2020

Permanent URI for this collection

https://escholar.umt.edu.pk/handle/123456789/15234

Browse

Now showing 1 - 20 of 26

Nested bee hive
(UMT Lahore, 2020) RANA MUHAMMAD NADIR
Several smart city ideas are introduced to manage various problems caused by overpopulation but, the futuristic smart cities is a concept of dense and artificial intelligence (AI) centric cities. Thus, a massive device connectivity with huge data traffic is expected in the future, that can make the situation more complicated for the existing networks. 6G is considered as the problem-solving network of futuristic cities with huge bandwidth and low latency. The expected sixth generation of the Radio Access Network (RAN) is on terahertz (THz) waves with the capability of carrying up-to one terabit per second (Tbps). THz waves have the capability of carrying a large number of data but these waves have some drawbacks like short-range and atmospheric attenuation. Hence, these problems with THz waves can create complications for 6G network. In this article, we have envisioned the futuristic smart cities using 6G and proposed a conceptual Terrestrial Network (TN) architecture for 6G. Moreover, we have designed the multilayer network infrastructure while considering the expectations from network of futuristic smart city and complications of THz waves. Furthermore we have performed simulations of different path finding algorithms in 3D multilayer domain to evaluate and set the dynamics of futuristic communication of 6G.
Performance optimization of 3d games using the profiler feedback
(UMT Lahore, 2020) HUZAIFA SAMI
Performance is the most crucial aspect of a computer game especially that is running on a mobile device. It becomes even more important if the game uses high-end graphics and the game is physical action-oriented. Key performance indicators (KPIs) include the frame rate (i.e. represented as frames per second FPS) and the draw call batches. The aim of performance optimization is to increase the frame rate and decrease the batches. Game performance optimization becomes quite a challenging task on a mobile device because a mobile device has limited resources as compared to the desktop systems. Use of profiler feedback could be helpful in this regard. A profiler is typically attached to a running game and it outputs the performance related data which could be used to identify the performance bottlenecks. This thesis focused on the use of profiler feedback in performance optimization of Unity 3D games on the development platform. Although we used Unity profiler but the methodology is generic and could be used with any gaming engine. We used following ten performance metrics to carry out this research work: (1) Frames per second (2) Batches (3) Drawcalls (4) CPU main (5) GPU main (6) Triangles (7) Vertices (8) Setpass calls (9) Used Textures, and (10) Render Textures. We selected five games from different categories like Action, Simulation, and Role Playing game (RPG) and applied different optimization techniques, mentioned in Chapter Error! Reference source not found. to optimize the performance of these games. After optimization of these games, we observed that the loading time was reduced up to 50%, number of batches reduced up to 95.7%, the CPU consumption decreased up to 72%, the GPU consumption decreased up to 76%, the frame rate increased up to 81.6% (which is five times the original cost), and the size of the game executable (APK) file reduced up to 51%.
