void saveUrl(String s) {
SharedPreferences prefs =
getSharedPreferences(getString(R.string.preference_file_key), Context.MODE_PRIVATE);
ArrayList<String> urls = new ArrayList<String>();
try {
urls =
(ArrayList<String>)
ObjectSerializer.deserialize(
prefs.getString(
getString(R.string.preference_file_key),
ObjectSerializer.serialize(new ArrayList<String>())));
} catch (IOException e) {
e.printStackTrace();
}
if (!urls.contains(s)) urls.add(s);
// save the url list to preferences
SharedPreferences.Editor editor = prefs.edit();
try {
editor.putString(getString(R.string.preference_file_key), ObjectSerializer.serialize(urls));
} catch (IOException e) {
e.printStackTrace();
}
editor.commit();
}
private static String train(
SQLContext sqlContext, DataFrame positives, DataFrame negatives, String modelFileName) {
// combine data sets
DataFrame all = positives.unionAll(negatives);
// split into training and test sets
DataFrame[] split = all.randomSplit(new double[] {.80, .20}, 1);
DataFrame training = split[0].cache();
DataFrame test = split[1].cache();
// fit logistic regression model
PipelineModel model = fitLogisticRegressionModel(training);
try {
ObjectSerializer.serialize(model, modelFileName);
} catch (IOException e1) {
// TODO Auto-generated catch block
e1.printStackTrace();
}
// predict on training data to evaluate goodness of fit
DataFrame trainingResults = model.transform(training).cache();
// predict on test set to evaluate goodness of fit
DataFrame testResults = model.transform(test).cache();
// predict on unassigned data mentions
StringBuilder sb = new StringBuilder();
sb.append(getMetrics(trainingResults, "Training\n"));
sb.append(getMetrics(testResults, "Testing\n"));
return sb.toString();
}