Can we predict the categorical variable of the given dataset? I have a dataset (3000 rows) as below, as one can see the dataset also contains few important string columns as Location,Country,Injury,Time. Can we predict the Survived feature using any predictive model? 
   Year     Location      Country    Age                Injury Survived Time
0  2016     chittwan      Nepal      09         attack on head     0    11h00
1  2016      siberia      Russia     19            Neck Injury     0    13h00
2  2016    Oba Hills      Nigeria    23              lower leg     1    04h43
3  2016     Edumanom      Nigeria    65             leg injury     0    11h00
4  2016        Kanha      India      NaN  attacked from behind     1    18h00
5  2016     Edumanom      Nigeria    09         attack on head     0    19h00
6  2016  Ranthambore      India      39            Neck Injury     0    11h00
7  2016     chittwan      Nepal      13   attacked from behind     1    09h12
8  2016     Edumanom      Nigeria    NaN            leg injury     1    11h00
9  2016        Bikin      Russia     NaN            leg injury     1    13h00

 A: Yes, a predictive model can be developed to predict the Survived feature. A problem similar to this is given in the link https://www.kaggle.com/c/titanic. It has some very good tutorials too. 
This is a binary classification problem. A logistic regression, Neural Networks or Random forests are suitable techniques. 
In this example, most of the features are categorical. Convert them accordindly, for eg. to numerical features (via one hot encoding) before inputting to a machine learning algorithm. Use cross validation to select the best technique.
A: I would suggest a logistic regression model, which can be implemented in R. Logistic regression is a type of generalised linear model, which is why in the code below the key function is glm
Here is some pseudo code:
library(MASS) # Load the MASS library, which contains the glm function

model = glm(Survived ~ Location + Country + Injury + Time) # Define a model of whether the victim survived, predicting on location, country, injury and time of attack

summary(model) # View the model coefficients to interpret the model

In terms of selecting a specific model, I would suggest investigating the stepAIC function in R, as some predictors may drop out/or be included in the model, for example interaction effects.
