With the development of the Internet, more and more apps have begun to make accurate recommendations. What followed was thousands of people, and the application of algorithms and the analysis of dimensions became more and more accurate. It's a good thing or a bad thing. Although it increases the immersive experience, on the entertainment platform, the extremes are counterproductive, and the high-quality content of spontaneous users is reduced, which is followed by boring and user experience fatigue.
But comprehensively speaking, the advantages and disadvantages of accurate recommendation outweigh the disadvantages. It is a good way to increase user stickiness. Today, I will only talk about some of the most basic algorithms. The advantages and disadvantages of precise recommendation will be analyzed in detail in the next article. The following is just my personal opinion, if you don't like it, don't spray it. thank you all. Let's take the label as an example to analyze. 1. The basis of the algorithm The Apriori algorithm is used to mine the basic algorithm of data association rules. It is used to find data sets that frequently appear in data values. Finding the patterns of these sets can help us make some decisions. For example, in common supermarket shopping datasets or e-commerce online shopping datasets, if we find frequently occurring datasets, then for supermarkets,
we can optimize the placement of products, and for e-commerce, we can optimize where the products are located. The location of the warehouse can save costs and increase economic benefits. The same user always browses the same form of content, so you can more accurately discover their points of interest and recommend other related products. The Aprior algorithm is a very classic mining algorithm. Many algorithms are based on the Aprior algorithm, including FP-Tree, GSP, and CBA. These algorithms use the idea of the Aprior algorithm, but make improvements to the algorithm. Isn't the true meaning of life just about making progress? 2. FP Tree algorithm This algorithm is my first choice when I just carried out the algorithm refinement, the structure is relatively simple, and it is suitable for the stage of just building.