1,721,293 research outputs found
Predicting the Role of Temozolomide Drug in Glioma by Integrating Available Genomic Databases and Computational Methods
No full textCancer is one of the most serious and harmful diseases that threatens humanity. Currently there is no robust treatment which leads to guaranteed cure from cancer. Thus, researchers from various domains are still working hard to identify molecules such as genes and proteins which could be handled and targeted as cancer biomarkers. Various methods have been developed and the research spans wide range of techniques from wet lab testing by biologists to computational methods by computer scientists. The latter research is promising because it greatly reduces the number of molecules as potential biomarkers. This project investigated existing literature data by integrating text mining, as well as gene–gene interactions. Different genes are highlighted in relationship to Glioma and temozolomide
Integrating Data Mining Techniques and Social Networking into Effective Recommendation Framework for Improved Shopping Experience
No full textThe application of data mining in the shopping domain has received a considerable attention for its key role in improving the marketing quality in the last two decades. The main data mining technique that can be used is association rules mining (ARM) though other techniques like clustering and classification are useful but they are beyond the scope of the work described in this thesis. Market basket analysis (MBA) is the most famous example as an application for ARM. MBA’s applications have designed from retail stores’ perspective to gain the benefit. In our thesis, we have designed and implemented a framework that considers the shopping process from consumers’ perspective to turn it into an interactive process, speed it up, save money, and keep the environment clean. Our proposed solution, backed by experimental results, discovers the frequent items that are usually purchased by the consumer; this helps us to introduce them as recommended items. Also, it helps in finding the nearest stores and introduces a navigational map to be used inside the store. Moreover, our proposed solution has been integrated with the social network analysis concept to improve the shopping process quality
A framework for effective web log mining and online prediction
No full textBibliography: p. 71-75Some pages are in colour
Multi-objective genetic algorithms based approach to clustering and its application to microarray data analysis
No full textBibliography: p. 76-8
A Heuristic Stock Portfolio Optimization Approach Based on Data Mining Techniques
No full textPortfolio optimization is the process of making investment decisions on holding a set of financial assets to meet various criteria. A variety of investment assets around the world make this multi-faceted decision problem very complicated. Econometric and statistical models as well as machine learning and data mining techniques have been used by many researchers and analysts to propose heuristic solutions for portfolio optimization.
However, a literature review shows that the existing models are still not practical as they do not always perform better than even the naïve strategy of investing in all available assets in the market. The methodology proposed in this thesis is an alternative heuristic solution to help investors make stock investment decisions through a semi-automated process. The proposed solution is based on the fact that the investment decision cannot be
fully automated because investors’ preferences that are the key factors in making investment decision, vary among different people. For this purpose, a semi-automated framework called SMPOpt (Stock Market Portfolio Optimizer) has been designed and implemented. In the proposed framework, the goal is to learn from the historical
fundamental analysis of companies to discover the optimum portfolio by considering
investors’ preferences. The Portfolio optimization problem is formulated and broken
down into steps to be able to apply data mining techniques such as Clustering and
Ranking, and Social Network Analysis. Some of these techniques are customized based
on the temporal behaviour of financial datasets. For instance, the ranking algorithm based on Support Vector Machine (SVMRank) is modified and a new algorithm called Time-
Series SVMRank is proposed. A comprehensive experimental study has been conducted
using the real stock exchange market datasets from the past recent decades to evaluate the proposed portfolio optimization solution. The obtained results confirmed the strength of
the proposed methodology
Effectiveness of Unique Grouping Techniques for Network Nodes in Serving Various Application Domains
No full textA network is an abstract representation of entities, which can be objects or concepts. Entities are generally represented by nodes, and connected to other entities in the model by links based on their relationship or interaction with the other entities. Networks are a simple but powerful tool for modeling and analyzing relationships between entities, which have become an important technique in many different fields of study. The semantics of the nodes and the links are determined based on the specific domain of study. Nodes in a network could be classified into groups. A group in a network is a subset of the nodes in the network that is being considered together for certain functions. Grouping network nodes refers to a technique of assigning labels to the nodes; grouping techniques are important for building an understanding of the network, and can be used in solving many problems in various domains.
