| 1. |
- Javeed, Ashir, 1989-, et al.
(författare)
-
Early Prediction of Dementia Using Feature Extraction Battery (FEB) and Optimized Support Vector Machine (SVM) for Classification
- 2023
-
Ingår i: Biomedicines. - : MDPI. - 2227-9059. ; 11:2
-
Tidskriftsartikel (refereegranskat)abstract
- Dementia is a cognitive disorder that mainly targets older adults. At present, dementia has no cure or prevention available. Scientists found that dementia symptoms might emerge as early as ten years before the onset of real disease. As a result, machine learning (ML) scientists developed various techniques for the early prediction of dementia using dementia symptoms. However, these methods have fundamental limitations, such as low accuracy and bias in machine learning (ML) models. To resolve the issue of bias in the proposed ML model, we deployed the adaptive synthetic sampling (ADASYN) technique, and to improve accuracy, we have proposed novel feature extraction techniques, namely, feature extraction battery (FEB) and optimized support vector machine (SVM) using radical basis function (rbf) for the classification of the disease. The hyperparameters of SVM are calibrated by employing the grid search approach. It is evident from the experimental results that the newly pr oposed model (FEB-SVM) improves the dementia prediction accuracy of the conventional SVM by 6%. The proposed model (FEB-SVM) obtained 98.28% accuracy on training data and a testing accuracy of 93.92%. Along with accuracy, the proposed model obtained a precision of 91.80%, recall of 86.59, F1-score of 89.12%, and Matthew’s correlation coefficient (MCC) of 0.4987. Moreover, the newly proposed model (FEB-SVM) outperforms the 12 state-of-the-art ML models that the researchers have recently presented for dementia prediction.
|
|
| 2. |
- Kerren, Andreas, 1971-, et al.
(författare)
-
CluMa-GO: Bring Gene Ontologies and Hierarchical Clusterings Together
- 2011
-
Annan publikation (refereegranskat)abstract
- Ontologies and hierarchical clustering are both important tools in biology and medicine to study high-throughput data such as transcriptomics and metabolomics data. Enrichment of ontology terms in the data is used to identify statistically overrepresented ontology terms, giving insight into relevant biological processes or functional modules. Hierarchical clustering is a standard method to analyze and visualize data to find relatively homogeneous clusters of experimental data points. Both methods support the analysis of the same data set, but are usually considered independently. However, often a combined view is desired: visualizing a large data set in the context of an ontology under consideration of a clustering of the data. This paper proposes a new visualization method for this task.
|
|
| 3. |
- Jusufi, Ilir, 1983-, et al.
(författare)
-
Visualization of Mappings between the Gene Ontology and Cluster Trees
- 2012
-
Ingår i: <em>Proceedings of the SPIE 2012 Conference on Visualization and Data Analysis (VDA '12)</em>. - : SPIE - International Society for Optical Engineering. - 9780819489418
-
Konferensbidrag (refereegranskat)abstract
- Ontologies and hierarchical clustering are both important tools in biology and medicine to study high-throughput data such as transcriptomics and metabolomics data. Enrichment of ontology terms in the data is used to identify statistically overrepresented ontology terms, giving insight into relevant biological processes or functional modules. Hierarchical clustering is a standard method to analyze and visualize data to find relatively homogeneous clusters of experimental data points. Both methods support the analysis of the same data set, but are usually considered independently. However, often a combined view is desired: visualizing a large data set in the context of an ontology under consideration of a clustering of the data. This paper proposes a new visualization method for this task.
|
|
| 4. |
- Javeed, Ashir, 1989-, et al.
