ann_technical.Rmd

---
title: "Annex6_Technical_NN"
author: "Alejandro Vaca"
date: "4/4/2018"
output: word_document
---


```{r}
library(readr)

btc <- read_csv('btc_usd_22_03_18.csv')
eth <- read_csv('eth_usd_22_03_18.csv')

btc <- btc[ , 2:ncol(btc)]
eth <- eth[ , 2:ncol(eth)]

colnames(btc) <- c("date", "close", "high", "low", "open", "volfrom", "volto")
colnames(eth) <- c("date", "close", "high", "low", "open", "volfrom", "volto")

library(dplyr)
btc$date <- as.Date(btc$date)
eth$date <- as.Date(eth$date)
btc1 <- btc %>% 
  filter(date > "2013-05-01")

eth1 <- eth %>%
  filter(date > "2016-05-01")
```

First for bitcoin: 

```{r}
price = btc1$close
HLC = matrix(c(btc1$high, btc1$low, btc1$close), 
             nrow = length(btc1$high))

library("TTR")
library(forecast)
bitcoin.lr = diff(log(price))
```


```{r}
for(i in 1:nrow(btc1)) {
  if (btc1$date[i] == "2017-10-01") {
    print(i)
  } else {
    next
  }
}
```

```{r}
rsi = RSI(price)
macd = MACD(price)
macd= macd[ , 1]
will = williamsAD(HLC)
cci = CCI(HLC)
STOCH = stoch(HLC)
stochK = STOCH[,1]
stochD = STOCH[ , 1]
```

```{r}
#input and target matrix for training and validation dataset.
input = matrix(c(rsi[599:1784], cci[599:1784],  #observation 600 is 2014-12-22
                 macd[599:1784], will[599:1784], 
                 stochK[599:1784], stochD[599:1784]),
                 nrow = 1186)

target = matrix(c(bitcoin.lr[600:1785]), nrow = 1186)
date = as.Date(btc1$date[599:1784])
trainingdata = cbind(input, target)
colnames(trainingdata) = c("RSI", "CCI", "MACD", "WILL", 
                           "STOCHK", "STOCHD","Return")
trainingdata = data.frame(trainingdata)
trainingdata$date = as.Date(date)
```

```{r}
for(i in 1:nrow(trainingdata)) {
  if (trainingdata$date[i] == "2017-10-01") {
    print(i)
  } else {
    next
  }
}
```


```{r}
trainingdata = trainingdata[ , -8]
bitcoin.train = trainingdata[1:1014, ]
bitcoin.test = trainingdata[1015:nrow(trainingdata), ]

```

```{r}

#install.packages("nnet")
library(nnet)
#let´s start with neural networks 
best.network = matrix(c(5, 0.5))
best.rmse = 1

for (i in 5:15) 
  for(j in 1:3) {
    bitcoin.fit = nnet(Return ~ RSI + CCI + MACD + WILL + STOCHK + STOCHD,
                       data = bitcoin.train, maxit = 1000, size = i,
                       decay = 0.01 * j, linout = 1)
    bitcoin.predict = predict(bitcoin.fit, newdata = bitcoin.test)
    bitcoin.rmse = sqrt((mean(bitcoin.predict - bitcoin.lr[1614:1785])^2))
    if(bitcoin.rmse < best.rmse) {
      best.network[1,1] -> i 
      best.network[2,1] -> j
      best.rmse = bitcoin.rmse
    }
  }

```

```{r}
bitcoin.fit = nnet(Return ~ RSI + CCI + MACD + WILL +
                     STOCHK + STOCHD, data = trainingdata, maxit = 1000,
                   size = best.network[1,1], decay = 0.1*best.network[2,1],
                   linout = 1)

bitcoin.predict1 = predict(bitcoin.fit, newdata = bitcoin.test)

for ( i in 1:20) {
  bitcoin.fit = nnet(Return ~ RSI + CCI + MACD + WILL +
                       STOCHK + STOCHD,  data = trainingdata, maxit = 1000,
                     size = best.network[1,1], decay = 0.1*best.network[2,1],
                     linout = 1)
  
  bitcoin.predict = predict(bitcoin.fit, newdata = bitcoin.test)
  bitcoin.predict1 = (bitcoin.predict1 + bitcoin.predict) / 2
  
}

```

```{r}
plot(bitcoin.predict1, type = "l")
```

```{r}
write.csv(bitcoin.predict1, "ann_bitcoin.csv")
```