- Upgrade Mira export file creation to take into account the case where several conditions are present. (Pending for Q to act)
- Mira: clean the table to remove bad unit before working with it. Will make code more readable.
- Ace TCR/deg files should be robust with regard of first line (iteration=...)
- NBTI exportfile to script.
- Rewrite for loops in the functions in order to enhance readability.
- Create a function that estimate the lifetime of the device. par: Scale, Ea, n, A, DeviceID, ConditionTable. Make it stand alone if user want to use it to check something. And add a call in BlackModelization (optional parameter). Could be used to estimate quickly time for a new experiment.
- Test TDDB modelization when several area are provided.
- Compare datacruncher results with amsRel for TDDB.
- Robust modelisation: fit 1 followed by a study on outliners leading to a second fit (Cook's distance). see nlrob (robustbase)
- Implement hierachical models (mixed effect models) to check if they can be useful for self heating interference.
- Force fit option: avoid fitting again if the parameters are available
- Implement Kaplan-Meier and Exponential Greenwood Confidence Intervals
- Look into J.G. Stotvig codes. fitdistr is failing when censored samples are provided. Take advantage of know probability.
- Student's t-test function in order to analyze populations (eg: impact of lenght, or tool or wafer)
- Ficher's F-test to add to previous function.
- Copy the structure file locally (dimension of EM structures) and make comparison with time stamps. Would allow working without network.
- Find a way to store model in a file and directly call them. Avoid hard coding the models in the functions. Avoid redundancy (error prone). Ex: Black and TDDB lifetime.
- plyr strategie (split apply combine) on data with multiple structures inside the table?
- Use ldply to read a list of file instead of for loops.
- Create a function to compare experiments (by conditions, by wafer, by lot...)
- Create a function to save an exportfile corrected by a lifetime factor (useful to merge several wafer in 1 lot.)
- Enhance legend presentation to include:
- MTTF, Number of samples + number of censored samples<-- Create a table for this like in Minitab
- Model parameters <- addressed with annotate + clipping off (2nd table)
- Minimum 3 decades per charts
- Revert renaming CreateGraph function.
- Create new function for axis setting definition: deg scale or proba scale / linear or log scale
- Create a small GUI to select the analysis to be performed
- Add an update function to the GUI in order to update the package.
- Add a close after analysis check box.