| 摘要: | 深度學習是許多現代AI人工智慧（Artificial Intelligence）應用的基礎。自從在語音識別和圖像識別領域中展現出突破性的成果後，深度學習在其他領域的應用便以極快的速度成長。而在生物醫學領域也陸續可以看到其應用，像是癌症檢測，生物信息分析等等，在生物學中的衰老研究也有相當大的貢獻，在此論文中，使用了來自The Genotype-Tissue Expression(GTEx)的基因型組織表達，針對DNA去做次世代定序而得出的RNA-seq資料，因具有檢測速度快，通量(throughput)高且檢測的範圍廣的特點，所以在偵測基因表現量上有較佳的正確性。
在本篇論文裡，主要方向為三：
1. 分別對肺組織、肝組織、小腦組織、心臟組織以及血液所屬供體的年齡層進行分類預測
2. 實驗深度神經網路下在不同激發函數以及損失函數中比較結果 3. 透過統計分析的方法提取各組織關聯基因集以探討老化因子
在本篇論文裡，會使用像是嶺回歸(Ridge Regression)，決策樹(Decision Tree)，隨機森林(Random Forest)以及支援向量機(Support Vector Machine)等部分機器學習之方法來實驗，同時為了比較各個方法的辨識率，也加入深度神經網路(Deep Neural Network)、自編碼器(Auto-encoder)等深度學習的方法，最後利用統計分析的方法探討各組織間存在的衰老相關潛在因子。
;Deep learning is the foundation of AI Artificial Intelligence applications.
Since the achievement in the field of speech recognition and image recognition, DNN has grown with extremely fast rate in other fields. For biomedicine, the application of deep learning methods, such as cancer detection, bioinformatics analysis, etc., has also been widely used,
and aging research has also made significant contributions . In this paper,
The genotype tissue from The Genotype-Tissue Expression (GTEx) expresses RNA-seq data for DNA sequencing, with high detection speed, high throughput and wide range of detection. Characteristics, so there is better correctness in detecting gene expression.
In this paper, we have tree main directions:
1.Classification and prediction of age groups from normal tissues
2.Compare the results between activation function and loss functions
3.Extracting related gene sets of various tissues by statistical analysis
In this paper, we will use machine learning for experiment such like Ridge Regression , Decision Tree , Random Forest, and Support Vector Machine . In order to compare the recognition rates of each method,we also added deep neural network , auto-encoder , and other methods of deep learning. |