Discovery of antiviral phytochemicals for the treatment of coronavirus disease 19 (covid-19)
(UMT Lahore, 2020) MUHAMMAD SOHAIB ROOMI
Corona Virus Disease 19 (Covid-19) is an epidemic infection which is originated from Wuhan, China. Sars-Cov-2 is a virus that spreads this disease inside human body. Scientists made efforts to contain and examine the cause and severity of this Virus so that Antiviral drugs and precautionary measures could be suggested. One way to control this virus is to stop the reproduction and progression of this disease inside the human body. RNA-Dependent-RNA-Polymerase which is known as RdRp is the leading protein of coronavirus which is the main actor in the replication operation and leads the viral infection in the human body. Inhibition of RdRp can stop the replication and as result damage in the human body. To inhibit the working of RdRp, many antiviral drugs and compounds like Hesperidin, Remdesivir, Camostat Mesylate, and Famotidine have been proposed and which are in clinical trials to examine their effects on the large scale. Phytochemicals are natural and less reactive compounds and generally have very few side effects. These compounds can be found in different kinds of plants, fruits, and flavonoids. This study also focuses on phytochemicals and other natural compounds with antiviral properties so that potential compounds can be examined during interaction with the virus. A large number of natural compounds were selected from chemical and plants’ databases and most suitable compounds are selected for further evaluation. Numerous in-silico analysis tools and techniques are followed to carry out the standard procedures which are used worldwide. These techniques consist of Docking, Virtual Screening, Molecular Analysis, and Protein-Ligand Visualization. After following all these steps, combinations of most suitable natural substances were selected via in-silico analysis. A comparative analysis has been carried out between the proposed compounds and compounds which are already under trial. A study has shown that the proposed compounds have strong potential to be considered in place of already under trial compounds
Evaluation of customer behavior using social media (a statistical analysis)
(UMT Lahore, 2020) MUHAMMAD AFAQ
Now a days for the growth of any business we know that customer engagement is very important. So, social media marketing plays very important role for that purpose.it is very important to know that which company growth graph is in positive direction for customer engagement and which company graph is in down direction for customer engagement. The main purpose of this study to analyze that which and how many customers like and dislike the pizza brands on Facebook that is used for social media platform. So, in this study we collected one year data of selected five pizza brands. Using these pizza brand’s customer generated data we can analyze the customer behavior about any pizza brand that how much customers are engaged with any pizza brand. This engagement is analyze base on the customer generated comments and like and dislike action performed by the customer for any pizza brand pages on Facebook. So by this analyses we find that images and videos type content are more attractive than any other written content and.so it is analyze that videos or images type content provide help to engage more customer on Facebook pages. So in this study we analyze how we can measured Facebook usage for any pizza brand and we find that how pizza industry can grow their business using social media (Facebook). So with the help of this study pizza brands can plan their strategies for the marketing purpose.
A novel method to design the substitution box by an affine transformation matrix and aes block cipher
(UMT Lahore, 2020) MIAN MUHAMMAD UMAR SHABAN
Nowadays, data security and privacy have become a major challenge. To secure the information, most of the organizations use data encryption and decryption methods known as ciphers. Modern day ciphers use key-dependent dynamic substitution boxes (S-Boxes) to encrypt and decrypt data. An S-Box is generated from the key to make it dynamic. In this thesis, a novel method of S-Box design is presented. To prove the strength of the proposed S-Box, a critical evaluation is done with standard properties that include Bit Independence Criterion, Nonlinearity, Linear Probability, Differential Uniformity, and Strict Avalanche Criterion. The results are compared with other well-known S-Boxes. The results of the proposed S-Box are quite encouraging and hence can be used in future ciphers.
Learn programming through games at school level in rural areas of pakistan
(UMT Lahore, 2020) MUHMMAD ARSLAN HANIF
Learn programming through games has been an inspiring and challenging mission for the early learners of programming at primary, middle and high school level in rural areas of Pakistan. Due to improper guidance, the early learners of programming feel disappointed and they get less score. If support from their institutes and teachers is provided to students then results will be positive for the novice(early) leaners of programming at school level in Pakistan. Constructive interest is generated in students which will increase their learning ability by proper guidance from their teachers. The use of android mobile phones is increasing in young generation of Pakistan which encourages us to teach them computer programming through games at school level. An inclusive scaffolding framework has been designed for learning programming and programming environment is arranged outside the classroom for novice learners to teach them leaning through computer games. Two experiments have been done to recognize the efficiency of computer programs in smart phones of android and students from GHS for boys Taunsa sharif and GHS for boys mangrotha participated in this experimentation. The evaluation of various parameters was done such as learning ability of students, time required for the completion of task, task access, number of errors and efficiency of students
A novel s-box design technique based on modular inverse and square transformation
(UMT Lahore, 2020) MUHAMMAD AZFAR TARIQ BAIG
Cryptography is an important field for making data confidential. Modern block ciphers in cryptography are the crucial and worthwhile techniques employed to provide secure communication. Permutation and substitution enhance the effectiveness and efficacy of block ciphers. For the purpose of substitution, block ciphers utilize substitution boxes. An S-Box plays an essential role to produce chaos between plaintext as well as cipher text. Many techniques are applied to construct cryptographically robust S-Boxes. The proposed method is based on modular inverse and square transformation for designing more efficient and solid S-Box. S-Box is passed through various criteria such as Nonlinearity, Strict Avalanche Criterion (SAC), Bit Independence Criterion (BIC), Linear Probability (LP), as well as Differential Probability (DP) to analyze the durability of S-Box. Comparison between proposed S-Box and previously designed S-Boxes demonstrates that the proposed S-Box is more viable than compared S-boxes. Consequences collected from criteria suggest that the proposed S-Box can be employed in modern ciphers.
