As a result, major potential security risks, such as malicious insiders, data loss or leakage and policy violations have now invaded our digital world with worrying trends of digital and cyber-crimes. This, therefore, has made computer based information a primary source of digital evidence in many legal matters and digital investigations. The understanding of the different types of information generated by computer systems is thus an importance aspect of any digital forensic investigation process. For this reason, this paper reviews existing digital forensic research literature and highlights the different types of digital evidence that can potentially be admissible in our courts of law today. In conducting this research study, however, it was difficult for the authors to review all the existing research literature in the digital forensic domain; hence, sampling and randomization techniques were employed to facilitate the review of the gathered literature. The taxonomy classifies a large number of Digital Forensic Evidence (DFE) into a few well-defined and easily understood categories which can be useful, for example, the future developments of digital forensic tools. In addition, the taxonomy can also be helpful to practitioners, for example, in classifying the different types of DFE that can be admissible in courts. The main contribution of this research is, therefore, to propose a taxonomy for DFE that can assist digital forensic analysts and forensic practitioners to understand the different types of evidence with ease and their applicability in different legal matters.
UAV heading controller using reinforcement learning
(Strathmore University, 2017) Kimathi, Stephen; Kang’ethe, Samuel; Kihato, Peter
The control of heading of an Unmanned Aerial Vehicle is a vital operation. It is accomplished by employing a design of control algorithms that control its flying direction. The available autopilots exploit Proportional-Integral-Derivative (PID) based heading controllers. Here we propose an adaptive controller based on reinforcement learning. The heading controller will be designed in Matlab/Simulink for controlling a UAV in X-Plane test platform. Through this platform, the performance of the designed controller is compared with that of a well-tuned PID controller using real time simulations. The results show that the proposed method performs better.
An Elliptic curve digital signature algorithm (ECDSA) for securing data : an exemplar of securing patient's data
(Strathmore University, 2017) Waruhari, Philomena; Nderu, Lawrence
In this paper, we present the progress of our work in the creation and implementation of an Elliptic Curve Digital Signature Algorithm (ECDSA). We present the design of the algorithm and its implementation in encryption of medical data. ECDSA PHP ECC code has been used to implement the digital signatures over elliptic curve P-256. The work presented highlights practical implementation of ECDSA signature generation to secure and authenticate patient laboratory test results in a Laboratory Information System (LIS). Future work will demonstrate the implementation of decryption using the ECDSA. With the inherent superiority capability of Elliptic Curves (EC) in securing data, our algorithm is highly secure and can be adapted in many areas where data privacy and security is paramount.
Energy consumption in cloud computing environments
(Strathmore University, 2017) Kenga, Mosoti Derdus; Omwenga O, Vincent O.; Ogao, Patrick J.
Datacentres are becoming indispensable infrastructure for supporting the services offered by cloud computing. Unfortunately, they consume a great deal of energy accounting for 3% of global electrical energy consumption. The effect of this is that, cloud providers experience high operating costs, which leading to increased Total Cost of Ownership (TCO) of datacentre infrastructure. Moreover, there is increased carbon dioxide emissions that affects the universe. This paper presents a survey on the various ways in which energy is consumed in datacentre infrastructure. The factors that influence energy consumption within a datacentre is presented as well.
Vehicle exhaust emissions inspection system for roadworthiness enforcement
(Strathmore University, 2017) Mwenda, Reuben K.; Orero, Joseph Onderi
Air pollution has been a growing concern as Kenya tries to industrialize. Increase in the number of vehicles and factories as well as constructions in Nairobi make this all the more critical. This polluted air has far reaching consequences which include illnesses that lead to death. Measuring the concentration of air pollutants is necessary to establish the quality of air in the city. By extension, measuring the concentration of pollutants being emitted through vehicle exhaust fumes can help establish if the vehicle is worthy to be on the road. To best measure the degree of these pollutants, random on-the-road inspection of vehicle inspection of vehicle exhaust emissions is key. However, this has not been achieved by the Kenyan law enforcement agencies. The ability to inspect the emissions from cars on the road will help law enforcement remove unroadworthy vehicles from the roads and thus minimize air pollution caused by vehicles. Conventional inspection methods are done in controlled environments such as laboratories. Vehicles are driven in and are inspected while they remain stationary. These controlled tests fall short of revealing the true state of a vehicle’s exhaust emissions: the fumes emitted while a car is on open road are different in composition from those emitted in such a controlled environment. In addition, manufacturers can tweak their vehicles to emit gases that are within the prescribed thresholds as was done by Volkswagen in order to meet and exceed the US Environment Protection Agency standards. This study will present a model that utilizes sensors to assess the level of pollutants produced from a vehicle exhaust to the air and register these to back-end server hosted on the cloud. The model will have an LCD screen on which law enforcement can view levels of pollutants as measured by the sensors. The information will be stored in
Scalable dataspace construction
(Strathmore University, 2017) Shibwabo, Bernard K.; Wanyembi, Gregory N.; Ateya, Ismail L.; Omwenga, Vincent O.
This paper proposes the design and implementation of scalable dataspaces based on efficient data structures. Dataspaces are often likely to exhibit a multidimensional structure due to the unpredictable neighbour relationship between participants coupled by the continuous exponential growth of data. Layered range trees are incorporated to the proposed solution as multidimensional binary trees which are used to perform d-dimensional orthogonal range indexing and searching. Furthermore, the solution is readily extensible to multiple dimensions, raising the possibility of volume searches and even extension to attribute space. We begin by a study of the important literature and dataspace designs. A scalable design and implementation is further presented. Finally, we conduct experimental evaluation to illustrate the finer performance of proposed techniques. The design of a scalable dataspace is important in order to bridge the gap resulting from the lack of coexistence of data entities in the spatial domain as a key milestone towards pay-as-you-go systems integration
Provision of caller ring back tones for IP multimedia platforms
(Strathmore University, 2017) Nkinyili, Tiberius; Gavole, Vitalis
Customised Caller Ring Back Tones (CRBT) are used to entertain callers by playing a media clip while the callee’s phone is ringing. CRBT involves the mobile operator replacing the standard audio clip with a clip selected by the user, in this case the callee. The service may be offered by 3rd party application providers, but can also be offered by mobile operators themselves. CRBT service can be supported by different mobile network infrastructures including the circuit switched GSM networks and IP multimedia networks such as IMS. These networks need integration of additional components to provide the CRBT service.3GGP has standardized the IMS architecture, which comprises transport, control and application planes. SIP interface to application can enable 3rd party application providers to offer value added services such as IPTV and CRBT. RTP packets conveying media for these applications would be streamed across transport plane connections. This paper presents the design and implementation of CRBT on IMS networks. It presents considerations for deploying both CRBT and reverse CRBT. The design adopts the architecture where an IMS application server is used to control CRBT service, while the media is stored and served from an RTSP media server. We utilize the Fraunhofer Fokus open source IMS core and UCT IMS client for implementation. Test results are geared to proof of concept; performance tests show minimal added call setup delay of 15 millisecond.
Mapping of terrorist activities in Kenya using sentiment analysis
(Strathmore University, 2017) Ngoge, Lucas .A.; Orero, Joseph Onderi
Terrorism has become a subject of concern to many people in Kenya today. Corruption, porous border and lack of government in the neighboring Somali, have made Kenya a potential target for terrorists’.The advancement in technology has brought a new era in criminal activities where Online Social Networks such as Twitter,Facebook has driven the increase use of the internet by criminal organizations and their supporters for a wide range of purposes including recruitment, financing, propaganda, incitement and gathering and dissemination of information for criminal activities such as threats, incitement to imminent violence, harassing speech, libelous speech etc.Although the Kenya government improved its ability to fight terrorism the changing pattern of terrorist activities, human errors and delayed crime analyses have given criminals more time to destroy evidence and escape arrest.The main objective is to test and validate a technique that can be used to establish crime patterns associated with terrorist activities using sentiment information deduced from twitter data.The data collected will then be used as features for training and development of the algorithm which will then be used to carry out real time mapping of terrorist activities. The algorithm’s performance will be then measured for accuracy.
Techniques for evaluation and selection of partners for construction projects
(Strathmore University, 2017) Musumba, George Wamamu; Nyongesa, Henry O.; Kanyi, Patrick W.; Wario, Ruth D.
Inter-enterprise collaborations require careful evaluations of partner enterprises and their attributes. Evaluation of partners for a project is a multi-criteria decision making process. The project initiator defines multiple criteria to be used in the selection of suitable partners. This study compares three different multicriteria decision making techniques. Analytical Hierarchy Process (AHP) uses pairwise comparisons of crisp numerical values to derive weights of importance of partners. Fuzzy AHP (FAHP) uses pairwise comparisons of fuzzy values to derive weights of importance. Reduced Group Fuzzy AHP (RGFAHP) computes geometric mean of lower and upper bound fuzzy values to derive weights of importance. Eighty persons evaluated five companies to do structural engineering works for a large building. Their evaluation values were subjected to these algorithms. Total mean relative weights of partners were 0.9936, 0.9968 and 0.9866 with errors of 0.0064, 0.0032 and 0.0134 with time complexities of n(n+6), n(n-1)/2 and n(n-1) for AHP, FAHP and RGFAHP respectively. AHP is effective when dealing with crisp evaluation values while FAHP is effective for fuzzy evaluation values. RGFAHP combines fuzzy approximate reasoning with conventional AHP, reduces the number of comparisons when a large number of attributes are used and deals with imprecise evaluators' judgement.
