Your SlideShare is downloading. ×
PredictionIO - Scalable Machine Learning Architecture
Upcoming SlideShare
Loading in...5

Thanks for flagging this SlideShare!

Oops! An error has occurred.

Saving this for later? Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime – even offline.
Text the download link to your phone
Standard text messaging rates apply

PredictionIO - Scalable Machine Learning Architecture


Published on

PredictionIO's presentation slides for Data Science London on April 24, 2013 during the Big Data Week.

PredictionIO's presentation slides for Data Science London on April 24, 2013 during the Big Data Week.

Published in: Technology, Education

  • Be the first to comment

No Downloads
Total Views
On Slideshare
From Embeds
Number of Embeds
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

No notes for slide


  • 1. Simon Chansimon@prediction.ioData Science London - April 24, 2013Big Data Week
  • 2. Machine Learning is....computers learning to predictfrom data
  • 3. puttingMachine Learninginto practice
  • 4. challenge #1Scalability
  • 5. Big Data BottlenecksMachine Learning Processing
  • 6. PredictionIO has ahorizontally scalablearchitecture
  • 7. Async SDKClient client = new Client(appkey);// Adding user behaviorsreq = client.getUserRateItemRequestBuilder(uid, iid, rate);client.userRateItemAsFuture(req);
  • 8. PlayFramework‣ stateless - no server session‣ non-blocking web request
  • 9. Play: A Non-blocking Exampledef index = Action {val futureInt = scala.concurrent.Future { slowDataProcess() }Async { => Ok(views.html.result.render(i)))}}
  • 10. MongoDB‣ Read scaling: Replica Sets‣ Write scaling: Sharding‣ Indexes (e.g. geospatial){ geoSearch : "places", near : [33, 33],maxDistance : 6, search : { uid : "user1" } }
  • 11. HadoopHadoop&Cascading&(Java)&Scalding&(Scala)&
  • 12. MapReduce- Native Javapublic class WordCount {public static class Map extends Mapper<LongWritable, Text, Text, IntWritable> {private final static IntWritable one = new IntWritable(1);private Text word = new Text();public void map(LongWritable key, Text value, Context context) throws .....{String line = value.toString();StringTokenizer tokenizer = new StringTokenizer(line);while (tokenizer.hasMoreTokens()) {word.set(tokenizer.nextToken());context.write(word, one);}}}public static class Reduce extends Reducer<Text, IntWritable, Text, IntWritable> {public void reduce(Text key, Iterable<IntWritable> values, Context context)throws IOException, InterruptedException {int sum = 0;for (IntWritable val : values) { sum += val.get(); }context.write(key, new IntWritable(sum));}}public static void main(String[] args) throws Exception {Configuration conf = new Configuration();Job job = new Job(conf, "wordcount");job.setOutputKeyClass(Text.class);job.setOutputValueClass(IntWritable.class);job.setMapperClass(Map.class);job.setReducerClass(Reduce.class);job.setInputFormatClass(TextInputFormat.class);job.setOutputFormatClass(TextOutputFormat.class);FileInputFormat.addInputPath(job, new Path(args[0]));FileOutputFormat.setOutputPath(job, new Path(args[1]));job.waitForCompletion(true);}}
  • 13. MapReduce- Scaldingclass ScaldingTestJob(args: Args) extends Job(args) {Tsv(args(0), text).flatMap(text -> word) { text : String => text.split("s+") }.groupBy(word) { _.size }.write(Tsv(args(1))}
  • 14. Sample Code
  • 15. ### Sample PredictionIO Python SDK Codeclient = predictionio.Client(appkey="<your app key>")# Add Dataclient.create_user(uid=”user123”)client.create_item(iid=”itemXYZ”, itypes=(1,))client.user_view_item(uid=”user123”, iid=”itemXYZ”)# Get Predictionrec = client.get_itemrec(engine="<engine name>", uid=”user123”, n=5)
  • 16. GettingInvolved!- @PredictionIO- - Newsletter-
  • 17. Q&AQ: Selecting the right features is a big problem. Can PredictionIO solve this problem?A: Not at this moment.That’s why we focus on collaborative filtering algorithms right nowwhich don’t require the use of features.And we believe that the involvement of datascientists is needed for many specific problems. PredictionIO is positioned as a tool tomake their work easier, but not as a replacement.Q: How’s PredictionIO different from Weka?A:Weka, like Mahout, is a ML algorithm library.You can see PredictionIO as a layer on topof it, which helps you to implement algorithm into production environment by providing acomplete infrastructure.Q: How do you compare PredictionIO with RapidMiner?A: RapidMiner is a great product to define data engineering workflow visually.PredictionIO focuses on a different problem -- i.e. deploying ML solution into productionenvironment.Q: How does the algorithm evaluation metrics work in PredictionIO?A: At this moment, you can evaluate algorithms by some offline metrics, such as MeanAverage Precision, based on your existing data.Q:What’s the business model?A: We focus on making PredictionIO a useful open source product at this moment.