4mc-rf
(UMT Lahore, 2020) FAJAR ARSHAD
N4-methylcytosine 4mC is an essential epigenetic modification that occurs enzymatically by DNA methyltransferase. 4mC sites exist in prokaryotes and play a vital role in regulating gene expression, DNA replication, and cell cycle. The efficient and accurate prediction of 4mC sites has a significant role in the insight of 4mC biological properties and functions. Therefore, we have proposed a sequence-based predictor, namely 4mC-RF, for identifying 4mC sites in prokaryotes by integrating statistical moments along with position and composition dependent features. Relative and absolute position based features are computed to extract the optimal features. A popular machine learning classifier Random Forest was used to training the model. Validation results were obtained under rigorous processes of Self-consistency, 10-fold crossvalidation, Independent testing, and Jackknife testing giving 95.01%, 95.02%, 97.02%, and 95.36% accuracies. Our proposed model depicts the highest prediction accuracies as compared with the literature results. Thus, the developed 4mC-RF model was constructed into a web server. A significant and more accurate predictor of 4mC Methylcytosine sites helps experimental scientists gather results moderately.
Iptsw (3l)
(UMT Lahore, 2020) MAHA ASHRAF
The promoter sequence in genomic data are comprising of 81-1000base pairs, exists in the upstream part of the genic transcriptional start site. It modulates the transcription mechanism of many genes in association with a variety of transcriptional factors. With the revolution in the advanced-genomic era, it is essential to statistically classify the promoters. Such classification may help in drug development. Some forecasting methods were established. Most of them were restricted to the pure recognition of a DNA query sequence. However, based on their differing levels of strength for transcriptional activation, the advertising expression can be split into three layers: promoter vs non-promoter, types of promotors, and their strengths. A modern three-layer predictor, named "iPTSW (3L)" has been established by integrating the information of nucleotide abundance and physiochemical properties into position and composition-dependent features. Its first layer decides whether the query DNA sequence is a valid promoter, the second layer classifies the type of a promoter, while its third layer determines the strength of the promoter. A model is trained by using the Random forest technique to learn the pattern and sequence of the data for prediction. An accuracy of 97.57% for 10-Fold cross-validation, 99.8% for Jackknife validation, and 89.0% for Independent Testing was achieved. These results indicate that the proposed methodology plays a vital role in prediction instead of performing all tests in the laboratory utilizing conventional ways and also cost Efficient. This research can also help in solving relevant prediction problems. The web server http://biopred.org/promotersis developed for Sequence Prediction.