(författare)
-
An Intelligent Learning System for Unbiased Prediction of Dementia Based on Autoencoder and Adaboost Ensemble Learning
- 2022
-
Ingår i: Life. - : MDPI. - 2075-1729. ; 12:7, s. 1-18
-
Tidskriftsartikel (refereegranskat)abstract
- Dementia is a neurological condition that primarily affects older adults and there is stillno cure or therapy available to cure it. The symptoms of dementia can appear as early as 10 yearsbefore the beginning of actual diagnosed dementia. Hence, machine learning (ML) researchershave presented several methods for early detection of dementia based on symptoms. However,these techniques suffer from two major flaws. The first issue is the bias of ML models caused byimbalanced classes in the dataset. Past research did not address this issue well and did not takepreventative precautions. Different ML models were developed to illustrate this bias. To alleviate theproblem of bias, we deployed a synthetic minority oversampling technique (SMOTE) to balance thetraining process of the proposed ML model. The second issue is the poor classification accuracy ofML models, which leads to a limited clinical significance. To improve dementia prediction accuracy,we proposed an intelligent learning system that is a hybrid of an autoencoder and adaptive boostmodel. The autoencoder is used to extract relevant features from the feature space and the Adaboostmodel is deployed for the classification of dementia by using an extracted subset of features. Thehyperparameters of the Adaboost model are fine-tuned using a grid search algorithm. Experimentalfindings reveal that the suggested learning system outperforms eleven similar systems which wereproposed in the literature. Furthermore, it was also observed that the proposed learning systemimproves the strength of the conventional Adaboost model by 9.8% and reduces its time complexity.Lastly, the proposed learning system achieved classification accuracy of 90.23%, sensitivity of 98.00%and specificity of 96.65%.
|
|
| 5. |
- Jusufi, Ilir
(författare)
-
Towards the Visualization of Multivariate Biochemical Networks
- 2012
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Many open challenges exist when dealing with different biological networks. They are crucial for the understanding of living beings. Complete drawings of these typically large networks usually suffer from clutter and visual overload. In order to overcome this issue, the networks are divided into single, hierarchically structured pathways. However, this subdivision makes it harder to navigate and understand the connections between pathways. Another challenge is to visualize ontologies and hierarchical clusterings, which are important tools to study high-throughput data that are automatically generated nowadays. Both of these methods produce different types of large graphs. Although these methods are used to explore the same data set, they are usually considered independently. Therefore, a combined view showing the results of both methods is desired. Additionally, real life data sets, including biological networks, usually have additional attributes related to the considered network. Investigating means to visualize such multivariate data together with the network drawing is also one of the ongoing challenges in biology, but also in other fields.The aim of this thesis is to lay out the foundations towards defining techniques for the visualization of multivariate biochemical networks. An overall understanding of the problems related to biochemical networks should be acquired to achieve this aim. More importantly, a contribution to the aforementioned challenges is necessary.Two research goals have been defined to accomplish our aim: for the first goal, we should improve shortcomings of the approach of dividing larger biological networks into smaller pieces and contribute to the problem of a visualization of different types of interconnected biological networks. The second goal is a contribution for the visualization of multivariate biological networks.Initially, a brief survey on techniques to visualize multivariate networks is presented in this thesis. Then, various visualization and interaction techniques are presented that address the challenges in biochemical network analysis. Three different software tools were implemented to demonstrate our research efforts. We discuss all features of our systems in detail, describe the visualization and interaction techniques as well as disadvantages and scalability issues if present.
|
|
| 6. |
- Olsson, Anki, et al.
(författare)
-
Better platelet function, less fibrinolysis and less hemolysis in re-transfused residual pump blood with the Ringer’s chase technique : a randomized pilot study
- 2018
-
Ingår i: Perfusion. - : Sage Publications. - 0267-6591 .- 1477-111X. ; 33:3, s. 185-193
-
Tidskriftsartikel (refereegranskat)abstract
- Introduction: Residual pump blood from the cardiopulmonary bypass (CPB) circuit is often collected into an infusion bag (IB) and re-transfused. An alternative is to chase the residual blood into the circulation through the arterial cannula with Ringer’s acetate. Our aim was to assess possible differences in hemostatic blood quality between these two techniques.Methods: Forty adult patients undergoing elective coronary artery bypass graft surgery with CPB were randomized to receive the residual pump blood by either an IB or through the Ringer’s chase (RC) technique. Platelet activation and function (impedance aggregometry), coagulation and hemolysis variables were assessed in the re-transfused blood and in the patients before, during and after surgery. Results are presented as median (25-75 quartiles).Results: Total hemoglobin and platelet levels in the re-transfused blood were comparable with the two methods, as were soluble platelet activation markers P-selectin and soluble glycoprotein VI (GPVI). Platelet aggregation (U) in the IB blood was significantly lower compared to the RC blood, with the agonists adenosine diphosphate (ADP) 24 (10-32) vs 46 (33-65), p<0.01, thrombin receptor activating peptide (TRAP) 50 (29-73) vs 69 (51-92), p=0.04 and collagen 24 (17-28) vs 34 (26-59), p<0.01. The IB blood had higher amounts of free hemoglobin (mg/L) (1086 (891-1717) vs 591(517-646), p<0.01) and D-dimer 0.60 (0.33-0.98) vs 0.3 (0.3-0.48), p<0.01. Other coagulation variables showed no difference between the groups. Conclusions: The handling of blood after CPB increases hemolysis, impairs platelet function and activates coagulation and fibrinolysis. The RC technique preserved the blood better than the commonly used IB technique.
|
|
| 7. |
- Shamshad, Hasib, et al.
