Meeting Notes 20090604 Phone meeting (Dejing, Haishan, Gwen, & Bob; 6/4/2009)"

From Nemo

Jump to: navigation, search


NEMO phone conf (Haishan, Dejing, Bob, & Gwen)

Main topic: review results from Haishan's latest metric-mapping studies, as documented in ' SimDataPCAICA-06042009gf.doc.'


  1. Evaluation of the results so far. -- Good.
  2. Why are there only spatial metrics included in this study? -- Assume 1-1 mapping. GF: But please include TI-max1 and TI-max2, where there is a 1-1 mapping''.
  3. What are the intermediary steps (clustering, smoothing, normalization, etc) for?
  4. Explain the assumptions in the experiment.
    1. 1-1 mapping between metrics
    2. similar structures/patterns exist in both datasets
  5. Shall we prepare new datasets to conduct more experiments? -- Yes for paper'''.
  6. Mapping between patterns -- Probably requires a different method'''.
  7. Evaluate/revise the to-do list
    1. JUNE 4 -- Completed tech report based on old simulated ERP dataset & initial methods (modified 'SimDataPCAICA-05202009rf-v01.doc') √
    2. JUNE 10 -- Revised tech report that includes:
      • Table captions
      • A few sentences at beg of Mapping section (Sec 3) that explain goals, methods, & method assumptions.
      • New section ("Study 4") that reports results over multiple randomizations (n=5) for each of the 4 experiments -- with inclusion of TI-max1 and TI-max2 metrics.
    • JUNE 11 -- Completed draft of poster, ready for the rest of us to comment on and help edit.
    • JUNE 15 -- Completed draft of ICDM paper that is based on tech reports & on HBM poster presentation

I want to include two figures here from the draft of the paper which illustrate the idea of row alignment according to cluster label and subsequent normalization and smoothing processes:

Figure 1: Point-sequence curves of two attribute IN-LOCC (A) and IN-01 (B)
Figure 2: After reording the points according to the cluster label. Smoothing and normalization are also performed as pre-processing steps before calculating the similarity (distance) of the two curves.
/wiki/images/1/17/Fish1.png /wiki/images/e/ea/Fish2.png /wiki/images/f/fa/Fish3.png /wiki/images/f/ff/Fish4.png /wiki/images/4/40/Fish5.png /wiki/images/c/c5/Fish6.png
Personal tools