使用线性回归时不正确的Scord预测 [英] Incorrect Scord predictions when using Linear Regression
问题描述
我正在使用线性回归算法编写示例实验，但是在通过预测性Web服务进行测试时，我看到Trained模型返回的预测不正确.可能是我选择了错误的算法.需要有关如何调整的帮助
针对以下用例进行实验以预测正确的结果.
1.数据源是CSV文件，它具有4个功能(余额，年龄，因子，输出)，其中3个功能(余额，年龄，因子)的组合将决定输出值.输出要素和两个输入要素(平衡和因数)均为浮动类型
点和年龄是数字.
输入约有350条记录，其中70％作为训练数据传递给使用线性回归机器学习算法的训练模型.
部署了预测性Web服务，并且在测试模型时，我看到一些数据组合返回负输出值.数据的数据组合非常有限，正在返回接近结果.
如何解决此问题并返回接近输出值的计分预测?请提出建议.
Hi,
I was writing a sample experiment using Linear Regression algorithm but I see incorrect predictions being returned by the Trained model when testing through predictive webservice. May be I have choose the incorrect algorithm. Need help with how to tweak the
experiment for below mentioned use case to predict the right results.
1. Source of data is a CSV file and it has 4 features (balance, age, factor, output) in which combination of 3 features (balance, age, factor) will determine the output value. Output feature and two input features (balance & factor) are of type floating
point and age is numeric.
Input has got around 350 records and 70% of which is passed as trained data to train model that uses Linear Regression Machine Learning algorithm.
Deployed the predictive webservice and while testing the model, I see some combination of data is returning negative output values. Very limited data combination of data is returning close results.
How to resolve this issue and to return scored predictions close to the output value? Please suggest.
推荐答案
您可以检查随机数种子"是否设置为特定数字吗?
Could you check to see if the 'random number seed' is set to a specific number?
对于随机数种子，您可以选择键入一个值来为模型使用的随机数生成器播种.
For Random number seed, you can optionally type a value to seed the random number generator used by the model.
如果您希望在同一实验的不同运行之间保持相同的结果，则可以使用种子值.否则，默认值为使用系统时钟中的值.
Using a seed value is useful if you want to maintain the same results across different runs of the same experiment. Otherwise, the default is to use a value from the system clock.
此致，
Jaya
Regards,
Jaya
这篇关于使用线性回归时不正确的Scord预测的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
