2017.11.3
Short term goal: (1)Finish the experiments : build topic Evolution roads. (2) Participate in the construction of the Danmaku datasets. Build participles vocabulary and preprocess the datasets.
- Analysis the tweeter 2011 relults, the two baselines models are little worse than the the paper <Real-Time Top-R Topic Detection on Twitter with Topic Hijack Filtering> (KDD 15).
2.Finish analysis the statistics information of the Youku datasets.
2017.11.2
Short term goal: (1)Finish the experiments : build topic Evolution roads. (2) Participate in the construction of the Danmaku datasets. Build participles vocabulary and preprocess the datasets.
1.Dessicuss with Prof.Fang about the experiments results. Because of the perplixity is much larger than other papers(the baselines are also very large ). He suggest me try to use tweeter datasets and compare the results with other papers.Thanks ChenQin give me 4 years' tweeter datasets.
2.Analysis the Youku datasets.Mainly some statistics information.
- Talk about the Danmaku comments labeling work with prof.Hu and Song yang.
2017.11.03
Short term goal: revise the TKDE paper
Writing: rewrite the section of "Comparison with the State-of-the-art" by adding 2 baselines. Organize the experimental results and add a new section of "Comparison with the Microblog Track Runs"
2017.11.1
Short term goal: (1)Finish the experiments : build topic Evolution roads. (2) Participate in the construction of the Danmaku datasets. Build participles vocabulary and preprocess the datasets.
- Finish building the common vocabulary of the comments.
- Do the jiont NMF experiments.
2017.10.31
Short term goal: (1)Finish the experiments : build topic Evolution roads. (2) Participate in the construction of the Danmaku datasets. Build participles vocabulary and preprocess the datasets.
1.Collect and preprocess the common vocabulary of the comments. Include film,TV show,cartoon and so on, which can be used in danmaku comments word segmentation.
- Do the jiont NMF experiments. Get and analysis baselines results.
3.Help Meihong run her JavaWeb project.
3.Import thecomments into MySQL,
2017.10.30
Short term goal: (1)Finish the experiments : build topic Evolution roads. (2) Participate in the construction of the Danmaku datasets. Build participles vocabulary and preprocess the datasets.
1.Do the jiont NMF experiments. Get and analysis baselines results.
2.Read paper of <Crowdsourced Time-sync Video Tagging using Temporal and Personalized Topic Modeling> (KDD2014)
3.Participate in the construction of the Danmaku datasets: collect the common vocabulary of the comments.
