研究生: |
陳寧寬 Chen, Ning-Kuan . |
---|---|
論文名稱: |
以計量經濟和機器學習方法分析 iOS 和 Android 手機應用程式(APPs) Analyzing iOS and Android APPs with Econometric and Machine Learning Methods |
指導教授: |
林世昌
Lin, Eric S. |
口試委員: |
張焯然
Zhang, Zhuo-Ran 周大森 Zhou, Da-Sen |
學位類別: |
碩士 Master |
系所名稱: |
科技管理學院 - 經濟學系 Department of Economics |
論文出版年: | 2020 |
畢業學年度: | 108 |
語文別: | 中文 |
論文頁數: | 32 |
中文關鍵詞: | 蘋果作業系統 、安卓作業系統 、手機應用程式 、訂價 |
外文關鍵詞: | iOS, Android, APP, Pricing |
相關次數: | 點閱:2 下載:0 |
分享至: |
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
本研究透過 Kaggle 網站的資料分析 Android 和 iOS 平台 APP 定價、評價等因素是否不同 , 首先 , 本研究使用文字雲的視覺化方式呈現 Android 和 iOS 在標題用字上的差異 , 我們發現 iOS APP 用字以遊戲為主 , Android APP 用字則以免費為主。 我們接著使用計量經濟和機器學習的方法實證分析兩個平台影響 APP 的定價和評價的重要變數 , 以及透過有序羅吉斯迴歸分析 Android APP 的下載量。 最後我們利用詞袋模型結合帶有 Lasso 懲罰項的羅吉斯迴歸 , 成功建立出一個不需要人工判別 , 即可高準確判斷 APP 留言為正向或是負向的模型 , 該模型有效的縮減維度 , 使得
資料的維度從 177,448 下降到 2,707, 而透過 Lasso 懲罰項 , APP 開發商亦可快速得知何種關鍵字顯著影響 APP 評價。
This research analyzes the difference between the Android and iOS platform apps through the data on the Kaggle website. First of all, this research uses a word cloud to visualize the difference between Android and iOS in terms of title. We found that the iOS APP mainly
uses games, while the Android APP uses free words. We also use econometric methods and machine learning methods to analyze the differences between the two platforms that affect APP pricing and the important variables that affect APP evaluation. There is also an ordered logistic regression to analyze Android APP downloads. Finally, we use the bag-of-words model combined with Logistic regression with Lasso penalty term, Successfully established
a model that can judge whether the APP message is positive or negative without manual discrimination. The model successfully reduced the dimension, reducing the dimension of the data from 177,448 to 2,707. And through the Lasso penalty item, developers can also quickly know which keywords are the reasons for app evaluation.
[1] Breiman, L. (2001). Random Forests. Machine Learning, 45, 5-32.
[2] Choi, H. S. (2015). Pricing Strategies of Mobile Apps and Their Effects on Piracy: An
Empirical Study. 2015 48th Hawaii International Conference on System Sciences.
[3] Finkelstein, A., Harman, M., Jia, Y., Martin, W., Sarro, F., and Zhang, Y. (2017)
Investigating the relationship between price, rating, and popularity in the Blackberry
World App Store. Inf Softw Technol 87:119–139.
[4] Fu, B., Lin, J., Li, L., Faloutsos, C., Hong, J., and Sadeh, N. (2013). Why people
hate your app: Making sense of user feedback in a mobile app store. In Proceedings of
the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data
Mining, pp. 1276–1284.
[5] Kramer, A., and Kalka, R. (2017). How Digital Disruption Changes Pricing Strategies
and Price Models. in Khare A., Stewart B. and Schatz R., eds, Phantom Ex Machina-
Digital Disruption’s Role in Business Model Transformation, Springer, pp.87-103.
[6] Lee, G., Santanam, R. (2011). Product portfolio and mobile apps success: Evidence
from app store market. In 17th Americas Conference on Information Systems 2011,
AMCIS 2011 (pp. 3912-3921).
[7] Liu, W., Zou, Y., Yang, Y., Cheng, W., and Zhang, G. (2017). How app update affects
app download in iOS appstore. 2017 3rd IEEE International Conference on Computer
and Communications (ICCC).
[8] Mahmood, A. Identifying the influence of various factor of apps on google play apps
ratings. J. of Data, Inf. and Manag. 2, 15–23 (2020).
[9] Roma, P., Perrone, G., and Valenti, F. 2013. An empirical analysis of revenue drivers in
the mobile app market. Proceedings of POMS 24th Annual Conference 2013, May 3-6,
Denver, USA.
[10] Statista (2020). Mobile operating systems’ market share worldwide from January 2012
to December 2019. Retrieved May 10, 2020, from https://www.statista.com/statis-
tics/272698/global-market-share-held-by-mobile-operating-systems-since-2009/.