It is currently able to push Lavaan and Mirt models to AI Ninja for post hoc score estimation.
Install the package:
install.packages("remotes")
remotes::install_github("ianrothmann/ainr")You need to install.packages("dotenv") and create a .env file in your project.
The env file should look like this:
AIN_URL=https://api.aininja.dev
AIN_KEY=[Your key here]
#Add an empty line at the end
In your project, import dotenv and load your env file as follows:
library(dotenv)
load_dot_env(file = "~/Dev/r/ainr/.env")For Lavaan models:
library(lavaan)
library(ainr)
#Normally estimate your model through lavaan
fit <- sem(model, data=myData, estimator = "WLSMV", ordered = c(...))
#Upload your model to AI Ninja
ain.uploadLavaanModel('model_name',model.cons.fit) #replace model_name with a name and version of your model you would like to deploy.It would be a good idea to test score estimation on your model using lavPredict before uploading.
For Mirt models:
library(mirt)
library(ainr)
#Normally estimate your model through mirt
cmod <- mirt(myData, itemtype="Rasch", SE=TRUE, verbose=FALSE)
#Upload your model to AI Ninja
ain.uploadMirtModel('model_name',cmod) #replace model_name with a name and version of your model you would like to deploy.It would be a good idea to test score estimation on your model using fscores before uploading.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent