# Node attribute imputation using variational inference (NAIVI)
Author: **Simon Fontaine** (simfont@umich.edu)
## Repository structure
### NAIVI
Contains the source code to run various node attribute imputation methods.
- `vmp`: the proposed method (joint latent space model estimated with variational message passing)
- `mle`: the joint latent space model estimated using either maximum likelihood estimation with projection (MLE) or maximum a posteriori estimation (MAP) using the prior
- `mice`: imputation within the data matrix only using either MICE with linear models or KNN
- `constant`: imputaiton using the mean
- `gcn`: imputation using a graph convolutional network
### Pypet experiments
This folder contains utilities to run experiments using the pypet library.
### Experiments
This folder contains the scripts to run the experiments using `pypet`. Results and figures are also included in the
respective folders.
### Datasets
Datasets used in the real data experiments.
### Requirements
All requirements are contained in the `requirements.txt` file, or can be obtained using the `poetry` library.
Note: the latest version of `pypet` conflicts with newer versions of `numpy`. There are some basic types that were
removed from `numpy` that are referenced in `pypet`. This needs to be patched, otherwise you might not be able to import `pypet`.
Here is my fix: in `pypetconstants.py`, you need to change the following tuple at line 204:
```
PARAMETER_SUPPORTED_DATA = (numpy.int8,
numpy.int16,
numpy.int32,
numpy.int64,
# numpy.int,
numpy.int_,
# numpy.long,
numpy.uint8,
numpy.uint16,
numpy.uint32,
numpy.uint64,
# numpy.bool,
numpy.bool_,
numpy.float32,
numpy.float64,
# numpy.float,
numpy.float_,
numpy.complex64,
# numpy.complex,
numpy.complex_,
numpy.str_,
str,
bytes,
bool,
int,
float,
complex)
```