-- Mar 2. In-Class Exercise Thread

After Normalisation : cars difficulties and motorcrosses After stemming : car difficulti motorcross example: Boiling water would change to Boil water after stemming which results in two different meanings.

-- Mar 2. In-Class Exercise Thread

1) Car,difficulti motorcross 2) The word "news" will be stemmed as "new".Both words have different meanings and this will impact the search query results.

-- Mar 2. In-Class Exercise Thread

   stemmed: [car, difficulti, motorcross]

1. Normalized : [cars, difficulties, motorcrosses] stemmed: [car, difficulti, motorcross] 2. An example would be "runner" which would get stemmed to "run". Now "run" would return a lot of results that would match the relevant set but the precision would be low because we will get a resulting set that would have results spanning "run", "running", "runner". so this actually returns a lot of results that might not all be relevant to our actual search term that was "runner"

-- Mar 2. In-Class Exercise Thread

1. 
cars -> car
difficulties -> difficulti
motorcrosses -> motorcross 
 
2.
Running stemmers on our query terms might increase the number of documents retrieved on a search, since multiple words might boil down to have the same stem word. This, however, does not go to say that all of them would be relevant to the query. 
Eg. The word "marketing" will be stemmed to "market", both of which signify different things.

<pre> 1. cars -> car difficulties -> difficulti motorcrosses -> motorcross 2. Running stemmers on our query terms might increase the number of documents retrieved on a search, since multiple words might boil down to have the same stem word. This, however, does not go to say that all of them would be relevant to the query. Eg. The word "marketing" will be stemmed to "market", both of which signify different things. </pre>

-- Mar 2. In-Class Exercise Thread

  car

  difficulti

  motorcross

1. Cars, difficulties, and motorcrosses becomes: car difficulti motorcross 2. Stemming might improve recall and return an increased number of documents but it will decrease precision. Eg: connected, connection, connects might all stem to connect.

-- Mar 2. In-Class Exercise Thread

1. car, difficulti, motorcross 2. Caring will after stemming become car. These two words have very different meanings. This will impact the search results. Stemming would decrease precision but increase recall

-- Mar 2. In-Class Exercise Thread

1. Given: ”Cars, difficulties, and motocrosses" After normalization: Cars difficulties and motocrosses After stemming using Porter Stemming: car difficulti and motorcross 2. An example would be "winner" which would get stemmed to "win". Now "win" would return fewer results that would match the relevant set but the precision would be low because we will get a resulting set spanning "win", "winning", "winner".