Robert’s Working Blog

1. http://webconferencing-test.com/en/webconferencing_top5/webconference_compact_results.html
2. http://office.microsoft.com/en-us/livemeeting/FX101729061033.aspx

1. http://www.cellml.org/models/gonzalezheydrich_steingard_kohane_1994_version02
2. http://www.cellml.org/models/gonzalezheydrich_steingard_kohane_1994_version02

1. http://en.wikipedia.org/wiki/Homoscedasticity
2. http://en.wikipedia.org/wiki/Bartlett%27s_test

Principles of biostatistics / Marcello Pagano, Kimberlee Gauvreau.

1. curve fitting for individual woman and calculate the median trajectory, then curve fitting this median trajectory to confirm an equation
2. curve fitting for all existing data to choose an equation
3. curve fitting for individual woman (probably different woman has different equation), rank these equations in a “probability distributions” to find an equation with largest proportions of ranking.

We need to discuss some possible scenarios:
1. data is not well fitted

Discussion:
1. Curve fitting to find the best equation needs trade on “simplicity” and “goodness of fit”.
2. Needs to match the general trend of blood hormone
3. Needs to define some criteria, which we can’t predict this woman.

1. corrcoef() to calculate the “correlation coefficient” in matlab

1. homoscedasticity

http://www.le.ac.uk/bl/gat/virtualfc/Stats/variance.html