Motivations, challenges and recommendations for using machine learning in drug discovery and development
(UMT Lahore, 2020) ADIL PERVAIZ
Emergence of computational methods and technologies has produced a society that is fed by the data. There are tremendous amounts of data which should be sorted and characterized in order to avail the useful information. With the advancement in computational technologies especially the Machine Learning, new doors to data manipulation have been opened. Beside other domains, this has led to the change in traditional drug development process. Traditionally, drug development was target driven, whereas modern drug development process is data driven. At present, almost all the pharmaceutical companies are using Machine Learning techniques to develop new and novel drugs. Our work focuses on the review of Machine Learning techniques used in drug discovery and development and illustrates how they have changed the face of drug discovery in the last decade. The study is done by choosing more than 100 articles that meet our selection criteria, from six famous databases: Springer, IEEE, Elsevier, ChemInformatics, BMC and ACS. The information gained from these chosen articles is sorted under various headings and the main findings are presented pictorially. Since a long time, analysts have been following the pattern of Machine Learning (ML) and other computational technologies in drug disclosure in different ways, yet leaving territories or a corner for further consideration. We can conclude that a strong collaborative relationship between computer and biological scientists could boost the process of discovery and development of novel and improved drugs to cure various diseases.
Itransaminase- pseaac
(UMT Lahore, 2020) Mahreen Zainab
The major focus of this thesis is to identification of transaminase and non-transaminase through dataset of 3077 positive and 2500 negative sequences which was collected from uniport. We review new spectrophotometric determination methods to determine the transaminase activity. The main diseases which were diagnosis through transaminase levels in bloodstream are liver and heart it is very important to diagnose them in preliminary stage for better treatment. In present scenario to detect infection of liver devise like sensors were used or blood sampling in laboratory. In this context, this research utilizes neural network approaches for classification of the diseases due to transaminase in patients. We majorly discuss algorithm (SVM) and their results for the identification of transaminase and non-transaminase. Further to improve the accuracy of identification a new model was developed i-transaminase using Random forest Algorithm. This model was tested for its performance using feature extraction components like Statistical Moments Calculation, PRIM, RPRIM, AAPIV and RAAPIV. The new model resulted in prediction accuracy of 99.9% using testing like self-consistency, jackknife, cross validation. The accuracies for above mention testing is 99.9%, 99.93% and 99.87. The results ensure that development of this model improved the accuracy of identification. To serve the medicine community for identification of transaminase in liver and heart patients, a user interface also developed using Python. This GUI is deployed as a package in local repository or also on webserver.
Framework for security of health monitoring applications
(UMT Lahore, 2020) MUHAMMAD SAMI
Mobile phones have given application designers the chance of bringing actual period correspondence among doctors and patients. The proposed framework of Health monitoring apps comprises essentially of utilizations of software and work at the degree of equipment. This framework is for the most part used to sift through interminable illnesses, screen the senior and eager moms, and help the patients. iOS and the Google play stores are the two significant conveyance stages that underpin the Android and iPhone activity framework. The proposed framework of health monitoring apps helps to screen and record a patient's standard activities. The treatment includes the fundamental protection issues concerning various information of patients, specialists, and guardians. The security in the proposed framework of health monitoring apps is fundamental as their information is traded with remote systems. The proposed framework of Health monitoring apps that are likewise open in the online circulation stage which are called Android. Android stage deals with all the delicate information that is helpful for the expert (specialist, doctor, and advisor). The security framework of health monitoring apps structure configuration comprises two layers: A Monitoring Layer (ML) which is utilized to framework the modules of security-checks, an interface layer (IL) which givesinterface among (ML), and working framework Android. The approval of the proposed framework is apparent through the usage of real Android gadgets; this execution shows the system for all intents and purposes and assesses its presentation. The significance of tests indicated that they effectively secure the health care app framework against various assaults through the proposed structure.
