Im looking into the explanation of why certain results get ranked the way they do in my implementation of ElasticSearch. The thing I cannot figure out or find anywhere on the web is the following:
What is the motivation behind the formula of tfNorm? It includes the parameters termFreq, k1, b, avgFieldLength and fieldLength and how are these calculated?
Can someone please clarify how the algorithm works the way it does by explaining how you could manually calculate the aforementioned parameters.