Various techniques have been explored to group network nodes together, such that nodes in each group are highly connected, and nodes between groups have fewer connections. General grouping techniques will discover these high density groups in a wide variety of networks for further examination in numerous fields. The problem with general grouping techniques is that they are multipurpose tools, thus they produce groups of nodes with some characteristics that are commonly sought. Nevertheless, there may be situations that call for discovering groups that have an unusual characteristic. In these problems, a unique grouping technique that is designed specifically to address that particular problem would be a much more effective means to solve the problem. Accordingly, a general framework is proposed in this thesis to help guide the design of unique grouping techniques. This thesis will demonstrate the effectiveness and significance of unique grouping techniques through the development, and application of unique grouping techniques in four distinctive cases. This thesis will show that unique grouping techniques should be a serious consideration alongside general grouping techniques for research work dealing with networks
4D FMRI'dan multistage Alzheimer'in tespiti derin öğrenmeyi kullanan veriler
No full textThe application of machine learning techniques, which significantly improve the recognition of patterns in biomedical data, such as drug delivery systems and medical imaging, has emerged as one of the most important methods for assisting researchers in gaining a better understanding and resolving complex medical issues over the past few years. This has been one of the most significant developments in medical research in recent years. Deep learning is an effective technique for the classifications that extract low-level to high-level features from data. Utilizing a range of machine learning and deep learning algorithms to identify Alzheimer's disease has shown outstanding results. Alzheimer's dementia is progressive, a fatal disorder that turns out to be worse over time; therefore, it is important to diagnose it as early as possible to lessen its impact. To diagnose Alzheimer's disease, deep learning techniques perform significantly better than machine learning techniques by using MRI imaging data. MRI data is even hard to analyze for the physicians. In the literature, two techniques have been used for the identification of Alzheimer's: either by splitting the image into 2D/3D or translating it into functional connectivity or by using the 4D image data after the preprocessing. In this research, the 4D functional MRI data is used for the detection of Alzheimer's after preprocessing. Different preprocessing techniques are applied which include head motion correction, slice timing, slice normalizing, brain extraction, image smoothing, and normalization. The 3-dimensional (CNN) model is implemented and taught on the OASIS data. The transfer learning technique is used on the 3D CNN model and bidirectional long-short-term memory (LSTM) layers are added to understand the temporal information from data. The extended algorithm was named Conv3d-lstm and retrained on the preprocessed ADNI data. Two different datasets are used in this study to generalize the algorithm for the new data. Different 2D CNN models are also trained and tested to assess the performance of the proposed model. Finally, it is concluded that the suggested algorithm provides the finest results comparable to those of other trained algorithms and earlier studies. The algorithm has the highest accuracy and AUC with an AUC of 96% and 91.06% accuracy. The proposed algorithm achieves good results but still, there is space for improvement in the performance.İlaç dağıtım sistemleri ve tıbbi görüntüleme gibi biyomedikal verilerdeki kalıpların tanınmasını önemli ölçüde iyileştiren makine öğrenimi tekniklerinin uygulanması, araştırmacıların karmaşık tıbbi sorunları daha iyi anlamalarına ve çözmelerine yardımcı olmak için en önemli yöntemlerden biri olarak ortaya çıkmıştır. Son birkaç yıl. Bu, son yıllarda tıbbi araştırma alanındaki en önemli gelişmelerden biri olmuştur. Derin öğrenme, verilerden düşük seviyeden üst seviyeye kadar özellikler çıkaran sınıflandırmalar için güçlü bir tekniktir. Alzheimer hastalığını teşhis etmek için bir dizi derin ve makine öğrenimi öğrenme algoritmasının kullanılması olağanüstü sonuçlar göstermiştir. Alzheimer hastalığı, zamanla kötüleşen ilerleyici, ölümcül bir hastalıktır; bu nedenle, hastalığın etkisini azaltmak için mümkün olduğunca erken keşfetmek önemlidir. Alzheimer hastalığını teşhis etmek için derin öğrenme algoritmaları, MRI görüntüleme verilerini kullanan makine öğrenimi algoritmalarından önemli ölçüde daha iyi performans gösterir. MRG verilerini doktorlar için analiz etmek bile zor. Literatürde Alzheimer teşhisi için iki teknik kullanılmıştır: ya görüntüyü 2D/3D'ye bölerek ya da fonksiyonel bağlantıya çevirerek ya da ön işlemeden sonra 4D görüntü verilerini kullanarak. Bu araştırmada, ön işlemeden sonra Alzheimer teşhisi için 4D fonksiyonel MRI verileri kullanılmıştır. Dilim zamanlama, kafa hareketi düzeltme, dilim normalleştirme, beyin çıkarma, yumuşatma ve görüntü normalleştirmeyi içeren farklı ön işleme teknikleri uygulanır. 3D evrişimli sinir ağı (CNN) modeli, OASIS verileri üzerinde uygulanmış ve eğitilmiştir. 3D CNN modelinde transfer öğrenme tekniği kullanılmış ve buna uzun-kısa süreli bellek (LSTM) katmanları eklenerek verilerden zamansal bilgilerin öğrenilmesi sağlanmıştır. Genişletilmiş algoritmaya Conv3d-lstm adı verildi ve önceden işlenmiş ADNI verileri üzerinde yeniden eğitildi. Algoritmayı yeni veriler için genellemek için bu çalışmada iki farklı veri seti kullanılmıştır. Önerilen modelin performansını değerlendirmek için farklı 2D CNN modelleri de eğitilmiş ve test edilmiştir. Son olarak, önerilen modelin diğer eğitilmiş algoritmalar ve daha önceki çalışmalarla karşılaştırılabilir en iyi sonuçları verdiği sonucuna varılmıştır. Algoritma, %96 AUC ve %91.06 doğruluk ile en yüksek doğruluğa ve AUC'ye sahiptir. Önerilen algoritma iyi sonuçlar elde ediyor, ancak yine de performansta iyileştirme için alan var
- …
