# 醫學資訊概論4.Big Data Analytics and AI for Decision Making in BI [簡報在這裡](https://elearn2.fju.edu.tw/course/234105/learning-activity#/1242105) ## 小考題目及解答: - 選擇題 1. Which big data issue is to exam (test) large volumes of data from a variety of types and sourcess to uncover hidden patterns, correlation, trends,and so on? - [ ] big data management - [ ] big data control - [ ] big data gathering - [x] big data analytics 2. Which big data issue focuses on the organization, administration and governance of large volumes of both structured and unstructured data? - [x] big data management - [ ] big data analytics - [ ] big data cleaning - [ ] big data computing 3. Which are the major tasks of Map Phase in MapReduce operation? - [x] spilts & mapping - [ ] shuffling & reducing - [ ] mapping & reducing 4. Which are the major tasks of Reduce Phase in MapReduce operation? - [ ] spilts & mapping - [x] shuffling & reducing - [ ] mapping & reducing 5. Which does not a feature of big data? - [ ] variety - [ ] velocity - [x] volatile - [ ] veracity 6. Which tools are used to scan massive data for discovering meaningful new correlations, patterns, and trends? - [ ] visualization tools - [x] data mining tools - [ ] virtual reality tools 7. 以下哪一種人工智慧演算法具深度學習的特性? - [ ] 數學演算法 - [x] 類神經網路 - [ ] 專家系統 - [ ] 決策樹 8. 以下哪一種非深度學習的主要特徵? - [ ] 大量資料學習 - [ ] 嚴謹的學習演算法 - [ ] 需要快速運算能力 - [x] 需要快速繪圖能力 9. 哪一項不是應用類神經網路建立醫療決策系統的必要元素? - [ ] 歷史性案例 - [ ] 深度學習機制 - [x] 推論引擎 - [ ] 系統建立相關人員 10. 哪一種人工智慧技術具推論及解釋功能? - [ ] 決策樹 - [x] 專家系統 - [ ] 基因演算法 - [ ] 倒傳遞類神經網路 11. 哪一種人工智慧技術模仿生物競存法則而建立? - [ ] 機器人 - [ ] 專家系統 - [x] 遺傳演算法 - [ ] 到傳遞類神經網路 12. 哪一Google 雲端 BigTable (Google File System)存取架構之描述是不正確的? - [x] 他是一種關聯是SQL database - [ ] 較底層包含a set of distributed chunkservers - [ ] 儲存在chunkserver 的資料單元稱 "chunk" - [ ] 有一Master Process 維護metadata 13. 下列何者不是大數據分析常用到的相關技術? - [ ] Internet of Things - [ ] Cloud computing - [ ] High-speed data communication such as 5G - [x] File organization and management 14. 下列何者不是成為一位Data Scientist應具備的技能? - [ ] Data modeling and using AI tools for data analytics - [x] Natural science experiment - [ ] System deployment and data analytics on cloud platform such as Ammazon AWS - [ ] Information System Deevelopment and Coding skill for cloud computing 15. 下列何者不是成為一位Data Analyist應具備的技能? - [x] Machine building skills - [ ] Data warehouse modeling skills - [ ] Communication skills - [ ] Data Visualization skills 16. 下列何者不是大數據須面對之主要議題? - [ ] Big data analytics - [ ] Big data Technologies - [ ] Big data management - [x] 以上都須面對之議題 17. 下列何者不是因應大數據分散儲存及分析發展的雲端資料儲存模式? - [x] Microsoft SQL - [ ] Google BigTable - [ ] Amazon RDB - [ ] Microsoft SQL Azure 18. 下列何者不是NewSQL的產品? - [ ] Amazon RDB - [ ] MS SQL Azure - [x] my SQL - [ ] MySQL Cluster 19. Which task (or activity) is used to insure that the dixision model of conducting big data analytics is reliably implement? - [x] evaluation - [ ] validation - [ ] verification - [ ] auditing - 申論題 1. 請用一個可被清楚理解之架構圖(Diagram)描繪並簡要說明雲端運算機制MapReduce之平行運作模式,並簡要說明之。  2. 請分別簡述AI without Machine Learning, Machine Learning, Deep Machine Learning的概念,並個舉出幾個他們各自的主要應用。 - 應用: - AI:精準醫學、智慧醫院、心理健康、科技防疫、決策樹 - Machine Learning:動物分類、客戶細分、購買傾向、(信用卡)欺詐檢測、3 年、5 年、10 年腦/乳腺癌/結腸癌存活率預測 - Deep Machine Learning:AlphaGo、無人駕駛汽車 - 概念: - AI:計算機模仿人類思考進而模擬人類的能力/行為。 - 機器學習 概念:機器學習採用了人工智能的一些核心思想,並將其重點放在使用神經網路去模仿我們的決策(從資料中學習模型),來解決現實生活的問題 - 深度學習的概念:深度學習 更專注於機器學習的工具和技術,並將它們應用於解決幾乎任何需要“思考”(思考;推理)的問題。 3. 請用一個系統架構圖,表達專家系統建置所包含之系統功能元件(functional components)、領域專家、專家系統開發人員或知識工程師,他們在圖上有其功能之關聯性。  4. 請用圖描繪符合以下一倒傳遞類神經網路圖(Backpropagation Neural Network,BPNN):他至少包含兩層隱藏層(每層各含三個神經元(Neurons)),一可接受四種不同型態資料之輸入層,一可產生三種不同型態結果之輸出層。並說明其機器學習(machine learning)的原理 ```graphviz digraph graphname{ //spaces in1[label="input" shape="circle"] in2[label="input" shape="circle"] in3[label="input" shape="circle"] in4[label="input" shape="circle"] hl01[label="0X0" shape="circle"] hl02[label="0X1" shape="circle"] hl03[label="0X2" shape="circle"] hl11[label="1X0" shape="circle"] hl12[label="1X1" shape="circle"] hl13[label="1X2" shape="circle"] out1[label="output" shape="circle"] out2[label="output" shape="circle"] out3[label="output" shape="circle"] //input in1->hl01 in1->hl02 in1->hl03 in2->hl01 in2->hl02 in2->hl03 in3->hl01 in3->hl02 in3->hl03 in4->hl01 in4->hl02 in4->hl03 //hidden layer hl01->hl11 hl01->hl12 hl01->hl13 hl02->hl11 hl02->hl12 hl02->hl13 hl03->hl11 hl03->hl12 hl03->hl13 //output hl11->out1 hl11->out2 hl11->out3 hl12->out1 hl12->out2 hl12->out3 hl13->out1 hl13->out2 hl13->out3 } ``` 機器學習著重於訓練電腦從資料中學習,並根據經驗改進,而不是按照明確的程式碼運行作業。在機器學習中會訓練演算法尋找大型資料集的模式和關聯性,並根據該分析做出最佳決策和預測。機器學習應用程式會隨著使用不斷改善,存取的資料越多、準確度越高。 ###### tags: `Biomedical Informatics` `110` `2021` <style> .navbar-brand::after { content: " × FJUMIIA"; } </style>
×
Sign in
Email
Password
Forgot password
or
By clicking below, you agree to our
terms of service
.
Sign in via Facebook
Sign in via Twitter
Sign in via GitHub
Sign in via Dropbox
Sign in with Wallet
Wallet (
)
Connect another wallet
New to HackMD?
Sign up