Cloud computing security issues and services using enhanced virtualized Platform" ”
(UMT Lahore, 2020) HUSNAIN ZAHEER
By the passage of time computer technology has become more useful and of low cost . The change in computer technology has given rise to a new technology called cloud computing, computer resources have made data storage more easier using softwares and hardware tools. This new tech has been divided into two parts the inside view of system who control cloud systems and user friendly data keeping environment, customers are served in a inside database environment. This new technology has gave good attractive response to the industry .we use software to offer healthy support through internet. The cloud involves both I/o and outside software support. We openely and freely to all people we name it as public cloud; the result in response to user query in cloud is software computing. Examples of public software computing consist of all net services. We use cloud working session for inside working not open to all. The reverse order result will require a new platform for data protection. This studies will growth the service support for every layer activity and could maintain backup of all statistics offering a virtualized surroundings for every layer the statistics for every person is first being despatched to disbursed offerings introduction platform which passes it via assuarance platform which in addition transfers it to digital aid layer at bottom computer community storage is used then at last transport platform is signaled and message is going to cease person.
Early prediction of skin cancer using deep learning
(UMT Lahore, 2020) MUHAMMAD ATIF
The skin cancer is considered amongst one of the most fatal diseases, with high morbidity. By 2020, it is estimated that more than 100,000 people have been diagnosed by skin cancer and around 7000 mortalities have been recorded, in United States itself. Majorly, the skin cancers are divided in three main types which are basal cell carcinoma (BCC), squamous cell carcinoma (SCC), and melanoma. In the most cases of the skin cancers, these are squamous cell carcinoma and basal cell carcinoma. Although these are malignant, these are less likely to appear on the other parts of the body if treated at early stages. If not treated early, they can deform locally. The first step in diagnosing a malignant lesion by a dermatologist is a visual examination of the suspicious area of the skin. Accurate diagnosis is important because of the similarity of some lesion types. Manual visualization and inspection through the eye could be hectic and troublesome when the number of samples is large. Therefore, an automated classifier is always required to perform such tasks. As compared to image processing approaches and conventional machine learning, deep learning is more efficient and accurate. Thus, an automated deep learning-based classifier is required for skin cancer identification using images. In the present study, the early prediction of skin cancer has been aimed. Data for 9 skin cancer types have been collected which are basal cell carcinoma, dermatofibroma, melanoma, nevus, vascular lesion, actinic keratosis, seborrheic keratosis, squamous cell carcinoma, and pigmented benign keratosis. Based on this data, a deep learning residual neural network algorithm is incorporated for extensive and deep feature extraction from the images, and their classification. A trained prediction model for skin cancer classification has been developed and validated using v cross-validation approaches. Thus, an accuracy of 97.45% has been achieved which prove that the proposed predictor surpasses the existing models and can be served as a time and cost-effective stratagem for identifying skin cancer and its type.
Numerical modeling and simulation of engineering problems
(UMT Lahore, 2020) AHSAN RAZA
This work is presented in the study of boundary value problem for thin film flow of fourth grade fluid down a vertical cylinder. The practical usage of the exact flow is restricted as it involves very complicated integrals. The nonlinear problem that arises is solved by Galerkin’s finite element approach based on weighted-residual formulation which is used to find the approximate solutions of the fourth-grade problem. This approach utilizes a piecewise linear approximation using Linear Lagrange polynomials. Convergence of the solutions which briefly describes the flow characteristics to include the effects of the emerging parameters. The results obtained after implementation are not restrictive to small values of flow parameters. Numerical studies have shown the superior accuracy and lesser computational cost of this scheme in comparison to collocation, homotopy analysis method and homotopy perturbation method. The impact of the relevant parameters is examined through graphical results after implementing a number of iterations.
A cloud based multi-agent system for traffic exigency services
(UMT Lahore, 2020) MUHAMMAD SALMAN ZAHEER
A multi-agent system can work on different tasks in a distributed environment to achieve single global goal. These systems can be integrated with Cloud computing to get better performance. A multi-agent system can perform more efficiently by utilizing cloud storage resources and computing power. Until now, a lot of research work has been done on traffic signal automation by using multi-agent systems. This research presents a new architecture by utilizing the power of multi-agent system and cloud computing to handle the exigency services in the urban transportation system. Exigency services are the special cases in the transportation system, which may include an emergency ambulance movement, police movement, Fire Brigade, and other exigent situations. In the proposed architecture, agents are deployed at each signal node, which collect information at each signal and transmit it to the cloud services. These agents are also able to handle the signal nodes in case of emergency. Agents at different nodes share information about exigency. Cloud service plays the backbone role with multi-agent while handling exigency. It analyses the input data and determines the presence of an emergency service. We have defined a priority mechanism to handle multiple emergency services at a particular crossroad or chowk. We also have presented different case studies to show the utility of the proposed architecture. It gives a better understanding of exigency service handling in an automated urban transportation system.
