Assignments Session 5

1. Convert Data into a time series format and then calculate the returns from it.

Data set- CNX Mid-cap Index downloaded from NSE from August 2012-January 2013-10th reading to 95th reading. 

>nse6m<- read.csv(file.choose(),header=T)

> Close<- nse6m$Close

> Close.ts<- ts(Close)

> Close.ts<- ts(Close, deltat=1/252)

> Close.interval<-ts(data=Close.ts[10:95],frequency=1,deltat=1/252)

> Close.diff<- diff(Close.interval)

> Close.lag<- lag(Close.interval, K=1)

> Returns<- Close.interval/Close.lag

> Returns.table<- cbind(Close.interval,Close.lag,Close.diff,Returns)

> plot(Returns.table,main="Data from Index, Difference, Return")

 

 2.  First 700 data is available, Predict the data from 701-850, use the GLM estimation using LOGIT Analysis for the same.

> logit<-read.csv(file.choose(),header=T)

> logit.data<- logit[1:700,1:9]

> head(logit.data)

  age ed employ address income debtinc creddebt othdebt default

1  41  3     17      12    176     9.3    11.36    5.01       1

2  27  1     10       6     31    17.3     1.36    4.00       0

3  40  1     15      14     55     5.5     0.86    2.17       0

4  41  1     15      14    120     2.9     2.66    0.82       0

5  24  2      2       0     28    17.3     1.79    3.06       1

6  41  2      5       5     25    10.2     0.39    2.16       0

> logit.data$ed<-factor(logit.data$ed)

> logit.est<-glm(default ~ age + ed + employ + address + income, data=logit.data, family ="binomial")

> summary(logit.est)

Call:

glm(formula = default ~ age + ed + employ + address + income,

    family = "binomial", data = logit.data)

Deviance Residuals:

    Min       1Q   Median       3Q      Max 

-1.5229  -0.8196  -0.5408   1.0673   2.8521 

Coefficients:

             Estimate Std. Error z value Pr(>|z|)   

(Intercept) -0.894875   0.450781  -1.985 0.047126 * 

age          0.024982   0.015655   1.596 0.110541   

ed2          0.226624   0.216796   1.045 0.295868   

ed3          0.357120   0.291745   1.224 0.220921   

ed4         -0.007832   0.403119  -0.019 0.984499   

ed5         -1.032156   1.201660  -0.859 0.390372   

employ      -0.168476   0.024544  -6.864 6.69e-12 ***

address     -0.062393   0.018521  -3.369 0.000755 ***

income       0.013021   0.004144   3.142 0.001679 **

---

Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

(Dispersion parameter for binomial family taken to be 1)

    Null deviance: 804.36  on 699  degrees of freedom

Residual deviance: 711.89  on 691  degrees of freedom

AIC: 729.89

Number of Fisher Scoring iterations: 5

> forecast<-logit[701:850,1:8]

> forecast$ed<-factor(forecast$ed)

> forecast$probability<-predict(logit.est,newdata=forecast,type="response")

> head(forecast)

    age ed employ address income debtinc creddebt othdebt probability

701  36  1     16      13     32    10.9     0.54    2.94  0.04370623

702  50  1      6      27     21    12.9     1.32    1.39  0.11226674

703  40  1      9       9     33    17.0     4.88    0.73  0.17599440

704  31  1      5       7     23     2.0     0.05    0.41  0.24972447

705  29  1      4       0     24     7.8     0.87    1.01  0.37009683

706  25  2      1       3     14     9.9     0.23    1.15  0.44594623