LensKit デモ プログラムを実行すると、次のエラーが表示されます。
[main] ERROR org.grouplens.lenskit.data.dao.DelimitedTextRatingCursor - C:\Users\sean\Desktop\ml-100k\u - Copy.data:4: 無効な入力、行をスキップ
ML 100k データセットを作り直して、この行のみを保持するようにしましたが、これがどのように影響するかはわかりません。
196 242 3 881250949
186 302 3 891717742
22 377 1 878887116
244
これも私が使用しているコードです：
public class HelloLenskit implements Runnable {
public static void main(String[] args) {
HelloLenskit hello = new HelloLenskit(args);
try {
hello.run();
} catch (RuntimeException e) {
System.err.println(e.getMessage());
System.exit(1);
}
}
private String delimiter = "\t";
private File inputFile = new File("C:\\Users\\sean\\Desktop\\ml-100k\\u - Copy.data");
private List<Long> users;
public HelloLenskit(String[] args) {
int nextArg = 0;
boolean done = false;
while (!done && nextArg < args.length) {
String arg = args[nextArg];
if (arg.equals("-e")) {
delimiter = args[nextArg + 1];
nextArg += 2;
} else if (arg.startsWith("-")) {
throw new RuntimeException("unknown option: " + arg);
} else {
inputFile = new File(arg);
nextArg += 1;
done = true;
}
}
users = new ArrayList<Long>(args.length - nextArg);
for (; nextArg < args.length; nextArg++) {
users.add(Long.parseLong(args[nextArg]));
}
}
public void run() {
// We first need to configure the data access.
// We will use a simple delimited file; you can use something else like
// a database (see JDBCRatingDAO).
EventDAO base = new SimpleFileRatingDAO(inputFile, "\t");
// Reading directly from CSV files is slow, so we'll cache it in memory.
// You can use SoftFactory here to allow ratings to be expunged and re-read
// as memory limits demand. If you're using a database, just use it directly.
EventDAO dao = new EventCollectionDAO(Cursors.makeList(base.streamEvents()));
// Second step is to create the LensKit configuration...
LenskitConfiguration config = new LenskitConfiguration();
// ... configure the data source
config.bind(EventDAO.class).to(dao);
// ... and configure the item scorer. The bind and set methods
// are what you use to do that. Here, we want an item-item scorer.
config.bind(ItemScorer.class)
.to(ItemItemScorer.class);
// let's use personalized mean rating as the baseline/fallback predictor.
// 2-step process:
// First, use the user mean rating as the baseline scorer
config.bind(BaselineScorer.class, ItemScorer.class)
.to(UserMeanItemScorer.class);
// Second, use the item mean rating as the base for user means
config.bind(UserMeanBaseline.class, ItemScorer.class)
.to(ItemMeanRatingItemScorer.class);
// and normalize ratings by baseline prior to computing similarities
config.bind(UserVectorNormalizer.class)
.to(BaselineSubtractingUserVectorNormalizer.class);
// There are more parameters, roles, and components that can be set. See the
// JavaDoc for each recommender algorithm for more information.
// Now that we have a factory, build a recommender from the configuration
// and data source. This will compute the similarity matrix and return a recommender
// that uses it.
Recommender rec = null;
try {
rec = LenskitRecommender.build(config);
} catch (RecommenderBuildException e) {
throw new RuntimeException("recommender build failed", e);
}
// we want to recommend items
ItemRecommender irec = rec.getItemRecommender();
assert irec != null; // not null because we configured one
// for users
for (long user: users) {
// get 10 recommendation for the user
List<ScoredId> recs = irec.recommend(user, 10);
System.out.format("Recommendations for %d:\n", user);
for (ScoredId item: recs) {
System.out.format("\t%d\n", item.getId());
}
}
}
}
私はこれで本当に迷っており、助けていただければ幸いです。御時間ありがとうございます。