Prediction of rna-binding proteins with respect to their domains.
(UMT Lahore, 2020) ARJMAND MAJEED
Due to innumerability and diversity in biological processes, many cellular physiological processes depend on the decisive factors of interaction between RNA and protein. RNA-binding proteins (RBPs) habitually changes the functionality of RNA by making abound with one or more globular domains of RNA-binding with incredibly diverse structure and mechanisms. High productiveness of biological experiments come up with enough essential information, to initially identify the RNAprotein interaction, however, RPI networks complexities are so time-consuming and expensive. Hence, a reliable and high-speed prediction method becomes necessary. In this proposed work, we put forward a computation method that uses sequential information for the prediction of RNAprotein interactions and their domains like K- homology, RNA Recognition Motif, Ribosomal protein S! -like, THUMP domain, C2h2 Zinc Finger/CCCH zinc finger and PUA through which they interact. Position relative statistical moments were calculated for the training of neural networks. With the use of 10-fold cross-validation and jackknife testing accuracies of 94.97% and 96% have been achieved respectively whereas the accuracy of the overall system is 94.4% with a sensitivity value of 94%. The obtained result shows a high potential to play an important role among the existing method.
Emotion recognition in computer vision by using machine learning technique in static images
(UMT Lahore, 2020) Amna Arshad
An immense area of technology which is covering many area and huge part of discussion is a computer vision this area cover many techniques and many AI facts so that this area is a widely covered and used by different authors and many working areas. Computer vision is based on many techniques which has deep learning techniques which hits many areas so that computer vision area is also known as emotions recognition. Emotions recognition is a wide area of technology which is based on behavioral changes, environmental changes. Emotions recognition from static images is a bit challenging task. The modalities include in this task are static images which are having different type of emotions as emotions are intense part of behavior because these are nonverbal part of communication and this part has a key role to describing the behavioral changes, these changes are effect on a person by any situation. In recent few years when machine learning and AI become popular and authors has work on it than emotions recognition become very continuous discussion part of technology because these are covered by many areas of technology so that emotions in static images are very important in many facts and these facts are based on behavioral changes and these changes are the factors of emotions. In this study there are seven main and basic emotions type will be discussed which are (Natural, Happy, Sad, Surprise, Fear, Anger and Disgust) these emotions types are basic types which are also discussed by many authors although emotions are very vast area of image recognitions because these are recognized by the face expressions which tends to be emotions. So that these are behavioral phenomena which are based on nature environment and situations which effects on human. According to the artificial intelligence emotional study is based on different techniques and areas which are interlined with each other whereas those area which are connected are neural network iv which are used to describes specificity, image processing area which indicates the different area of images and emotional static images are one of them so that other part which is that part of techniques and that is based on the machine learning area of techniques which has an important part to complete and describes all that parts in a legitimate way. So that to describe accuracy rate in images in this study here are some techniques which can describes the clearly the emotional states. Techniques which are used are show that the accuracy of emotional states of images which is 98%. So, that Here this study is based on the emotions so that CNN (Convolutional Neural Network) has a major role to describes different part of emotions in static images those static images are taken from the well-known data bases ck+ (Cohan-Kanade) and JAFFE (Japanese female facial expression) both of these data bases are based on static images which are having seven different type of emotions. JAFFE and CK+ has emotional state images whereas these are using for describing the accuracy rate in the images.
