1, Analyze the three product datasets provided , To determine , Describe and support mathematical evidence , Meaningful quantitative sum / Or qualitative model , relationship , Measurement standard and star rating , Parameters between comment and benefit rating , This will help sunshine succeed in its three new online market products .
For the problem 1, Determine the star level , The relationship and parameters between comment and benefit rating , You can use neural network knowledge , Star , Reviews and beneficial ratings are used as input neurons , Whether the final product online sale of sunshine company is successful as output neuron .
2. Use your analysis to answer the following specific questions and requirements of the marketing director of sunshine company :
a. Based on ratings and comments that best provide the following information , Determine data metrics . Sunshine Company tracking , Once their three products are sold on the online market .
problem a The first step is to digitize the comments , utilize Python Of NLP Speech processing , Comment on 0 Negative comments ,1 Positive comments , Then the comment and star data are normalized , You can get rid of one column , Simplify data . It can also be achieved by weight ratio , For example, comments account for 30%, Star occupation 70%.
b. Identify and discuss time-based metrics and patterns within each dataset , These measures and patterns may indicate that a product's reputation in the online market is increasing or decreasing .
Time series model was used , See how ratings and reviews change over time .
c Identify a combination of text-based metrics and rating based metrics , These two measures best indicate whether the product is likely to succeed or fail .
Using fuzzy evaluation method to measure the success of products .
d. Will higher stars lead to more reviews ? for example , After seeing a series of low stars , Are customers more likely to write some types of comments ?
Or the relationship between stars and reviews . According to the star rating, how about the reviews
e Specific quality descriptors for text based comments , as “ enthusiasm ”,“ disappointment ” And others , Is it closely related to the rating level ?
According to the comments , How about the stars .
Technology