23-solutions-ann.Rmd

# Solutions ch. 10 - Artificial neural networks {#solutions-ann}

Solutions to exercises of chapter \@ref(ann).

## Exercise 1

```{r}
library("neuralnet")
 
#To create a neural network to perform square root

#Generate 50 random numbers uniformly distributed between 0 and 100
#And store them as a dataframe
traininginput <-  as.data.frame(runif(50, min=0, max=100))
trainingoutput <- sqrt(traininginput)
 
#Column bind the data into one variable
trainingdata <- cbind(traininginput,trainingoutput)
colnames(trainingdata) <- c("Input","Output")
 
#Train the neural network
#Will have 10 hidden layers
#Threshold is a numeric value specifying the threshold for the partial
#derivatives of the error function as stopping criteria.
net.sqrt <- neuralnet(Output~Input,trainingdata, hidden=10, threshold=0.01)
print(net.sqrt)
 
#Plot the neural network
plot(net.sqrt)
 
#Test the neural network on some training data
testdata <- as.data.frame((1:10)^2) #Generate some squared numbers
net.results <- compute(net.sqrt, testdata) #Run them through the neural network
 
#See what properties net.sqrt has
ls(net.results)
 
#see the results
print(net.results$net.result)
 
#Display a better version of the results
cleanoutput <- cbind(testdata,sqrt(testdata),
                         as.data.frame(net.results$net.result))
colnames(cleanoutput) <- c("Input","Expected Output","Neural Net Output")
print(cleanoutput)
```

*Acknowledgement: this example excercise was from* http://gekkoquant.com/2012/05/26/neural-networks-with-r-simple-example/