Algorithmic prediction of internet technology utilization in learning
(Strathmore University, 2017) Khakata, Esther Nyokabi; Msanjila, Simon Samuel; Omwenga, Vincent Oteke
Internet technology has been revolutionary over the years especially in the educational sector. However, the utility of internet technology in the learning process of a student in a higher learning institution has not been determined over the years. This has been due to the evolution that has taken place in education. This paper aims at helping in the development of an algorithmic model that will be used for the prediction of internet technology utilization in learning. Specifically, the research will focus on modelling the Cobb- Douglas production theorem to predict the learning output of a given student considering the utility of the internet technology, the infrastructural investment made by the institution of higher learning and the effort of the student. The results of this ongoing research will eventually be of great importance in helping institutions of higher learning determine their returns after investing in internet technology. The students will also be informed on how to use the internet technology in a better way in order to get the best out of the resource.
An Application of association rule learning in recommender systems for e-Commerce and its effect on marketing
(Strathmore University, 2017) Mbugua, Anne W.; Omondi, Allan O.
High annual customer churn rates and low customer attractions caused by poor marketing recommendations inhibit enterprises from making as much profit as they should. The purpose of this research was to derive a more optimized association rule learning algorithm that can be used in a web-based recommender system for small-scale enterprises. The method used was a case study approach on a small-scale enterprise called Makewa Hardware located in Ruiru, Kenya. Having access to the enterprise supported the use of the agile methodology, more specifically, extreme programming in the development of the system that applied the algorithm. A sample of training data consisting of transactions made in the past was obtained from the enterprise in order to create the machine learning aspects of the algorithm. The results howed that the derived association rule learning algorithm was able to learn and generate its own frequent-item-set and use this to give appropriate recommendations to customers. The results revealed the system’s ability to make more accurate recommendations.This was based on the pattern of purchases made from the hardware store by various customers. The recommendations were given on a weekly basis. The implication of the results on the subjects showed that more business owners are open to having intelligent systems help make and predict their sales. The findings can be applied not only in hardware stores but also in other retail stores. Future research can ensure that a normal dataset can be transformed into a market basket without it losing important information.
Embedded system for vehicle speed monitoring
(Strathmore University, 2017) Murakaru, Anne Wamuyu; Orero, Joseph Onderi
This paper investigated the impact of current approaches taken to curb speeding of public service vehicles in Kenya. A qualitative research pointed out that the existing systems are inefficient and ineffective in monitoring speeding and reporting speeding offenses to the relevant authorities. In addition, public service vehicle drivers are not aware of the current speed limit zones in various locations given that the National Transport Service Authority (NTSA) periodically changes speed limit regulations along particular roads. An embedded system for vehicle speed monitoring was proposed and tested. The objective was to design a real time microcontroller based system for mapping speed limit zones and reporting cases of speeding violations to the relevant authorities through an android mobile application. An LCD Screen was integrated to the microcontroller to provide a visual display of the vehicle location and speed limit within the location. In the event of speeding, an audio alert is triggered to notify the driver and an SMS is sent from the GSM module to a central server. Through this system, public service drivers are aware of the speed limit zones on various roads and are alerted once the speed limit is exceeded. The real time reporting system enables the transport agencies and other regulatory bodies to equally monitor speeding vehicles on the roads.
The Secure distributed data exchange protocol
(Strathmore University, 2017) Mabusi, Erick; Shibwabo, Bernard K.
Distributed protocols implementations over a large network is a well-studied problem that converges asymptotically; however, existing protocols do not provide a way for each node to distributively detect the level of trust of another node. In this paper a method is developed to distributively determine whether a certain node should be trusted or not. In absence of such a method all nodes in the network keep communicating and running various computations even a certain node is known to be the origin of unwanted traffic, which is not preferable as in large-scale distributed networks resources like power are limited. Moreover, this additional communication can cause signal interference with other critical information. This distributed data security protocol is expected to take finite time and occurs at each node simultaneously.
Overview of computational intelligence application on prediction of global solar radiation
(Strathmore University, 2017) Fashoto, Stephen Gbenga
Computational Intelligence is not just about robots. It is also about understanding the nature of intelligent thought and action using computers as experimental devices. New applications using computational intelligence are still being developed, although computational intelligence is an established field. The essence of this keynote address is to give a general picture of the research directions which may give an insight into the future of this research area. Meanwhile, an attempt to comprehensively address how computational intelligence may enhance the progress of global solar radiation can be addressed in near future.

Browse

Browsing Pan African Conference on Science, Computing and Telecommunications (PACT) 2017 by Issue Date

Results Per Page

Sort Options