(författare)
-
Forecasting and Trading of the Stable Cryptocurrencies With Machine Learning and Deep Learning Algorithms for Market Conditions
- 2023
-
Ingår i: IEEE Access. - : Institute of Electrical and Electronics Engineers (IEEE). - 2169-3536. ; 11, s. 122205-122220
-
Tidskriftsartikel (refereegranskat)abstract
- The digital market trend is rapidly expanding due to key characteristics like decentralization, accessibility, and market diversity enabled by blockchain technology. This study proposes a Predictive Analytics System to provide simplified reporting for the three most popular cryptocurrencies with varying digits, namely ADA Cardano, Ethereum, and Binance coin, for ten days to contribute to this emerging technology. Thus, this proposed system employs a data science-based framework and six highly advanced data-driven Machine learning and Deep learning algorithms: Support Vector Regressor, Auto-Regressive Integrated Moving Average (ARIMA), Facebook Prophet, Unidirectional LSTM, Bidirectional LSTM, Stacked LSTM. Moreover, the research experiments are repeated several times to achieve the best results by employing hyperparameter tuning of each algorithm. This involves selecting an appropriate kernel and suitable data normalization technique for SVR, determining ARIMA's (p, d, q) values, and optimizing the loss function values, number of neurons, hidden layers, and epochs in LSTM models. For the model validation, we utilize widely used evaluation techniques: Mean Absolute Error, Root Mean Squared Error, Mean Absolute Percentage Error, and R-squared. Results demonstrate that ARIMA outperforms the other models in all cases, accurately projecting the price variability within the actual price range. Conversely, Facebook Prophet exhibits good performance to some extent. The paper suggests that the ARIMA technique offers practical implications for market analysts, enabling them to make well-informed decisions based on accurate price projections. © 2013 IEEE.
|
|
| 8. |
|
|
| 9. |
- Rakus-Andersson, Elisabeth
(författare)
-
Hybridization of Immunological Computation and Fuzzy Systems in Surgery Decision Making
- 2011
-
Konferensbidrag (refereegranskat)abstract
- From the domain of Computational Intelligence we have selected immunological computation and fuzzy systems to combine them in a new hybrid model. This novel numerical method has been tested on patient data strings to make decisions about the choices of surgery types. The model input clinical data concerns patients who suffer from gastric cancer.
|
|
| 10. |
- Mhathesh, T. S. R., et al.
(författare)
-
A 3d convolutional neural network for bacterial image classification
- 2021
-
Ingår i: Advances in Intelligent Systems and Computing. - Singapore : Springer. - 9789811552847 ; , s. 419-431
-
Konferensbidrag (refereegranskat)abstract
- Identification and analysis of biological microscopy images need high focus and years of experience to master the art. The rise of deep neural networks enables analyst to achieve the desired results with reduced time and cost. Light sheet fluorescence microscopies are one of the types of 3D microcopy images. Processing microscopy images is tedious process as it consists of low-level features. It is necessary to use proper image processing techniques to extract the low-level features of the biological microscopy images. Deep neural networks (DNN) are efficient in extracting the features of images and able to classify with high accuracy. Convolutional neural networks (CNN) are one of the types of neural networks that can provide promising results with less error rates. The ability of CNN to extract the low-level features of images makes it popular for image classification. In this paper, a CNN-based 3D bacterial image classification is proposed. 3D images contain more in-depth features than 2D images. The proposed CNN model is trained on 3D light sheet fluorescence microscopy images of larval zebrafish. The proposed CNN model classifies the bacterial and non-bacterial images effectively. Intense experimental analyses are carried out to find the optimal complexity and to get better classification accuracy. The proposed model provides better results than human comprehension and other traditional machine learning approaches like random forest, support vector classifier, etc. The details of network architecture, regularization, and hyperparameter optimization techniques are also presented. © Springer Nature Singapore Pte Ltd 2021.
|
|
