As more systems and sectors are driven by predictive analytics, there is increasing awareness of the possibility and pitfalls of algorithmic discrimination. In what ways do you think Recommender Systems reinforce human bias? Reflecting on the techniques we have covered, do you think recommender systems reinforce or help to prevent unethical targeting or customer segmentation? Please provide one or more examples to support your arguments.
Every algorithms, machine learning techniques or recommendation systems have always inaccuracy or biaseness that’s why models are always supposed to be checked through different evaluation tools. Even after normalization and evaluation for accuracy, it still persists inaccuracy and biasness. Also, predictions are always dependent on the data which is ratings in this case for the movies. If the users gives garbage reviews then obviously it will recommend inaccurate movies. The data needs to be monitored for inaccurate ratings. I have seen this problem especially from the users on Amazon. Amazon customers often times give high ratings if asked in a good way by company. Likewise, if customer got angry due to some other factor and not product then they give bad ratings. This is always the case and needs to be monitored. Personal biasness should be identify through classification techniques and identify the biased ratings. I remember 2016’s election. All the models were showing Hillary’s victory but the result came opposite with Trump winning the elections.