timecorr

timecorr copied to clipboard

Adding a function to calculating Spearman correlation

1

Hello, I am currently using this toolbox. Because of the data requirements of the Pearson correlation, I wanted to add the ability to calculate the Spearman correlation. After reading the...

BinglinW

Cannot install the latest package

3

## Summary The installation stuck in looking for compatible requirements. ## Problem I tried both installation commands ```bash pip install timecorr ``` and ```bash pip install --upgrade git+https://github.com/ContextLab/timecorr.git ``` Here...

const7

Working on memoizing `timecorr`—need to have hashable parameters (then we can simply use Python's `@lru_cache`). I've written a function that converts np.array and lists of np.array (non-hashable) into tuples (hashable...

F1LM1

ipyvolume visualization (static timecorr visualization)

2

Create a 3D correlation matrix (feature x feature x volume) and change the opacity to increase with stronger (absolute value) correlations. If the features are voxels, you can arrange them...

lucywowen

There may be a very quick way to implement the main timecorr algorithm using some sort of kernel method or convolutions (using the Gaussian weight function as the kernel). Need...

jeremymanning

Detect state changes using HMMs

2

Use brainiak implementation of event segmentation

jeremymanning

https://en.wikipedia.org/wiki/Multitaper

andrewheusser

Instead of assuming the user wants to reduce the data back to the original shape, allow the user to specify their own size (possibly dependent on level). Another idea would...

andrewheusser

consistent params for timecorr

2

For some parameters (rfun) you pass a string, but for other paramters (cfun, combine, weights_fun, etc.) you pass a function. Should we discuss this?

lucywowen

Add a `levels` argument to predict to optionally use the given list of levels of correlational info to predict future states. (Default: `levels=0`)

jeremymanning