CMU 11642 's curriculum notes outline. Involves a lot of algorithms, detailed see the specific links, the code is not posted. Welcome to the discussion, please correct me ~
Jamie Search Engine This course, but also very fruitful, in addition to some basic concepts and algorithms in the class, there are many new research, covering a very wide range, absolutely more than one book. According to Jamie, in the Yahoo and other companies search department students come back to say now do work feeling is the homework, whether there is exaggeration do not know, but we can feel. Has passed the elective stage, as to the next want to choose a small pot friends a little workload Information Bar: reading notes. There are a lot of reading every week, maybe a textbook or a paper. Note that reading notes is binary, 1 or 0, do not think that on the blackboard to see their 80 points on the thought of 0.8, 80 points = 0 points. Homework. Five times to complete a perfect search engine, the language is Java, about thirty or forty classes, each time on the basis of a further improvement, so in addition to the last job, you do every homework performance will deeply affect the next time. If you find that your running time is much longer than Callen, be sure to optimize it. The work is not difficult, usually lets you realize each kind of algorithm, commonly used as well as some papers, however the score is very strict, many corner case should notice.
Each job is completed have a report, need to do a lot of experiments (forty or fifty at least, do not write a script can feel from dark to do Dawn), and do "profound" summary, the reason for saying "deep" is because sometimes I racked my brains to write something to get the comment is shallow. A sad tear. In general, write the algorithm one day after all the test case, and finally do the experiment write report. Exam Midterm final two examinations, on the text analytics people know, Callen's consistent style, examination breadth First, the question quantity is big, relies on the instinct, you make time to think you lose, gives the cent low.
But don't be afraid to say so much. Your final score will look good even if your test scores are lower.
About being able to live hold, let's just say I chose Machine Learning (11601A), distributed Systems (95702), and Data structures for application last semester. Programmers (08722), feel 4 courses frankly classes can only focus one to two doors, if you have to brush questions to find work, or suggest p/f or audit a door.
Then back to the point, a high sum, this course on two questions, one is how to accurately match the query and document, one is how to quickly return the search results, is the effect vs efficiency of a trade-off. The following summary summarizes the focus of the course, which will involve a number of specific algorithms, but this is a simple outline, not the formula/algorithm are listed, specific to see the link below or read/handout. Reveal a point: most algorithmic projects you need to implement, and do not need to implement the algorithm, Jamie will not easily let you off, so where do you think it will appear.