# Regression of exponential data

Question/Issue: I have been trying to make a regression model using edge impulse with exponential data. The accuracy rate is 0%. How can we make a regression with that data?

|1|999|
|2|790.625|
|3|351.3888889|
|4|197.65625|
|5|126.5|
|6|93.55|
|7|81.65|
|8|76.7|
|9|63.15|
|10|57.4|
|11|52.15|
|12|48.45|
|13|44.25|
|14|42.55|
|15|41.4|
|16|39.5|
|17|32.05|
|18|30.4|
|19|28.45|
|20|26.8|
|21|25.7|
|22|23.25|
|23|21.56|
|24|18.952|
|25|16.86|
|26|15.24|
|27|14.5|
|28|12.96|
|29|11.92|
|30|10.5|
|31|9.8|
|32|9.19|
|33|8.94|
|34|8.42|
|35|7.94|
|36|7.51|
|37|7.11|
|38|6.74|
|39|6.4|
|40|6.08|
|41|5.8|
|42|5.52|
|43|5.266|
|44|5.03|
|45|4.89|
|46|4.69|
|47|4.5|
|48|4.32|
|49|4.15|
|50|3.98|
|51|3.82|
|52|3.67|
|53|3.52|
|54|3.4|
|55|3.28|
|56|3.17|
|57|3.06|
|58|2.96|
|59|2.86|
|60|2.77|
|61|2.68|
|62|2.6|
|63|2.52|
|64|2.44|
|65|2.36|
|66|2.29|
|67|2.22|
|68|2.16|
|69|2.1|
|70|2.04|
|71|1.98|
|72|1.92|
|73|1.88|
|74|1.83|
|75|1.78|
|76|1.73|
|77|1.68|
|78|1.64|
|79|1.6|
|80|1.56|
|81|1.52|
|82|1.48|
|83|1.44|
|84|1.4|
|85|1.36|
|86|1.32|
|87|1.29|
|88|1.26|
|89|1.23|
|90|1.2|
|91|1.17|
|92|1.14|
|93|1.11|
|94|1.08|
|95|1.06|
|96|1.03|
|97|1.01|
|98|0.98|
|99|0.96|
|100|0.94|

Hello @Meom30,

In the model testing page, you can adjust your confidence threshold. Maybe it is too strict.
Could you share your project ID please so I can have a look to understand how you set this up

Best,

Louis

Thank you for replying, by mistake I deleted my project so I have created it again… this is project id 238947. Now the accuracy is showing 95%. However, when I do live classification it’s not showing the same result… Can you please help?