2015年12月11日 星期五

[Spark] Spark TW 2015年最後一場活動


2015年的最後一場,也是SparkTW成立一週年,特別有意義的場子.不用特別報名,但是晚來就只能坐地板了!
2015年底最強壓軸檔 -- 時間:12/16號(三)晚上 7:30 -- 地點:大安站燦坤旁資策會九樓

非常有榮幸邀請到在Kaggle以及業界都相當有經驗的Mark Peng大大來跟我們分享他在參與比賽時的寶貴經驗.

講者資訊: Mark Peng got 7th place of 1334 teams in Search Results Relevance competition and is ranked as top 300th place of 393,443 users with a Master title on Kaggle.

He is the R&D Manager and Head of Search Engineering in Uitox Corporation. He received his Master degree of Information Management from National Taiwan University. He had several years of experience in Web mining, NLP, sentiment analysis (opinion mining) and search engine applications. He is also familiar with machine learning tools, such as R, Python, Mahout and Spark.
分享主題: General Tips for participating Kaggle Competitions
Agenda:
* Data science process on Kaggle
* Data cleaning techniques
* Mostly used ML models
* Feature engineering methods
* Why cross-validation matters
* Ensemble learning
* How to team up
* Recommended books, MOOCs and resources