A survey of prediction method using feature selection in genomics
(UMT Lahore, 2020) ATIF ANWAR MIRZA
Microarrays have enabled the scientists to study expression of hundreds and thousands of genes in a single experiment with very low number samples and biological replicates, most of microarray gene expression studies are used to classify human cancer using expression levels of genes in cancer cells compared with normal cells. The data scientists have huge amount of data submitted in the repositories for analysis and development of algorithms for efficient classification and selection of genes. This huge amount of data with fewer number of biological samples has huge dimensionality in the data due to a number of biological and technical reason. Feature selection in this scenario can not only reduce the dimensionality of data but it can also reduce the number of features to be analyzed without affecting the prediction capability, thus reduce the computational load and save time being used for data analysis. Two data sets having samples from three leukemia subclasses acute lymphoblastic leukemia from B and T cell and acute myeloid leukemia. While training data set comprised of 30 samples (10 acute lymphoblastic leukemia -B, 10 acute lymphoblastic leukemia –T and 10 acute myeloid leukemia). The test data set consisted of 38 samples (19 acute lymphoblastic leukemia -B, 8 acute lymphoblastic leukemia –T and 11 acute myeloid leukemia). The metagene factors were extracted and evaluated by the method described by Tamayo et al., 2007. Feature selection is a process where you automatically select those features in your data that contribute most to the prediction variable or output in which you are interested. Feature selection algorithms like Information Gain, Gain ratio, Gini Index, ReliefF, The Fast Correlation-Based Filter, Analysis of Variance (ANOVA) and Chi2 were used. For testing of hypothesis that feature selection may improve prediction efficiency of metagene classification the method proposed by Tamayo et al., 2007 was modified with hybrid approach. The feature selection was performed in python using orange 3.25 GUI and the selected data were analyzed using metagene projection model. Feature selection algorithms application was planned with for types of data selection using each algorithm i.e., 100%, 75%, 50% and 25% of total data were selected for metagene prediction using each of the seven feature selection methods. The feature selection method was applied on original data having 5571 features/genes in python using Orange 3.25 graphic user interface and 100% features were selected i.e., 5571 features were analyzed for metagene prediction SVM algorithm in R graphic user interface and the results are shown in bio plots of model and test samples, hierarchical clustering of original data and projected data and heat maps of original and projected data were developed. To evaluate the prediction ability of the model with special reference to data filtering using various feature selection categories two parameters were calculated i.e., brier score and error percentages were calculated. ReliefF was proved to be the best method for feature selection based on lowest brier score and lowest error percentage.
Prediction model developed through rf, ann and svm classifier for angiogenesis &tumor angiogenesis process
(UMT Lahore, 2020) RABIA KHAN
A crucial biological process called Angiogenesis plays a vital role in migration, growth, and wound healing of endothelial cells and other processes that are controlled by chemical signals. The balancing of these signals is necessary for the proper working of angiogenesis. Unbalancing of these signals increase blood vessel formation, which causes abnormal growth or several diseases including cancer. The proposed work focuses on developing a two-layered prediction model using different classifiers like Random Forest, Neural Network, and Support Vector Machine. The performance of the model is evaluated through various validation techniques. The first layer of the model predicts whether the given primary structure corresponds to angiogenesis proteins or not. If it is determined as an angiogenesis protein then the second layer of the model will classify whether or not the protein plays a role in tumor angiogenesis. The model was trained on RF, ANN, and SVM and it was evaluated by using k-fold crossvalidation, independent, self-consistency, and jackknife testing. The overall accuracy of RF for angiogenesis is 99.3% and tumor angiogenesis is 99.7%, ANN showed 96.23% accuracy for angiogenesis and 78.65% for tumor angiogenesis, and the accuracy of SVM for angiogenesis is 78.65% and for tumor angiogenesis is 65.19%.

Browse

Browsing 2020 by Issue Date

Results Per Page

Sort Options