Team project recommendation: Gift recommendation-by xizhou

It has been nearly two months since I came to Microsoft for an internship, and the advanced software engineering (ASE) Course has started. At the end of last week's course, we formed a team. After thinking and preliminary research, I recommended "gift recommendations" for our team project as our subject.

Chinese people always advocate courtesy and exchange, and mutual gift in communication is an indispensable part of our social life. Based on the nature of the event, the purpose of the gift, and the expected effect, the gift should be determined by the object, self-identification, and information. The selection of gifts should be treated differently for different recipients. In general, it is better to be practical and cost-effective for the poor. Rich people, with exquisite creativity preferred. It is better for couples to commemorate romance. Gifts are good for friends. It is good for the elderly. It is better for children to be smart and novel. For foreign guests, special features are preferred.

The significance of a gift is much greater than we think. A gift is a language and a signal. It is implicit, but unforgettable. People who receive gifts from others will be happy, even if they don't like what you give.

Therefore, I hope to implement the following functions through the "gift recommendation" project:

1. Based on the relationship between the user and the recipient, the age of the recipient, the personality of the recipient, the festival/event of the gift, and the emotion to be conveyed, the user-based recommendation is carried out.
2. Similar gifts are recommended for item-based recommendations based on what users think is appropriate.
3. Recommended gifts can be presented based on the user's rating feedback to give recommendations that the system thinks are more reasonable.
4. Users can search for gifts to obtain the gifts that are commonly used.

Now we use the nabc method for analysis:

• Need

As mentioned above, we usually have a great demand for gift selection. I believe many people (especially engineering men ), I have had the same experience of searching on the Internet when I select a gift. The final gift is often unsatisfactory. If there is such an "gift recommendation" app, it will definitely make everyone look forward to it.

• Approach

A simple idea is to manually mark the appropriate gift categories (TAGS) for various gift occasions and purposes based on the opinions of books and the Internet as an initial solution, the information may be missing, and then a mathematical model (such as a probability graph model) is constructed. The model parameters are modified based on the feedback of each recommendation user to obtain a more reasonable recommendation gift category (TAG ), finally, you can obtain the e-commerce website search result based on the tag.

• Benefits

Even if the initial mark is unreasonable or missing, you can learn more suitable gift recommendations based on user feedback and interactions.

• Competition

At this stage, there are no apps that meet our needs. The market for gift recommendations is mainly in Q & A (such as Baidu Knows, zhihu, etc, we have obvious defects (which cannot be customized in such depth). The advantage is that Q & A often reports low-quality responses and gives enough feedback, it also provides professional recommendations.

After the above analysis, I suggest using "gift recommendation" as our project subject. Of course, we will go through more discussions to determine a suitable subject.

 

Team project recommendation: Gift recommendation-by xizhou